wenz iD - Proefschrift Ronald Willemse

Functional mapping of the sensorimotor cortex Clinical studies with MEG and fMRI

Ronald B. Willemse

Functional mapping of the sensorimotor cortex: Clinical studies with MEG and fMRI Š Ronald Willemse, 2016 No part of this thesis may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or otherwise without permission of the author.

ISBN: 9789462953291 Cover : Design: Printed by:

Aurora borealis, polar light (photograph by Dr. R.A.R. Gons) Wendy Schoneveld || www.wenz iD.nl Proefschriftmaken.nl || Uitgeverij BOXPress

The publication of this thesis was financially supported by Stichting Neurochirurgische Ontwikkeling Amsterdam

VRIJE UNIVERSITEIT

Functional mapping of the sensorimotor cortex: Clinical studies with MEG and fMRI

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. V. Subramaniam, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Geneeskunde op woensdag 8 juni 2016 om 13.45 uur in de aula van de universiteit, De Boelelaan 1105

door Ronald Bernardus Willemse geboren te Haarlem

promotoren: prof.dr. W.P. Vandertop

prof.dr. C.J. Stam

CONTENTS

Chapter 1

Introduction

Chapter 2

Magnetoencephalographic study of posterior tibial nerve stimulation in patients with intracranial lesions around the central sulcus

21

Chapter 3

Topographical organization of mu and beta band activity associated with hand and foot movements in patients with perirolandic lesions

39

Chapter 4

Slowing of M1 oscillations in brain tumor patients in resting state and during movement

53

Chapter 5

Magnetoencephalographic study of hand and foot sensorimotor organization in 325 consecutive patients evaluated for tumor or epilepsy surgery

71

Chapter 6

Spatiotemporal imaging of somatosensory cortical activity with identical paradigms: comparison of fMRI and MEG.

89

Chapter 7

Localisation of the central sulcus region in glioma patients with three-dimensional fluid-attenuated inversion recovery and volume rendering: comparison with functional and conventional magnetic resonance

111

Chapter 8

General discussion and future perspectives

127

Chapter 9

Summary

137

7

Chapter 10 Dutch Summary | Nederlandse samenvatting

143

Appendices List of abbreviations List of publications Acknowledgements | Dankwoord About the author Dissertations Brain Tumor Center Amsterdam

150 151 153 155 156

CHAPTER 1

Introduction

CHAPTER 1

BACKGROUND The aim of a neurosurgical procedure in a patient with a brain tumor is to maximize tumor resection, while minimizing significant postoperative neurological deficit. Therefore, knowledge of the anatomical relationship between a lesion and eloquent cortex is essential in the decision-making process prior to not only tumor resections, but also in procedures requiring resection of epileptogenic tissues, other intracranial pathology or a combination of these. In the specific case where lesions approach or involve the primary sensorimotor cortex, surgery can lead to postoperative motor and somatosensory deficits (28).

Magnetic resonance (MR) imaging is the standard imaging technique and can be used for identification of the central sulcus (CS) region. However, cortical anatomy is complex with large intra- and intersubject variability, making identification of specific cortical regions based on morphological features challenging (43). In addition, space-occupying lesions, such as brain tumors, or vascular malformations can complicate this procedure by distorting normal anatomy, making cortical identification difficult or even impossible.

The classical procedure for cortical identification is intraoperative electrical cortical stimulation (ECS) in awake patients, as first described by Wilder Penfield in 1959. However, this is a time-consuming procedure with prolongation of the operation time and patients have to be awoken during surgery, with the potential risk of inducing seizures with brain swelling (41).

However, despite major advances in functional neuroimaging, the gold standard in the treatment of certain brain tumors at this moment is still the awake-craniotomy with ECS (40). Preferably, presurgical evaluation with functional imaging should facilitate an optimal decision on the best surgical approach. Since intracranial lesions not only affect the brain structurally but also functionally, functional imaging can also be used to assess the influence of brain lesions on the sensorimotor network. For practical purposes and because this thesis is about cortical somatosensory and motor function we will focus on the sensorimotor cortex and the imaging modalities (MEG and fMRI) which were used.

SENSORIMOTOR CORTEX The primary sensorimotor cortex consists of the precentral gyrus (frontal lobe) and the postcentral gyrus (parietal lobe), which are separated by the CS. The anterior bank of the CS corresponds to the primary motor cortex (M1, Brodmann area 4) and the posterior bank corresponds to the primary somatosensory cortex (S1, Brodmann areas 1 - 3).

The primary motor cortex is involved in the initiation of movements and the activity of different body parts, which are topographically arranged (somatotopy). Other regions involved in motor control consist of the premotor cortex (PMC), the supplementary motor area (SMA) and the anterior cingulate motor cortex (CMC), regions generally involved in motor planning. The descending pathways from M1, PMC and SMA give rise to the corticospinal tract. At the level of the medulla, approximately 90% of the corticospinal

8

INTRODUCTION

fibers decussate to the contralateral spinal cord until they reach the Îą-motoneurons.

The primary somatosensory cortex (S1) consists of four different cyto-architectonic regions and each one of these regions shows a clear somatotopic organization (sensory homunculus), already present in the ascending pathways and thalamus (24).

Other important somatosensory cortical areas are: the secondary somatosensory cortex (S2), located on the superior bank of the lateral fissure (Sylvius), which can be activated bilaterally by unilateral stimulation and the posterior parietal cortex (PPC), which has an integrative function.

It is important to realize that the traditional cortical motor and sensory representations are not strictly separated but can be diffuse, overlapping and even changeable over time (7). For instance, unilateral movements are usually associated with activity in the contralateral M1, however ipsilateral or bilateral activity may also occur, depending on the complexity of the movement (59). However, the presence of intracranial disease may also contribute to altered activation patterns (19,37,56).

STRUCTURAL IMAGING Structural MR-imaging of the brain and to a lesser extent computed tomography (CT), are still the most important imaging modalities for presurgical identification and localization of intracranial lesions in relation to surrounding eloquent cortex. Morphological landmarks and patterns in these imaging modalities can be used for identification of the sensorimotor cortex.

The motor hand function is the most robust and frequently used anatomical landmark of M1, which can be identified by visual inspection of the magnetization prepared rapid angle gradient echo (MPRAGE) sequences of the MR scan, in a specific segment of the precentral gyrus (see Figure 1). The classic variant as described by Yousry et al. (1997) is knob-like, most often with the form of an inverted omega (Ί) or a horizontal epsilon (ξ) (63).

Several new variants have been described, along with gender differences (5). Reliable landmarks for motor foot and tongue areas are not available and can only be estimated from the hand motor area and the expected somatotopical distribution.

Unfortunately, localization of the (M1) hand area on the sole basis of anatomical landmarks may be unreliable in patients with lesions close to the CS due to possible lesion-induced anatomical displacement, large intra- and interindividual variability, or functional plasticity and reorganization (57). We studied an alternative structural imaging technique (3D-FLAIR) with functional properties due to differences in signal intensities, which could be used as an alternative imaging modality in the presence of intracranial lesions.

Alternatively, intraoperative electrical cortical stimulation or reversal of the somatosensory evoked potential (SEP) could be performed to localize the sensorimotor cortex, but advances in functional neuroimaging nowadays make routine pre-operative functional localization possible.

9

1

CHAPTER 1

Figure 1. Axial T1-weighted image demonstrating features of the sensorimotor cortex: the primary motor cortex (M1) located at the precentral gyrus, the primary somatosensory cortex (S1) at the postcentral gyrus, the central sulcus (yellow) and the motor hand knob with the form of an inverted omega.

FUNCTIONAL NEURO-IMAGING Functional neuroimaging provides a wide range of different noninvasive imaging modalities, making it possible to investigate the functional organization of the brain under normal and pathological conditions. These techniques usually offer complementary information with different spatial and temporal resolution.

Magnetoencephalography Magnetoencephalography (MEG) is a non-invasive functional imaging technique that provides a direct measure of neuronal activity with a millisecond time scale by recording magnetic fields associated with electrical neuronal activity (15).

Magnetic source imaging (MSI) is a combination of MEG with structural MR-imaging that estimates the location of electrical sources at the origin of the recorded magnetic fields with a high spatial resolution. This technique has been studied for identifying eloquent areas of the brain for neurosurgical planning and for use in localization of epileptic foci (13,30,50). The main advantage of MEG is the insensitivity by surrounding brain structures in contrast to electrical activity recorded by the electroencephalogram (EEG). For MEG instrumentation and technology we refer to available review articles (6,14,55).

10

INTRODUCTION

Source models Several source localization techniques are available to estimate the origin of non-invasively recorded magnetic brain activity. We will shortly outline the two most frequently used source models in preoperative functional localization of the sensorimotor cortex using MEG. In short, the interpretation of electromagnetic fields from the brain involves solving the inverse problem, i.e. the reconstruction of the distribution of the sources in the brain based on the measured signals. This problem has (by definition) no unique solution and therefore, assumptions have to be made to estimate source localizations. The most straightforward solution to this problem is presuming a very simplified model of neuronal generators by using the equivalent current dipole (ECD) model. This model determines the source current element (dipole) that would most closely explain the registered magnetic field pattern. Obviously, measured cortical activity is never a point-like dipole, but for clinical purposes the single ECD model turned out to be rather successful in presurgical localization of early cortical evoked responses (22,48).

The second source model is the beamformer, a spatial filtering technique that localizes spectral power in a pre-selected frequency band (e.g., beta band for motor function) in a certain time window. By contrasting activity that is recorded during task performance with the resting state one can identify regions, which are significantly activated as a result of task performance. Subsequently, signals can be projected from sensor space to source space in order to estimate the activity at the source (17).

Somatosensory evoked fields Somatosensory evoked magnetic fields (SEFs), can be used to map the somatosensory cortex by using direct electrical stimulation of a peripheral nerve (e.g. median nerve) or mechanical stimulation of the skin. The first cortical component of the median nerve SEF is called N20m, which is localized on the posterior bank (Brodmann areas 1 - 3) of the CS, with an accuracy of a few millimeters (27). Controversy exists about the generator of the middle-latency (P35m) component: some authors have suggested a contribution from M1 (27), while others have suggested sources deep in S1 (61).

The somatotopic organization can also be examined by using different stimulation points (44) for instance, stimulation of the posterior tibial nerve (PTN) at the ankle can be localized at the medial part of the CS with a latency of approximately 35 ms (P35m).

Electrical stimulation of the median nerve (MN) has been widely studied, however studies with respect to posterior tibial nerve (PTN) are limited especially in patients with intracranial lesions.

Movement-related magnetic activity Voluntary movements are associated with task-related oscillatory changes within the mu (8 – 13 Hz) and beta (13 – 30 Hz) frequency bands around the sensorimotor cortex, which can be measured with MEG using spatial filtering techniques (see Figure 2). We studied the effect of intracranial lesions on these mu and beta band activity and brain oscillations.

11

1

CHAPTER 1

Figure 2. Beta band power reduction for right hand movement overlaid on the surface-rendered brain (viewed from above, arbitrary threshold).

Functional MRI The fMRI technique is still the most widely used neuroimaging technique for noninvasive presurgical evaluation of sensorimotor systems (13,20,57) and has several advantageous characteristics. These include, noninvasiveness, relatively high spatial resolution and excellent imaging of the underlying disease (60). In addition, fMRI can be performed on most conventional MR-scanners and has enough sensitivity to create high-quality functional maps within minutes. However, the temporal resolution, in the order of seconds, is limited by the intrinsic hemodynamic response and finite signal-to-noise ratio (29).

For methods and technology of fMRI we refer to available textbooks/literature (3,16,38,39,46). The most widely used endogenous contrast mechanism in fMRI will be outlined shortly below.

Blood-oxygenation-level-dependent (BOLD) fMRI is derived from local hemodynamic changes, which accompany task activation. Brain activity is associated with increased energy requirements (increased oxidative glucose metabolism). After onset of brain activity a signal is sent to the feeding arteriole, to dilate. This dilation leads to an increase in cerebral blood flow (CBF) in downstream capillaries. However, the increase in blood flow is larger than the oxygen consumption, hence oxygen concentration increases, especially at the venous side. The resulting decrease in deoxy-hemoglobin is the basis of the BOLD signal (2,47). The relationship between fMRI and BOLD contrast was studied in nonhuman primate brains during visual stimulation suggesting that BOLD contrast is more correlated with postsynaptic activity (local field potentials) than with spiking activity (38), therefore it is assumed that the BOLD signal is a composite of both neural and vascular reactivity.

Several publications with carefully done correlation analysis between preoperative fMRI and intraoperative cortical mapping have shown good correspondence between the two (35,49,62) with respect to presurgical localization of the primary motor cortex, which is the most employed application of presurgical fMRI. Motor-mapping using fMRI can be done with repetitive movements of the hand or foot and this is the most frequently used

12

INTRODUCTION

paradigm. Sensory mapping using fMRI is more difficult than the reliable motor activation as the BOLD signal is low and frequently shows overlap with M1 (52).

Since fMRI is based on the hemodynamic response, it is thought to be less powerful in lesions that lead to changes in the vascular auto-regulation, such as gliomas or cerebral ischemia (18,37). This could result in false positive and false negative detection of cortical sensorimotor areas in some patients (20,21,32,34) as compared with intraoperative mapping. In addition, its temporal resolution makes discrimination between premotor, motor, and somatosensory components associated with hand movements difficult (32).

Consequently, interpretation of fMRI maps may be more challenging in patients with brain lesions than in healthy subjects (21). Furthermore, it is still unclear to what extent the resection border could approach the functional MR-representation, due to the lack of a standardized approach or multicenter randomized trials addressing this issue.

In cases where fMRI is not possible or feasible, especially in pediatric patients, patients with claustrophobia, carrying ferromagnetic material, or harboring vascular abnormalities other brain mapping modalities can be considered as an option.

MULTIMODAL IMAGING MEG detects electrical activity in the brain with millisecond temporal resolution and fMRI measures hemodynamic phenomena. As a result of measuring different physiological functions, discrepancies may be expected. Therefore, it is important to be aware of the strengths and limitations of each modality when comparing different imaging modalities.

The first comparison of MEG and fMRI was performed with a visual motion paradigm in 1999 (1). Later, other combined both modalities for sensorimotor research in healthy subjects (4,33,53) but also in patients (31,32).

In neurosurgical patients, studies showed that combining (sensorimotor) fMRI and MEG may increase the localization reliability and that MEG may be superior to fMRI in some patients with unclear fMRI localization (20,23,32). Usually prominent activation of nonprimary areas in fMRI could lead to false localization of the CS in some cases. However, despite these differences, the increased reliability of combined sensorimotor mapping may be useful in neurosurgical practice to avoid postoperative deficits.

FUNCTIONAL SENSORIMOTOR REORGANIZATION MEG can be used to localize the primary sensorimotor cortex in relation to intracranial lesions and address whether so-called functional reorganization or cerebral plasticity exists. In healthy subjects, simple motor tasks elicit contralateral activation of M1 and somatosensory stimulation causes early S1 activation. Patients with intracranial lesions can show altered patterns of activation, varying from subtle shifts from the functional region involved to complete lateralization (10,54,58).

13

1

CHAPTER 1

Indications of functional reorganization may have implications for pre-operative mapping of resection areas and may explain the diversity of postoperative deficits in patients who harbor tumors close to or involving the CS area. Especially in patients with slow-growing tumors other brain areas can be recruited to compensate for expected neurological deficits (9). Further studies of the functional alterations in the sensorimotor network as a result of intracranial pathology may aid in the understanding of the complex network involved in sensorimotor functioning and possibly in the management of neurosurgical patients. In this thesis we studied the effects of intracranial lesions and epilepsy on the MEG mapping results of the sensorimotor cortex, especially in cases where the expected functional areas may have shifted or become lateralized.

AIMS OF THIS THESIS The primary objective was to investigate how brain lesions affect the results of sensorimotor cortical activity as studied with MEG. A second objective was to assess the relationship between fMRI and MEG with respect to the activation of somatosensory cortical activity. Third, we investigated whether volume rendering of a structural MR scan could be used as an alternative method to localize the CS. The following research questions were addressed in this thesis: 1. do patients with brain tumors around the CS show interhemispheric differences with respect to neuromagnetic aspects of PTN stimulation in relation to clinical findings?

2. do patients with brain tumors show altered patterns of movement-related oscillatory activity?

3. in what respect do hand and foot sensorimotor cortical activity differ in presurgical candidates and can differences be attributed to the underlying intracranial pathology?

4. what is the spatiotemporal relationship of somatosensory cortical activity after MN stimulation measured with MEG and fMRI in healthy subjects?

5. can the CS be reliably localized in brain tumor patients using volume rendering of an isotropic 3D FLAIR MR technique?

OUTLINE OF THE THESIS Sensorimotor MEG and brain lesions The main goal of this thesis was to assess the influence of brain lesions on the primary sensorimotor cortical activity, using MEG. Despite enormous advances in functional neuroimaging, including MEG, a number of aspects have not received much attention. Neuroimaging studies with respect to intracranial lesions and CS localizations are usually based on the somatosensory or motor aspects of the hand. Few studies have been published with respect to the foot. From a functional point of view and its importance in walking,

14

INTRODUCTION

this may be as important as the hand. Therefore, knowledge of possible changes in cortical activity of the foot area in brain tumor patients could facilitate clinical decision making and planning in patients with intracranial lesions and may support hypotheses regarding cortical reorganization. In Chapter 2 we examined whether unilateral tumors around the CS cause interhemispheric differences with respect to SEF characteristics and spatial distribution after PTN stimulation in relation to clinical findings.

Altered activation patterns of the motor cortex have been described in the presence of intracranial lesions using fMRI (11,37,45). Functional localization of motor cortex activity using MEG with single ECD modeling was found to be clinically insufficient (36). Cortical sources of motor activity could also be studied by measuring movement-related oscillatory changes, but studies in patients are limited (12,42,56) and data about the foot motor cortex are lacking. In Chapter 3 we investigated whether intracranial tumors around the CS affect movement-related oscillatory changes of the hand and foot with respect to topographical organization. A more in-depth analysis of the influence of brain tumors on the motor network was performed in Chapter 4. Since brain tumors can disrupt white matter fibers, communication between different brain regions can be affected. In the absence of motor deficits we studied whether mu and beta oscillations of M1 are affected by the presence of a brain tumor.

The previous three chapters mostly deal with intracranial glioma. It is conceivable that different lesions affect sensorimotor cortical activity in different ways. Chapter 5 evaluates the results of MEG sensorimotor mapping in a large heterogeneous group of patients eligible for tumor or epilepsy surgery, to see whether differences in localization and lateralization results of the hand and foot could be explained by the underlying pathology.

Multimodal imaging Previous studies comparing fMRI versus MEG and somatosensory stimulation used different stimulation techniques for fMRI and MEG (8,33). Another study used identical stimulus frequencies with pneumatically driven impulses but only to one side (51).

Moreover, these studies focused on the short-latency responses. In Chapter 6 we combined information from MEG and fMRI to assess which cortical areas and in which temporal order show activation after well-controlled identical MN stimulation in healthy subjects.

Structural imaging Structural MR imaging is essential in the evaluation of neurosurgical patients and different anatomical landmarks can be used for identification of the sensorimotor cortex. An alternative identification of the sensorimotor cortex is possible due to the lower signal intensities as observed on turbo fluid-attenuated inversion recovery (FLAIR) MR images in the normal population (25,26). With the development of single-slab methods with T2 and FLAIR contrast it was possible to obtain 3D FLAIR images with isotropic voxels. Volume rendering of this 3D FLAIR showed hypo-intense cortical regions corresponding to the CS region and occipital cortex. In Chapter 7 we studied whether volume rendering of an

15

1

CHAPTER 1

isotropic 3D FLAIR MR scan could be used as an alternative or complimentary method to localize the CS in comparison to fMRI and conventional structural MR-images.

Finally, in Chapter 8, we summarize and discuss the results presented in Chapters 2 to 7 against recent developments and present future directions of research.

16

INTRODUCTION

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Ahlfors SP, Simpson GV, Dale AM, et al. Spatiotemporal activity of a cortical network for processing visual motion revealed

by MEG and fMRI. J Neurophysiol 1999;82:2545-55.

Bandettini PA, Wong EC, Hinks RS, Tikofsky RS & Hyde JS. Time course EPI of human brain function during task activation.

Magn Reson Med 1992;25:390-7.

Belliveau JW, Rosen BR, Kantor HL, et al. Functional cerebral imaging by susceptibility-contrast NMR. Magn Reson Med 1990;14:538-46.

Bourguignon M, Jousm채ki V, Marty B, et al. Comprehensive functional mapping scheme for non-invasive primary sensorimotor cortex mapping. Brain Topogr 2013;26:511-23.

Caulo M, Briganti C, Mattei PA, et al. New morphologic variants of the hand motor cortex as seen with MR imaging in a large study population. AJNR Am J Neuroradiol 2007;28:1480-5.

Cheyne DO. MEG studies of sensorimotor rhythms: a review. Exp Neurol 2013;245:27-39.

Crone NE, Miglioretti DL, Gordon B, et al. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. I. Alpha and beta event-related desynchronization. Brain 1998;121:2271-99.

Del Gratta C, Della Penna S, Ferretti A, et al. Topographic organization of the human primary and secondary somatosensory cortices: comparison of fMRI and MEG findings. Neuroimage 2002;17:1373-83.

Desmurget M, Bonnetblanc F & Duffau H. Contrasting acute and slow-growing lesions: a new door to brain plasticity. Brain 2007;130:898-914.

Duffau H. Lessons from brain mapping in surgery for low-grade glioma: insights into associations between tumour and brain plasticity. Lancet Neurol 2005;4:476-86.

Fandino J, Kollias SS, Wieser HG, Valavanis A & Yonekawa Y. Intraoperative validation of functional magnetic resonance imaging and cortical reorganization patterns in patients with brain tumors involving the primary motor cortex. J Neurosurg 1999;91:238-50.

Gaetz W, Cheyne D, Rutka JT, et al. Presurgical localization of primary motor cortex in pediatric patients with brain lesions by the use of spatially filtered magnetoencephalography. Neurosurgery 2009;64:177-85; discussion 186.

Ganslandt O, Fahlbusch R, Nimsky C, et al. Functional neuronavigation with magnetoencephalography: outcome in 50 patients with lesions around the motor cortex. Neurosurg Focus 1999;6:e3.

Hari R & Salmelin R. Magnetoencephalography: From SQUIDs to neuroscience. Neuroimage 20th anniversary special edition. Neuroimage 2012;61:386-96.

H채m채l채inen M, Hari R, Ilmoniemi RJ, Knuutila J & Lounasmaa OV. Magnetoencephalographytheory, instrumentation, and applications to noninvasive studies of the working human brain. Reviews of modern Physics 1993;65:413.

Heeger DJ & Ress D. What does fMRI tell us about neuronal activity? Nat Rev Neurosci 2002;3:142-51.

Hillebrand A, Singh KD, Holliday IE, Furlong PL & Barnes GR. A new approach to neuroimaging with magnetoencephalography.

Hum Brain Mapp 2005;25:199-211.

Holodny AI, Schulder M, Liu WC, Wolko J, Maldjian JA & Kalnin AJ. The effect of brain tumors on BOLD functional MR imaging activation in the adjacent motor cortex: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 2000;21:1415-22.

Hou BL, Bradbury M, Peck KK, Petrovich NM, Gutin PH & Holodny AI. Effect of brain tumor neovasculature defined by rCBV on BOLD fMRI activation volume in the primary motor cortex. Neuroimage 2006;32:489-97.

Inoue T, Shimizu H, Nakasato N, Kumabe T & Yoshimoto T. Accuracy and limitation of functional magnetic resonance imaging for identification of the central sulcus: comparison with magnetoencephalography in patients with brain tumors.

Neuroimage 1999;10:738-48.

Jiang Z, Krainik A, David O, et al. Impaired fMRI activation in patients with primary brain tumors. Neuroimage 2010;52:53848.

Kakigi R, Hoshiyama M, Shimojo M, et al. The somatosensory evoked magnetic fields. Prog Neurobiol 2000;61:495-523.

Kamada K, Houkin K, Takeuchi F, et al. Visualization of the eloquent motor system by integration of MEG, functional, and anisotropic diffusion-weighted MRI in functional neuronavigation. Surg Neurol 2003;59:352-61; discussion 361-2.

Principles of neural science. McGraw-Hill Companies, Inc., 2012, 2012.

Karaarslan E & Arslan A. Perirolandic cortex of the normal brain: low signal intensity on turbo FLAIR MR images. Radiology 2003;227(2):538-41.

Karaarslan E & Arslan A. ROI measurement of the signal intensity of precentral cortex in the normal brain. Eur J Radiol 2004;52:221-3.

Kawamura T, Nakasato N, Seki K, et al. Neuromagnetic evidence of pre- and post-central cortical sources of somatosensory evoked responses. Electroencephalogr Clin Neurophysiol 1996;100:44-50.

17

1

CHAPTER 1

28. Keles GE, Lundin DA, Lamborn KR, Chang EF, Ojemann G & Berger MS. Intraoperative subcortical stimulation mapping for hemispherical perirolandic gliomas located within or adjacent to the descending motor pathways: evaluation of morbidity and assessment of functional outcome in 294 patients. J Neurosurg 2004;100:369-75.

29. Kim SG, Richter W & Uğurbil K. Limitations of temporal resolution in functional MRI. Magn Reson Med 1997;37:631-6.

30. Knowlton RC. Can magnetoencephalography aid epilepsy surgery? Epilepsy Currents 2008;8:1-5.

31. Kober H, Nimsky C, Möller M, Hastreiter P, Fahlbusch R & Ganslandt O. Correlation of sensorimotor activation with functional magnetic resonance imaging and magnetoencephalography in presurgical functional imaging: a spatial analysis.

Neuroimage 2001;14:1214-28.

32. Korvenoja A, Kirveskari E, Aronen HJ, et al. Sensorimotor cortex localization: comparison of magnetoencephalography, functional MR imaging, and intraoperative cortical mapping. Radiology 2006;241:213-22.

33. Korvenoja A, Huttunen J, Salli E, et al. Activation of multiple cortical areas in response to somatosensory stimulation: combined magnetoencephalographic and functional magnetic resonance imaging. Hum Brain Mapp 1999;8:13-27.

34. Krings T, Reinges MH, Erberich S, et al. Functional MRI for presurgical planning: problems, artefacts, and solution strategies.

J Neurol Neurosurg Psychiatry 2001;70:749-60.

35. Lehéricy S, Duffau H, Cornu P, et al. Correspondence between functional magnetic resonance imaging somatotopy and individual brain anatomy of the central region: comparison with intraoperative stimulation in patients with brain tumors.

J Neurosurg 2000;92:589-98.

36. Lin PT, Berger MS & Nagarajan SS. Motor field sensitivity for preoperative localization of motor cortex. J Neurosurg 2006;105:588-94.

37. Liu W-C, Feldman SC, Schulder M, et al. The effect of tumour type and distance on activation in the motor cortex.

Neuroradiology 2005;47:813-9.

38. Logothetis NK, Pauls J, Augath M, Trinath T & Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature 2001;412:150-7.

39. Menon RS. The great brain versus vein debate. Neuroimage 2012;62:970-4.

40. Miller KJ, denNijs M, Shenoy P, Miller JW, Rao RPN & Ojemann JG. Real-time functional brain mapping using electrocorticography. Neuroimage 2007;37:504-7.

41. Motamedi GK, Okunola O, Kalhorn CG, et al. Afterdischarges during cortical stimulation at different frequencies and intensities. Epilepsy Res 2007;77:65-9.

42. Nagarajan S, Kirsch H, Lin P, Findlay A, Honma S & Berger MS. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109:228-37.

43. Naidich TP, Grant JL, Altman N, et al. The developing cerebral surface. Preliminary report on the patterns of sulcal and gyral maturation--anatomy, ultrasound, and magnetic resonance imaging. Neuroimaging Clin N Am 1994;4:201-40.

44. Nakamura A, Yamada T, Goto A, et al. Somatosensory homunculus as drawn by MEG. Neuroimage 1998;7:377-86.

45. Niu C, Zhang M, Min Z, et al. Motor network plasticity and low-frequency oscillations abnormalities in patients with brain gliomas: a functional MRI study. PLoS One 2014;9:e96850.

46. Ogawa S, Menon RS, Tank DW, et al. Functional brain mapping by blood oxygenation level-dependent contrast magnetic resonance imaging. A comparison of signal characteristics with a biophysical model. Biophys J 1993;64:803-12.

47. Ogawa S, Lee TM, Kay AR & Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation.

Proc Natl Acad Sci U S A 1990;87:9868-72.

48. Ossenblok P, Leijten FSS, de Munck JC, Huiskamp GJ, Barkhof F & Boon P. Magnetic source imaging contributes to the presurgical identification of sensorimotor cortex in patients with frontal lobe epilepsy. Clin Neurophysiol 2003;114:221-32.

49. Roessler K, Donat M, Lanzenberger R, et al. Evaluation of preoperative high magnetic field motor functional MRI (3 Tesla) in glioma patients by navigated electrocortical stimulation and postoperative outcome. J Neurol Neurosurg Psychiatry 2005;76:1152-7.

50. Schiffbauer H, Berger MS, Ferrari P, Freudenstein D, Rowley HA & Roberts TPL. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97:133342.

51. Schulz M, Chau W, Graham SJ, et al. An integrative MEG-fMRI study of the primary somatosensory cortex using cross-modal correspondence analysis. Neuroimage 2004;22:120-33.

52. Schweizer R, Voit D & Frahm J. Finger representations in human primary somatosensory cortex as revealed by highresolution functional MRI of tactile stimulation. Neuroimage 2008;42:28-35.

53. Stippich C, Freitag P, Kassubek J, et al. Motor, somatosensory and auditory cortex localization by fMRI and MEG. Neuroreport 1998;9:1953-7.

54. Stoeckel MC & Binkofski F. The role of ipsilateral primary motor cortex in movement control and recovery from brain damage. Exp Neurol 2010;221:13-7.

18

INTRODUCTION

55. Stufflebeam SM. Clinical magnetoencephalography for neurosurgery. Neurosurg Clin N Am 2011;22:153-67, vii-viii.

56. Taniguchi M, Kato A, Ninomiya H, et al. Cerebral motor control in patients with gliomas around the central sulcus studied with spatially filtered magnetoencephalography. J Neurol Neurosurg Psychiatry 2004;75:466-71.

57. Towle VL, Khorasani L, Uftring S, et al. Noninvasive identification of human central sulcus: a comparison of gyral morphology, functional MRI, dipole localization, and direct cortical mapping. Neuroimage 2003;19:684-97.

58. Tozakidou M, Wenz H, Reinhardt J, et al. Primary motor cortex activation and lateralization in patients with tumors of the central region. Neuroimage (Amst) 2013;2:221-8.

59. Verstynen T, Diedrichsen J, Albert N, Aparicio P & Ivry RB. Ipsilateral motor cortex activity during unimanual hand movements relates to task complexity. J Neurophysiol 2005;93:1209-22.

60. Wengenroth M, Blatow M, Guenther J, Akbar M, Tronnier VM & Stippich C. Diagnostic benefits of presurgical fMRI in patients with brain tumours in the primary sensorimotor cortex. Eur Radiol 2011;21:1517-25.

61. Wikstrรถm H, Huttunen J, Korvenoja A, et al. Effects of interstimulus interval on somatosensory evoked magnetic fields (SEFs): a hypothesis concerning SEF generation at the primary sensorimotor cortex. Electroencephalogr Clin Neurophysiol 1996;100:479-87.

62. Yetkin FZ, Mueller WM, Morris GL, et al. Functional MR activation correlated with intraoperative cortical mapping. AJNR Am J Neuroradiol 1997;18:1311-5.

63. Yousry TA, Schmid UD, Alkadhi H, et al. Localization of the motor hand area to a knob on the precentral gyrus. A new landmark. Brain 1997;120 ( Pt 1):141-57.

19

1

CHAPTER 2

Magnetoencephalographic study of posterior tibial nerve stimulation in patients with intracranial lesions around the central sulcus Neurosurgery. 2007; 61(6): 1209-17

Ronald B. Willemse Jan C. de Munck Dennis van ’t Ent Peterjan Ris Johannes C. Baayen Cornelis J. Stam W. Peter Vandertop

CHAPTER 2

ABSTRACT Objective To study interhemispheric differences of somatosensory evoked field (SEF) characteristics and the spatial distribution of equivalent current dipole (ECD) sources in patients with unilateral hemispheric lesions around the central sulcus (CS) region.

Methods In seventeen patients with perirolandic lesions, averaged somatosensory responses after posterior tibial nerve (PTN) stimulation at the ankle were recorded with magnetoencephalography (MEG). Dipole source solutions in the affected (AH) and unaffected (UH) hemisphere were analyzed and compared for latency, ECD strength, root mean square (RMS) and spatial distribution in relation to clinical findings.

Results Three main SEF components, P45m, N60m, and P75m were identified in the hemisphere contralateral to the stimulated nerve. Dipole strength for the P45m component was significantly higher in the AH compared with the UH. SEF characteristics in the AH and UH showed no significant differences with respect to component latency or dipole strength of the N60m and P75m component. Inter-dipole location asymmetries exceeded 1.0 cm in 71% of the cases.

Comparison of the PTN evoked responses (P45m and N60m) in patients with motor deficits and patients without deficits showed that these responses are enlarged in the AH when perirolandic lesions are present. Patients with motor deficits also showed an increased response for P45m in the UH. Conclusions The results of posterior tibial nerve SEFs suggest spatial and functional changes in the somatosensory network as a result of perirolandic lesions, with a possible relationship with clinical symptoms. The results can provide further basis for the evaluation of cortical changes in the presence of perirolandic lesions.

22

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

INTRODUCTION For presurgical planning or intra-operative neuronavigation, somatosensory evoked responses to electrical stimulation of the median nerve at the wrist or pneumatical stimulation of the fingers, can be applied to localise the primary somatosensory cortex (SI) in patients with intracranial lesions (5,43,48). In magnetoencephalography (MEG), the early component (N20m) of the median nerve somatosensory evoked field (SEF) can be localized to the contralateral SI using a single equivalent current dipole (ECD) model (11,12,18,23,32,40). Intra-operative recordings have shown this localization to be very reliable (12,40).

In contrast to median nerve SEFs, there are only a few MEG studies of lower limb stimulation (8,13,20,31,38,40). MEG allows better spatial localization than EEG, because it is less distorted by the tissue layers between the electrical source and recording sensors (19,41) but intra-operative recordings of the lower limb area in humans are difficult due to the fact that the lower limb representation is located deep within the principal sulcus.

MEG studies on SEF responses following posterior tibial nerve (PTN) stimulation at the ankle in normal controls, report four main components, at 37 ms, 45 ms, 60 ms and 75 ms (13,20). The underlying sources of these components were identified around the foot area of the primary somatosensory (SI) cortex contralateral to the stimulated nerve, with dipole orientations rotating as a function of post-stimulus-latency (13). However, the spatiotemporal SEF characteristics of PTN stimulation in the presence of intracranial lesions are largely unknown.

In recent studies, MEG has been specifically applied to investigate the changes in cortical response to sensory stimulation in patients with unilateral pathological processes (33,37).

It is conceivable that different lesions affect the somatosensory network in different ways.

Therefore, knowledge about the structural as well as functional changes of the network in the presence of intracranial lesions has a potentially clinical significance for presurgical planning (4).

The objective of this study was to investigate interhemispheric differences in SEF components following PTN stimulation in patients harbouring perirolandic lesions in relation to clinical findings. To our knowledge, this is the first systematic study with respect to PTN-SEF characteristics in a population with perirolandic lesions.

METHODS Patients From the patients with intracranial lesions, referred to the department of Neurosurgery of the VU University Medical Center, seventeen consecutive patients with unilateral intracranial lesions around the central sulcus (CS) region eligible for treatment (eight female and nine male, age range 34 - 69 years; mean ± SD: 48.1 ± 11.6 years ) with a Karnofsky Performance Scale score ≥ 70, age ≤ 70 and successful MEG test-procedure.

23

2

CHAPTER 2

All patients had a neurological and radiological examination and volumetric data of the lesions was calculated from the MR segmentation images (Vector Vision Planning, BrainLAB AG, Heimstetten, Germany). The study was approved by the Medical Ethics Committee of the VU University Medical Center and informed consent from the participants was obtained prior to inclusion.

MEG Recordings Stimulation was applied to the ankle and delivered to the skin surface by two nickel electrodes (Electrical Stimulator: Grass, USA; model S48). The stimulation frequency was 2 Hz and electric pulse duration was 0.2 ms. Stimulation intensity was adjusted individually, starting at a low value (3-4 mA) and progressively increased until clear twitches of the toe were observed. During stimulation, subjects were lying or seating comfortably with eyes open, inside the three-layer magnetically shielded room (Vacuum Schmeltze Gmbh, Germany).

MEG was recorded with a system of 151 first-order axial gradiometers (VSM MedTech Ltd., Canada), with a helmet shape detector array covering the whole head. For the SEF recordings a pre-stimulus baseline time of 100 ms was used. About 500 epochs were measured with a sampling frequency of 1250 Hz. Data were online lowpass filtered at 400 Hz. Structural magnetic resonance imaging (MRI) scans were acquired with a 1.5 T MR scanner (Siemens Sonata, Erlangen, Germany) using T1-weighted magnetization prepared rapid acquisition gradient echo (MPRAGE) sequence. In cases with known contrast-enhancement of the lesion, gadolinium contrast was used.

Matching of functional and anatomical data Co-registration of MEG and MRI scan was performed using fiducial markers on the nasion and left- and right preauricular points. The markers were contained in specially developed marker sockets (BrainLAB AG) and kept in place until both MEG and MRI measurements had been performed to minimize the errors involved in data co-registration. In MEG small coils energized by AC currents and in MRI vitamin E capsules were used as markers. MRI registration was performed after MEG registration, with MRI markers in the same sockets and positions as the MEG coils, enabling matching of the datasets with an estimated precision of 2 mm (7).

The fiducial points were used to define a nasion, left ear, right ear (NLR) coordinate system.

In NLR the midpoint between left and right pre-auricular points defines the origin. The x and y axes in the plane formed by the three fiducials, with the x-axis directed through the nasion (positive values towards nasion) and y axis perpendicular to the x axis (positive values towards the left ear). The z axis is perpendicular to the x-y plane, with positive values upward.

Source localization procedure Trials or channels with artefacts (eye movements, muscle activity) were excluded. A conventional single equivalent current (moving) dipole was fitted in a least-squares sense

24

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

to the average SEFs using a homogeneous spherical volume conductor model which is usually sufficient to obtain a satisfactory solution (26). A single moving dipole model was used as dipole model because one of our goals was to compare source strengths.

The stationary multiple dipole model, often yields nearby and opposing dipoles of unrealistically large amplitudes. The spatial coordinates (x, y, z positions), ECD strength (Q; nanoamperemetre, nAm) and the root mean square (RMS; femtotesla, fT) of a bestfitting single ECD were estimated for each SEF component. SEF components were selected in the time interval of 30 to 130 ms post-stimulus with residual errors of fitted dipoles < 25 %. Since field patterns of clear SEF deflections in the 130 ms poststimulus window frequently show a clear dipolar pattern, ECDs explaining at least 75% of the field variance were used. The expected dipole location in the affected hemisphere compared with homologeous dipole sources in the unaffected hemisphere was used as a measure of spatial displacement in the affected hemisphere. For computing these inter-dipole distances, dipoles after right side stimulation in the left hemisphere were translated to the right hemisphere by inverting the sign of the y-coordinate (│y│). Then the distance could be calculated by taking the square root of the sum of the squares of the dipole differences of x, │y│ and z.

Statistical Analysis Latencies, ECD strengths, RMS and spatial coordinates of the dipoles were compared and tested statistically, using paired two-tailed t-tests. Statistical significance was determined at an alpha level of 0.05 and data are presented as mean ± standard deviation. The relationship between tumor volume and SEF characteristics was measured by correlation analysis. All statistics were performed using SPSS software (SPSS, Inc., Chicago, IL).

RESULTS Table 1 summarizes the clinical data of the patient group. Nine patients had an astrocytoma (six World Health Organization [WHO] grade II, three WHO grade IV), two patients had an oligodendroglioma, two a cavernoma, one a meningioma, one an arteriovenous malformation (AVM), and one multiple sclerosis (MS), initially suspected to be a glioma.

One patient had a lesion without tissue diagnosis with radiological features of a low-grade glioma (LGG). Lesions were located on the right side in eight cases, on the left in nine cases. Lesion volume ranged from 2 – 110 ml (mean ± SD: 35.5± 27.7 ml). Five patients had neurological deficits consisting of slight to moderate muscle weakness. The remaining twelve patients presented with seizures. Two patients had no surgery, one had a biopsy procedure (case 4) and fourteen had surgery. The postoperative neurological status was unchanged in 8 (53 %), 2 (13 %) had a transient motor deficit, 3 (20 %) had a new permanent mild deficit and 2 (13%) improved.

Table 2 lists the first three SEF components in the short- and middle latency range and the number of patients in which the SEF contained this component. The corresponding

25

2

CHAPTER 2

Patient No.

Age (years) / Sex

Diagnosis

Location, Lateralization

Lesion volume (ml)

Symptoms

Postop.

Table 1. Clinical data patient groupa.

1

43 / F

astrocytoma II

frontoparietal, R

32

seizures

permanent mild deficit

2

41 / F

AVM

frontotemporal, L

36

weakness L-arm/leg

no surgery

3

64 / M

oligodendroglioma

frontal, L

61

seizures

transient deficit

4

35 / F

MS

parietal, L

21

weakness R-leg

improved

5

52 / M

astrocytoma IV

frontotemporal, L

25

weakness R-arm/leg

unchanged

6

42 / M

astrocytoma IV

frontal, L

85

seizures

unchanged

7

60 / M

cavernoma

parietal, R

2

seizures

unchanged

8

38 / F

astrocytoma II

frontal, R

48

seizures

unchanged

9

38 / M

astrocytoma II

parietal, L

110

weakness R-arm/leg

new mild deficit

10

68 / M

astrocytoma IV

frontal, L

19

seizures

permanent mild deficit

11

69 / F

meningeoma

frontoparietal, L

25

weakness R-arm/leg

improved

12

47 / M

astrocytoma II

parietal, R

35

seizures

unchanged

13

49 / F

oligodendroglioma

frontoparietal, R

21

seizures

transient deficit

14

48 / F

astrocytoma II

parietal, R

32

seizures

unchanged

15

34 / M

astrocytoma II

parietal, R

26

seizures

unchanged

16

35 / F

LGG

parietal, R

23

seizures

no surgery

17

54 / M

cavernoma

frontal, L

2

seizures

unchanged

a

F, female; M, male; AVM, arteriovenous malformation; MS, multiple sclerosis; LGG, low-grade glioma; R, right; L, left.

Table 2. Somatosensory evoked field characteristicsa.

Deflection

N

Variable

AHb

UHb

P-value

P45m

17

Latency (ms)

47.9 ± 4.6

47.6 ± 5.9

0.765

Q (nAm)

20.6 ± 14.6

15.2 ± 8.0

0.021

RMS (fT)

33.8 ± 19.5

26.7 ± 11.6

0.020

Latency (ms)

67.1 ±13.6

67.0 ± 14.3

0.980

Q (nAm)

23.1 ± 16.7

23.9 ± 14.6

0.834

RMS (fT)

34.4 ± 19.4

31.8 ± 13.7

0.531

Latency (ms)

88.3 ± 15.1

80.9 ± 9.1

0.318

Q (nAm)

26.1 ± 28.3

24.2 ± 17.4

0.895

RMS (fT)

37.5 ± 40.6

37.6 ± 27.9

0.995

N60m

P75m

14

7

N, number of patients showing corresponding deflection; AH, affected hemisphere; UH, unaffected hemisphere; Q, equivalent current dipole strength; nAm, nanoamperemetre; RMS, root mean square; fT, femtotesla. b mean ± standard deviation. a

26

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

mean latency, ECD strength (Q) and RMS for the affected (AH) and unaffected (UH) hemisphere are also listed. All patients had a first component with a mean latency of 47.9 and 47.6 ms in the AH and UH respectively. Fourteen patients also had the second component with a latency of 67.1 ms in the AH and 67.0 ms in the UH. Seven patients in the affected group and seven in the unaffected group had all three components. The mean latencies of all three components showed no statistical significant difference between groups. With respect to the first SEF component (P45m), a statistical significant (P < 0.05) enlarged ECD strength and RMS were observed in the AH compared to the UH. The second (N60m) and third (P75m) SEF component showed no statistical difference in ECD strength and RMS. Dipole coordinates (x, │y│, z) and inter-dipole distances between homologous dipoles are shown in Table 3. The spatial characteristics of all three SEF components showed no interhemispheric statistical significant differences. No systematic displacement direction of the SEF components was found. However, homologous dipoles did show a considerable inter-dipole distance. Location asymmetries exceeding 1.0 cm were observed in 12/17 (71%) patients for the first SEF component, in 12/14 (86 %) patients for the second and 4/5 (80 %) for the third component. The mean residual error of the dipoles was 10.5 %.

Analysis of the subgroup of five patients with motor deficits in relation to the patients without neurological deficit showed significant differences for the P45m and N60m component with respect to ECD strength in the AH, showing enlarged dipole strengths in the patients with motor deficits (Table 4). In patients with motor deficits the UH exhibited an increased ECD strength compared to the UH in patients without motor deficits. No relationship was found between tumor volume and SEF characteristics.

Table 3. Spatial characteristics of dipole source solutiona.

AH Deflection

X (cm)

│Y│(cm)

UH

Location Difference (D)

Z (cm)

X (cm)

│Y│ (cm)

Z (cm)

9.56 ± 0.93

-0.24 ± 1.34

1.09 ± 0.66

9.37 ± 0.90

1.33 ± 0.97 12/17 (71%)c

D (cm)

D > 1.0 cmb

P45m

-0.58 ± 0.71b 0.97 ± 0.44

N60m

-0.39 ± 1.03

0.89 ± 0.50

9.34 ± 1,00

-0.25 ± 0.89

1.06 ± 1.48

8.36 ± 2.21

2.08 ± 1.70

12/14 (86%)

P75m

-0.47 ± 0.73

1.13 ± 0.81

8.77 ± 1.28

-0.67 ± 1.16

1.21 ± 0.64

8.97 ± 0.84

1.58 ± 0.82

4/5 (80%)

AH, affected hemisphere; UH, unaffected hemisphere; D, inter-dipole distance between homologous dipoles. Data are given in mean ± standard deviation.b Number of patients (%) with location asymmetries exceeding 1.0 cm a

Table 4. Dipole strength and clinical motor functiona .

Deflection

Variable

Motor deficit (N = 5)b

No deficit (N = 12)b

P-value

P45m

Q AH

33.4 ± 22.1b

15.3 ± 5.4

0.015

Q UH

22.2 ± 8.1

13.9 ± 5.7

0.030

Q AH

34.9 ± 22.4

16.6 ± 8.2

0.043

Q UH

29.8 ± 17.1

20.6 ± 12.9

0.277

N60m a

Q, equivalent current dipole strength; AH, affected hemisphere; UH, unaffected hemisphere. b mean ± standard deviation.

27

2

CHAPTER 2

Illustrative Cases Patient 4 A 35-year-old woman had suffered from progressive weakness of her right leg, and discrete motor disturbances of her right hand. An anatomic MRI scan showed a left-sided, medioparietal lesion, partially cystic, with ring-like contrast-enhancement after gadolinium and slight perilesional edema. Before treatment the PTN dipole localization showed activation within the lesion for the first SEF component (Figure 1A) and localization at the lesion margin for the second SEF component (Figure 1B).

Because an attempt to resect the lesion was judged inappropriate given the high risk of increased neurological deficit, a neuronavigation-assisted biopsy was performed with aspiration of cystic fluid. The histological diagnosis was consistent with a demyelinating disorder. Postoperatively the weakness of her right leg improved and she was referred back to her neurologist. Initially, the diagnosis of multiple sclerosis (MS) could not be confirmed on clinical grounds, but when the patient later presented with an optic neuritis, the diagnosis of MS was established and she was treated with steroids. Follow-up MR imaging showed a small remnant of the atypical MS-lesion (Figure 1C).

Figure 1. Summated SEF waveforms after right-sided PTN stimulation in patient 4 with corresponding magnetic field maps (A, anterior; P, posterior; L, left; R, right) for the first (41.6 ms) and second (54.4 ms) components in the affected hemisphere are depicted. The double peak after the stimulus onset is due to filtering the stimulus artefact. A â&#x20AC;&#x201C; C, magnetic source images (MSI) showing a partially cystic mass, with ring-like enhancement and posterior tibial nerve dipole localization in the lesion (A) and at the margin (B). Follow-up T2-MRI revealed a small remnant of the MS lesion (C).

28

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

Patient 11 Figure 2 shows summated SEF waveforms, corresponding magnetic field maps and magnetic source images (MSI) after left and right PTN stimulation in a 69-year-old patient with a left-sided frontal meningioma presenting with slight motor weakness of the right arm and leg. The MSI show the typical orientation of the first component, directed toward the mesial wall of the hemisphere and the second component with a more lateral direction. Enlarged dipole strengths were observed for the first and second component in the affected hemisphere.

Figure 2. Summated SEF waveforms after left and right PTN stimulation in Patient 11 with corresponding magnetic field maps (A, anterior; P, posterior; L, left; R, right) for the first (C1) and second (C2) component and MSIs. Note the latency differences and larger amplitude in the affected hemisphere.

29

2

CHAPTER 2

Patient 12 Figure 3 illustrates location asymmetry of homologous dipoles in a 47-year-old patient with seizures due to a right-sided parietal astrocytoma, WHO grade II. Mass effect of this postcentral tumor causes the localization of the first ECD to be displaced more anteriorly and laterally with an inter-dipole distance of 1.36 cm.

Figure 3. MSIs of Patient 12 after right (A) and left (B) posterior tibial nerve stimulation, showing location asymmetry of homologous dipoles. Mass effect of the postcentral tumor causes the localization of the first ECD to be displaced more anteriorly and laterally with an inter-dipole distance of 1.36 cm.

Figure 4. MSIs of Patient 15, showing the first two components in the affected (right) and unaffected (left) hemisphere.

Dipole orientation, latency and localization are shown in relation to the right-sided postcentral astrocytoma, WHO grade II. The first component on the right side suggests a safe distance between tumor and somatosensory cortex; however, the second component shows activation at the margin of the lesion. The inter-dipole distance between the first components is 1.19 cm and 2.33 cm between the second component.

30

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

Patient 15 Figure 4 shows that dipole localizations of the first and second component at different areas relative to the lesion have potential clinical value in a 34-year-old patient with a right-sided parietal astrocytoma, WHO grade II. The first SEF component in the affected hemisphere with a latency of 48.0 ms is located anteriorly compared to the second SEF component with a latency of 88.8 ms, which is located at the tumor-margin. Intraoperative electrocorticography confirmed the tumor localization adjacent to the postcentral gyrus.

DISCUSSION In the present study we compared interhemispheric differences in the activation and localization of the cortical somatosensory network to posterior tibial nerve stimulation in patients with unilateral perirolandic lesions. We found significantly increased dipole strength for the P45m component in the affected hemisphere of patients with unilateral intracranial lesions. Within the affected hemispheres, there was an additional significantly elevated dipole strength for the N60m component in the presence of motor deficits.

Patients with motor deficits also showed an increased P45m response in the contralesional hemisphere. No interhemispheric differences in the latencies of the SEF waveforms could be observed, only dipole location asymmetries.

SEF Characteristics MEG has been used mostly for localizing neuronal activity and the SEF characteristics and dipole strength after PTN stimulation have not received much attention in patients with intracranial pathology. In agreement with findings in normal controls, we found cortical activation patterns after PTN stimulation on the mesial wall of the contralateral hemisphere with the dipole directed horizontally to the contralateral hemisphere. Studies have shown that activation of the foot area shows rotating field patterns as a function of time, not only after tibial, but also after peroneal and sural nerve stimulations (13,17,20). Stimulation of the right PTN, for example, shows a counterclockwise rotation, which is explained by two approximately orthogonal dipoles with fixed locations and orientations but with varying relative strengths as a function of time. Although we did not study rotations of the PTN, the middle-latency component was oriented mostly in opposite directions compared to the first component.

Activity within or at the border of the lesion was found in three cases (18 %). Comparable findings were described by Schiffbauer et al. in a larger patient group, especially in the presence of low-grade gliomas (39). However, the clinical significance of these findings is subject to discussion. First, the source localization error of MEG in combination with possible conductivity heterogeneity of intracranial lesions can affect dipole localization.

31

2

CHAPTER 2

Secondly, the prediction of functional deficits after total resection when MEG sources are within the lesion remains to be established.

Location asymmetries of more than 1.0 cm were observed in more than 70% of the patients for each component. Although cortical displacement was not studied in detail, these location asymmetries are mostly due to the space-occupying effect of the lesions and their associated edema, which is in agreement with previous studies (5,34). Location asymmetries less than 1.0 cm were especially observed in the two cases with a small cavernoma in which there was virtually no cortical displacement. However, others have reported interhemispheric spatial differences of SEF components in the presence of cortical lesions without cortical displacement and hypothesized an altered cortical somatosensory network in the presence of lesions (3,9,14,34).

In studies on the topography of SEF following PTN stimulation in normal subjects, the main deflections (N37m, P45m, N60m and P75m) were identified in the hemisphere contralateral to the stimulated nerve, in the foot area of SI, mainly area 3b (13,17,20).

Our findings show latencies corresponding with P45m, N60m and P75m, but the N37m component was not identified. The first component in our group had a mean latency of approximately 48 ms which is in agreement with the findings of Mäkela et al. (29). Bilateral absence of the N37m response is unlikely to be related to the intracranial lesion but probably results from an insufficient dipolar pattern around 37 ms to calculate the ECD with a residual error smaller than 25 %. In addition, differences in the experimental settings with regard to number of sensors, pulse duration, stimulus frequency, interstimulus interval, number of trials and filter settings may also be influential.

We found certain variability in the number of SEF components, but this was equally distributed between the affected and unaffected hemisphere with a high intra-subject interhemispheric consistency. This finding has not been described for PTN activation, but has been shown earlier for median nerve activation in normal controls (45).

Interhemispheric Differences In agreement with Roberts et al. (37), who studied median nerve SEFs in neurosurgical patients, we observed an increased dipole strength in the affected hemisphere and no latency differences for the SEF components between the affected and unaffected hemisphere after PTN stimulation. The conclusion of Roberts et al. of increased neuromagnetic responses in patients with brain tumors was based on the absence of systematic displacement of dipole y-coordinates between the hemispheres, which could have affected the measured responses. However, the x- and z-coordinates were not taken into account. Since the estimated ECD strength value is greatly affected by the distance from the sphere center, it is possible that systematic displacement of the dipole location towards the surface could affect the ECD strength (47). Since there was no significant difference in the x- and z-coordinates in our group, this could not have influenced the observed ECD strength differences.

Some authors advocate the study of the dipole strength to study excitatory and inhibitory influences of cerebral lesions (9,33,37,47). Since the dipole strength value represents the

32

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

sum of excitatory and inhibitory postsynaptic potentials, neuronal lesions can alter the synchronisation process after somatosensory stimulation. In patients with stroke, hyperexcitability of the primary somatosensory cortex in the affected hemisphere has been found in the presence of cortical and cortico-subcortical lesions and is possibly related to a reduction of intracortical inhibition (6,33), in contrast to subcortical lesions, which show enhanced inhibition (25). Whether or not this applies to brain tumors is unclear. It has been suggested that increased dipole strength in patients with tumors close to the CS is a result of hyperactive and more synchronized neurons in surrounding tissue due to altered concentrations of inhibitory and excitatory neurotransmitters (37). Patients with motor deficits not only had increased responses in the AH, but also a substantially increased P45m response in the UH compared with the UH of patients without motor deficits. In patients with stroke motor cortex disinhibition (or hyperexcitability) has also been described in the unaffected hemisphere (25,42) and is presumed to be related to suppression of transcallosal inhibition. Others suggest increased contralesional responses as a result of recruitment reflecting cerebral plasticity (2). Our study has a relatively small and heterogeneous patient group, which makes it difficult to assess the underlying mechanism and relevance of the interhemispheric differences found. Future studies with respect to the differential effects of location and type of lesion on cerebral reorganization are needed to elucidate the mechanisms involved in cortical reorganization.

Lesion volume as a possible contributing factor to increased neuromagnetic responses was not demonstrated in our study, which is in accordance to others (33,44).

In healthy subjects, SEF characteristics after median nerve stimulation have a high intersubject variability in combination with an intra-subject interhemispheric consistency (45,46). The somatosensory cortex of the leg is known to be highly variable between subjects and therefore affects theoretically the inter-subject variability of the tibial nerve data (13) to a large extent. This study lacks the presence of a control group, but since intra-individual characteristics are more consistent it is conceivable that despite a high inter-subject variability of the PTN data (which reduces the chance of finding significant interhemispheric differences), the results have potential clinical value.

Clinical Relevance The dipole strength has been described as a valuable quantitative index of cortical response to somatosensory stimuli in patients with different neurological diseases (33,47).

In our study, patients with motor deficits showed larger dipole strengths in the affected hemisphere compared with patients without motor deficits. If SEF characteristics are related to clinical motor symptoms, MEG might be used as a potential quantitative measure of lesion involvement in the motor pathways. However, in patients with intracranial lesions like in other conditions such as epilepsy, multiple sclerosis and stroke, various abnormalities of SEFs and dipole strenght may exist (9,22,27,30), but the clinical relevance still has to be established.

The clinical usefulness of somatosensory evoked magnetic responses in preoperative surgical planning and intra-operative guidance using neuronavigation has been shown to

33

2

CHAPTER 2

be valuable and accurate in many reports (11,12,16,21,28,35,40) with respect to activation of the median nerve or digits. We found only find three reports, describing experiences with lower limb activation in functional neuronavigation (1,28,40). Alberstone et al. found that tibial and median nerve magnetic source imaging was an important tool in the preoperative assessment of patients with intracranial lesions and adjusted their surgical strategy in case of a close relationship of the tumor with the localization of the dipole (1).

Mäkela et al. (28) found that PTN-SEFs with electrical stimulation were useful for tumors in the vicinity of the central sulcus region in 10 out of 12 patients, not only for preoperative planning, but also for intra-operative orientation and facilitation of brain mapping. Schiffbauer et al. (40) found successful localization of the somatosensory representation of the toes in 82% of patients with supratentorial intra-axial brain lesions, by using compressed air-driven diaphragm clips attached to the toes in patients. It is unclear however if the unsuccessful cases were related to tumor localization close to the CS. The presence of intracranial lesions in our patient group did not affect successful dipole localization and the results can yield important information regarding the spatial relation and displacement of the somatosensory cortex of the lower limb in relation to the lesion. Furthermore, both latter studies only used ECDs in the 30-70 ms latency range (28,40). As shown by Figure 4, it is important to identify not only the first component but also middle-latency components that may show activation in SI, otherwise incorrect localising information might give the impression that the tumor is at a safe distance from SI. The presence of dipole location asymmetries observed with different SEF latencies, could have clinical significance in the evaluation and treatment of patients harbouring intracranial lesions. Therefore, for neurosurgical applications and surgical risk assessment it is recommended not only to evaluate the early-latency component but middle-latency components with activation in SI as well.

The use of MEG for cortical localization is a matter of debate as a result of its lower spatial resolution compared to functional MR imaging (fMRI) techniques (11,24,40). However, in the presence of intracranial lesions in or adjacent to the central sulcus, the fMRI results can show significantly lower activation in the affected hemisphere, reducing its reliability (10,15,18). Additional activation in multiple nonprimary areas may confound the results of fMRI as well (24). Both techniques have their qualities and limitations and probably information from both modalities will yield the best functional information to avoid neurological deficits (18,36).

The accuracy of PTN stimulation with respect to intraoperative evoked potential recordings was shown to be reliable in earlier studies (1,28,40). Our present results can be used for further studies elucidating the somatosensory network changes in the presence of intracranial lesions in combination with clinical findings.

34

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

CONCLUSIONS Posterior tibial nerve evoked magnetic responses in patients with unilateral supratentorial lesions around the central sulcus show significantly increased neuromagnetic responses in the affected hemisphere, especially in the presence of motor deficits. Dipole location asymmetries for homologous SEF components are frequently observed in the presence of intracranial lesions near the CS region. The results suggest spatial and functional changes in the somatosensory network as a result of intracranial lesions, with a possible relationship with clinical symptoms. Because fMRI results can be less reliable in the presence of perirolandic lesions, functional localization of the somatosensory cortex of the lower limb using PTN-SEFs could facilitate clinical decision making in patients with intracranial lesions.

35

2

CHAPTER 2

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

36

Alberstone CD, Skirboll SL, Benzel EC, et al. Magnetic source imaging and brain surgery: presurgical and intraoperative planning in 26 patients. Journal of neurosurgery 2000;92:79-90.

Altamura C, Torquati K, Zappasodi F, et al. fMRI-vs-MEG evaluation of post-stroke interhemispheric asymmetries in primary sensorimotor hand areas. Exp Neurol 2007;204:631-9.

Bartolomei F, Bosma I, Klein M, et al. Disturbed functional connectivity in brain tumour patients: evaluation by graph analysis of synchronization matrices. Clin Neurophysiol 2006;117:2039-49.

Bittar RG, Olivier A, Sadikot AF, Andermann F & Reutens DC. Cortical motor and somatosensory representation: effect of cerebral lesions. J Neurosurg 2000;92:242-8.

Buchner H, Adams L, Knepper A, et al. Preoperative localization of the central sulcus by dipole source analysis of early somatosensory evoked potentials and three-dimensional magnetic resonance imaging. J Neurosurg 1994;80:849-56.

Cicinelli P, Pasqualetti P, Zaccagnini M, Traversa R, Oliveri M & Rossini PM. Interhemispheric asymmetries of motor cortex excitability in the postacute stroke stage: a paired-pulse transcranial magnetic stimulation study. Stroke 2003;34:2653-8. de Munck JC, Verbunt JP, Van’t Ent D & Van Dijk BW. The use of an MEG device as 3D digitizer and motion monitoring system. Phys Med Biol 2001;46:2041-52.

Del Gratta C, Della Penna S, Ferretti A, et al. Topographic organization of the human primary and secondary somatosensory cortices: comparison of fMRI and MEG findings. Neuroimage 2002;17:1373-83.

Forss N, Hietanen M, Salonen O & Hari R. Modified activation of somatosensory cortical network in patients with righthemisphere stroke. Brain 1999;122 ( Pt 10):1889-99.

Fujiwara N, Sakatani K, Katayama Y, et al. Evoked-cerebral blood oxygenation changes in false-negative activations in BOLD contrast functional MRI of patients with brain tumors. Neuroimage 2004;21:1464-71.

Gallen CC, Schwartz BJ, Bucholz RD, et al. Presurgical localization of functional cortex using magnetic source imaging. J Neurosurg 1995;82:988-94.

Ganslandt O, Fahlbusch R, Nimsky C, et al. Functional neuronavigation with magnetoencephalography: outcome in 50 patients with lesions around the motor cortex. Neurosurg Focus 1999;6:e3.

Hari R, Nagamine T, Nishitani N, et al. Time-varying activation of different cytoarchitectonic areas of the human SI cortex after tibial nerve stimulation. Neuroimage 1996;4:111-8.

Hedström A, Malmgren K, Hagberg I, Jönsson L, Silfvenius H & Rydenhag B. Cortical reorganisation of sensory, motor and language functions due to early cortical damage. Epilepsy Res 1996;23:157-67.

Holodny AI, Schulder M, Liu WC, Wolko J, Maldjian JA & Kalnin AJ. The effect of brain tumors on BOLD functional MR imaging activation in the adjacent motor cortex: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 2000;21:1415-22.

Hund M, Rezai AR, Kronberg E, et al. Magnetoencephalographic mapping: basic of a new functional risk profile in the selection of patients with cortical brain lesions. Neurosurgery 1997;40:936-42; discussion 942-3.

Huttunen J, Kaukoranta E & Hari R. Cerebral magnetic responses to stimulation of tibial and sural nerves. J Neurol Sci 1987;79:43-54.

Inoue T, Shimizu H, Nakasato N, Kumabe T & Yoshimoto T. Accuracy and limitation of functional magnetic resonance imaging for identification of the central sulcus: comparison with magnetoencephalography in patients with brain tumors.

Neuroimage 1999;10:738-48.

Kakigi R & Shibasaki H. Scalp topography of the short latency somatosensory evoked potentials following posterior tibial nerve stimulation in man. Electroencephalogr Clin Neurophysiol 1983;56:430-7.

Kakigi R, Koyama S, Hoshiyama M, Shimojo M, Kitamura Y & Watanabe S. Topography of somatosensory evoked magnetic fields following posterior tibial nerve stimulation. Electroencephalogr Clin Neurophysiol 1995;95:127-34.

Kamada K, Takeuchi F, Kuriki S, Oshiro O, Houkin K & Abe H. Functional neurosurgical simulation with brain surface magnetic resonance images and magnetoencephalography. Neurosurgery 1993;33:269-72; discussion 272-3.

Karhu J, Hari R, Paetau R, Kajola M & Mervaala E. Cortical reactivity in progressive myoclonus epilepsy. Electroencephalogr Clin Neurophysiol 1994;90:93-102.

Kober H, Nimsky C, Möller M, Hastreiter P, Fahlbusch R & Ganslandt O. Correlation of sensorimotor activation with functional magnetic resonance imaging and magnetoencephalography in presurgical functional imaging: a spatial analysis.

Neuroimage 2001;14:1214-28.

Korvenoja A, Kirveskari E, Aronen HJ, et al. Sensorimotor cortex localization: comparison of magnetoencephalography, functional MR imaging, and intraoperative cortical mapping. Radiology 2006;241:213-22.

Liepert J, Restemeyer C, Kucinski T, Zittel S & Weiller C. Motor strokes: the lesion location determines motor excitability changes. Stroke 2005;36:2648-53.

PERIROLANDIC LESIONS AND TIBIAL MAGNETOENCEPHALOGRAPHIC RESPONSES

26. Lopes da Silva F. Functional localization of brain sources using EEG and/or MEG data: volume conductor and source models.

Magn Reson Imaging 2004;22:1533-8.

27. Mäkelä JP. Neurological application of MEG. Electroencephalogr Clin Neurophysiol Suppl 1996;47:343-55.

28. Mäkelä JP, Kirveskari E, Seppä M, et al. Three-dimensional integration of brain anatomy and function to facilitate intraoperative navigation around the sensorimotor strip. Hum Brain Mapp 2001;12:180-92.

29. Mäkelä JP, Illman M, Jousmäki V, et al. Dorsal penile nerve stimulation elicits left-hemisphere dominant activation in the second somatosensory cortex. Hum Brain Mapp 2003;18:90-9.

30. Mima T, Nagamine T, Ikeda A, Yazawa S, Kimura J & Shibasaki H. Pathogenesis of cortical myoclonus studied by magnetoencephalography. Ann Neurol 1998;43:598-607.

31. Naka D, Kakigi R, Koyama S, Xiang J & Suzuki H. Effects of tactile interference stimulation on somatosensory evoked magnetic fields following tibial nerve stimulation. Electroencephalogr Clin Neurophysiol 1998;109:168-77.

32. Oishi M, Fukuda M, Kameyama S, Kawaguchi T, Masuda H & Tanaka R. Magnetoencephalographic representation of the sensorimotor hand area in cases of intracerebral tumour. J Neurol Neurosurg Psychiatry 2003;74:1649-54.

33. Oliviero A, Tecchio F, Zappasodi F, et al. Brain sensorimotor hand area functionality in acute stroke: insights from magnetoencephalography. Neuroimage 2004;23:542-50.

34. Ossenblok P, Leijten FS, de Munck JC, Huiskamp GJ, Barkhof F & Boon P. Magnetic source imaging contributes to the presurgical identification of sensorimotor cortex in patients with frontal lobe epilepsy. Clin Neurophysiol 2003;114:221-32.

35. Roberts TP, Ferrari P, Perry D, Rowley HA & Berger MS. Presurgical mapping with magnetic source imaging: comparisons with intraoperative findings. Brain Tumor Pathol 2000;17:57-64.

36. Roberts TP & Rowley HA. Mapping of the sensorimotor cortex: functional MR and magnetic source imaging. AJNR Am J Neuroradiol 1997;18:871-80.

37. Roberts TP, Tran Q, Ferrari P & Berger MS. Increased somatosensory neuromagnetic fields ipsilateral to lesions in neurosurgical patients. Neuroreport 2002;13:699-702.

38. Rogers RL, Basile LF, Taylor S, Sutherling WW & Papanicolaou AC. Somatosensory evoked fields and potentials following tibial nerve stimulation. Neurology 1994;44:1283-6.

39. Schiffbauer H, Ferrari P, Rowley HA, Berger MS & Roberts TP. Functional activity within brain tumors: a magnetic source imaging study. Neurosurgery 2001;49:1313-20; discussion 1320-1.

40. Schiffbauer H, Berger MS, Ferrari P, Freudenstein D, Rowley HA & Roberts TP. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97:133342.

41. Seyal M, Emerson RG & Pedley TA. Spinal and early scalp-recorded components of the somatosensory evoked potential following stimulation of the posterior tibial nerve. Electroencephalogr Clin Neurophysiol 1983;55:320-30.

42. Shimizu T, Hosaki A, Hino T, et al. Motor cortical disinhibition in the unaffected hemisphere after unilateral cortical stroke.

Brain 2002;125:1896-907.

43. Sutherling WW, Crandall PH, Darcey TM, Becker DP, Levesque MF & Barth DS. The magnetic and electric fields agree with intracranial localizations of somatosensory cortex. Neurology 1988;38:1705-.

44. Tecchio F, Zappasodi F, Tombini M, et al. Brain plasticity in recovery from stroke: an MEG assessment. Neuroimage 2006;32:1326-34.

45. Tecchio F, Pasqualetti P, Pizzella V, Romani G & Rossini PM. Morphology of somatosensory evoked fields: inter-hemispheric similarity as a parameter for physiological and pathological neural connectivity. Neurosci Lett 2000;287:203-6.

46. Theuvenet PJ, van Dijk BW, Peters MJ, van Ree JM, Lopes da Silva FL & Chen AC. Whole-head MEG analysis of cortical spatial organization from unilateral stimulation of median nerve in both hands: no complete hemispheric homology.

Neuroimage 2005;28:314-25.

47. Tsutada T, Ikeda H, Tsuyuguchi N, et al. Detecting functional asymmetries through the dipole moment of magnetoencephalography. J Neurol Sci 2002;198:51-61.

48. Wood CC, Cohen D, Cuffin BN, Yarita M & Allison T. Electrical sources in human somatosensory cortex: identification by combined magnetic and potential recordings. Science 1985;227:1051-3.

37

2

CHAPTER 3

Topographical organization of mu and beta band activity associated with hand and foot movements in patients with perirolandic lesions Open Neuroimag J 2010; 4: 93 - 99

Ronald B. Willemse Jan C. de Munck Jeroen P.A. Verbunt Dennis van â&#x20AC;&#x2122;t Ent Peterjan Ris Johannes C. Baayen Cornelis J. Stam W. Peter Vandertop

CHAPTER 3

ABSTRACT To study the topographical organization of mu and beta band event-related desynchronization (ERD) associated with voluntary hand and foot movements, we used magnetoencephalographic (MEG) recordings from 19 patients with perirolandic lesions.

Synthetic aperture magnetometry (SAM) was used to detect and localize changes in the mu (7 â&#x20AC;&#x201C; 11 Hz) and beta (13 â&#x20AC;&#x201C; 30 Hz) frequency bands associated with repetitive movements of the hand and foot and overlaid on individual coregistered magnetic resonance (MR) images.

Hand movements showed homotopic and contralateral ERD at the sensorimotor (S/M) cortex in the majority of cases for mu and to a lesser extent for beta rhythms. Foot movements showed an increased heterotopic distribution with bilateral and ipsilateral ERD compared to hand movements. No systematic topographical segregation between mu and beta ERD could be observed.

In patients with perirolandic lesions, the mu and beta band spatial characteristics associated with hand movements retain the expected functional-anatomical boundaries to a large extent. Foot movements have altered patterns of mu and beta band ERD, which may give more insight into the differential functional role of oscillatory activity in different voluntary movements.

40

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

INTRODUCTION Source localization of motor evoked magnetic activity using magnetoencephalography (MEG) can be based on the detection of task-related modulations of cortical oscillatory activity (35-37). During the performance of voluntary movements, suppression or eventrelated desynchronization (ERD) of the mu (7-11 Hz) and beta (13-30 Hz) rhythm has been observed in the contralateral sensorimotor (S/M) areas of healthy subjects (31,32,37).

Spatial mapping of the mu-rhythm shows different sources in the primary somatosensory (SI) and primary motor (MI) cortex (37). The beta oscillations have been mainly observed in MI, but in SI as well (19,38). Mu and beta ERD is frequently found over the primary hand area during finger movement, however ERD over the primary foot area during toe movement is more difficult to find (31,32).

Localization of motor evoked magnetic activity was traditionally based on equivalent single dipole (ECD) modelling. Recently, a new spatial filtering technique, called synthetic aperture magnetometry (SAM) has been described as a new tool for neuromagnetic source localization (41,43). Statistical differences in the power of the selected frequency band can be evaluated between the active and control state on a per-voxel basis. The areas showing statistical differences can then be displayed on the individual coregistered magnetic resonance (MR) images (17).

Altered activation patterns of the motor cortex have been described in the presence of intracranial lesions, using functional magnetic resonance imaging (fMRI) (12,15,16,26) and positron emission tomography (PET) (4) and may be explained by partial neurovascular de-coupling and changes in metabolism. Functional localization of motor cortex using MEG measures neuronal activity directly with a high temporal resolution, but source localization with single ECD modelling was found to be clinically insufficient (25). Cortical sources involved in motor control can also be studied using SAM-analysis of oscillatory changes of the mu and beta frequency bands, but studies in patients with intracranial lesions are limited (13,27,41). MEG data of the spatial distribution of mu and beta ERD in combination with repetitive foot movements in patients with intracranial lesions have not been described earlier. In this study, we used SAM analysis to study movement-related oscillatory changes in the mu and beta band to identify systematic patterns of topographical organization in the presence of perirolandic lesions.

MATERIALS AND METHODOLOGY Patients From the patients with intracranial lesions, referred to the department of Neurosurgery of the VU University Medical Center (Amsterdam, The Netherlands), nineteen consecutive patients were selected with unilateral intracranial lesions around the central sulcus (CS) region eligible for treatment (nine female and ten male, mean age: 42.8 years; age range 29 - 70; Karnofsky Performance Scale score â&#x2030;Ľ 70). The study was approved by the Medical

41

3

CHAPTER 3

Ethics Committee of the VU University Medical Center and informed consent from the participants was obtained prior to inclusion.

MEG Recordings During voluntary movements, subjects were lying or sitting comfortably with eyes closed, inside the three-layer magnetically shielded room (Vacuum Schmeltze Gmbh, Germany).

MEG was recorded with a system of 151 third-order axial gradiometers (VSM MedTech Ltd., Canada), with a helmet shape detector array covering the whole head. Each trial consisted of 30 epochs of 10 seconds with movement (15 epochs) alternated by 10 seconds with no movement (15 epochs). The movement consisted of self-paced repetitive non-clenching opening and closing of the (right or left) hand at about 1 Hz. For foot movements patients were instructed to alternate flexion and extension at the (right or left) ankle at about 1 Hz. The epochs were indicated by presenting a short tone (movement) or short burst of noise (no movement). The sampling rate was 625 Hz and on-line low pass filtering of 100 Hz. Movements were monitored on video and head position was measured before and after the task.

Structural MR images were acquired with a 1.5 T MR scanner (Siemens Sonata, Erlangen, Germany) using T1-weighted magnetization prepared rapid acquisition gradient echo (MPRAGE) sequence. Coregistration of MEG and MR imaging was performed using fiducial markers on the nasion and left and right pre-auricular points. In MEG small coils energised by AC currents and in the MR imager vitamin E capsules were used as markers.

MR image registration was performed after MEG registration, with MR markers in the positions as the MEG coils, enabling matching of the datasets with an estimated precision of 2 mm (11).

SAM analysis Synthetic aperture magnetometry (SAM) is a spatial filtering technique for threedimensional source localization of cortical oscillations. Brain activity can be localized by creating differential images of source power changes over discrete time-intervals for both the active and resting-state, relative to their noise variance (34,42). Prior to SAM analysis the MEG data were filtered into two frequency bands: mu (7 to 11 Hz) and beta (13 to 30 Hz). The region of interest (ROI) was set to include the whole cerebral cortex with a 2.0 mm isotropic voxel resolution. The statistical evaluation of the ratio of the power differences between the active and resting state to the sum of the powers of noise was assessed for each voxel and expressed as pseudo-T statistics (18). The pseudo-T values were then displayed and overlaid on individual co-registered MR images. In accordance to previous studies (41), SAM pseudo-T images were thresholded for peak pseudo-T values of 2.5 and the highest peak-value was evaluated for its anatomical localization. The SAMimages were then analysed for laterality of activation patterns and the localization of the maximum peak.

42

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

RESULTS The clinical data of the patients are summarized in Table 1. Seven patients had an astrocytoma (six WHO grade II and one WHO grade III), five had an oligodendroglioma (three WHO grade II and two WHO grade III), one a mixed oligo-astrocytoma (WHO grade II), two a meningioma (WHO grade I), two a cavernoma, one a pilocytic astrocytoma (WHO grade I) and one had a hamartoma. Lesions were located on the left-side in nine cases, on the right in ten cases. Four patients had slight to moderate muscle weakness and fifteen patients had presented with seizures.

Table 1. Clinical data patient group.

Case No.

Age (years) / Gender

Location / Lateralization

Histology

Symptoms

1

57 / F

f-R

oligodendroglioma, II

paresis L-arm (4/5)

2

48 / M

p-R

astrocytoma, II

seizures

3

70 / F

f-R

meningioma

hemiparesis L (4+/5)

4

41 / F

fp-R

meningioma

hemiparesis L (4+/5)

5

36 / F

fp-R

astrocytoma, II

seizures

6

39 / M

p-L

astrocytoma, II

hemiparesis R (4/5)

7

35 / F

p-R

cavernoma

seizures

8

48 / F

p-L

oligodendroglioma, II

seizures

9

35 / M

p-L

astrocytoma, II

seizures

10

54 / M

f-L

cavernoma

seizures

11

35 / F

p-R

astrocytoma, II

seizures

12

34 / M

f-R

astrocytoma, II

seizures

13

37 / M

f-L

astrocytoma, III

seizures

14

36 / F

f-L

oligo-astrocytoma, II

seizures

15

57 / F

f-R

oligodendroglioma, II

seizures

16

42 / M

f-R

oligodendroglioma, II

seizures

17

50 / M

fp-L

oligodendroglioma, III

seizures

18

31 / M

f-L

pilocytic astrocytoma

seizures

19

29 / M

p-R

hamartoma

seizures

F, female; M, male; f, frontal; p, parietal; fp, frontoparietal; L, left; R, right.

Table 2 shows the topography of mu and beta ERD associated with laterality of hand and foot movements (left versus right) and laterality of tumor (affected versus unaffected hemisphere). All 19 patients had performed separately a motor task of the left and right hand and 13 patients had performed a motor task of the left and right foot. For all 19 patients, beta ERD associated with hand movements was found, resulting in 19 datasets

43

3

CHAPTER 3

for left and 19 datasets for right hand movements. For mu band activity associated with hand movements one dataset was excluded due to pseudo-T values less than 2.5, resulting in 14 datasets for left and 13 datasets for right hand movements. For voluntary foot movements, seven datasets were excluded with pseudo-T values less than 2.5, resulting in 13 left and 12 right foot datasets for beta band activity. The mu band results with pseudo-T values larger than 2.5 were obtained in 10 left and 10 right foot datasets.

Most patients showed contralateral desynchronization patterns (Figure 1) associated with left or right hand movements. Left hand, presumably non-dominant hemispheric, motor tasks result in contralateral mu band activity in 85.7 % of the trials, compared to 69.2 % of contralateral activity during right hand motor tasks. Ipsilateral mu band activity occurred during right-sided motor tasks for both hand (15.4 %) and foot (20 %) movements. Mu band activity associated with hand movements was located contralaterally in the majority of cases and not related to tumor laterality. The bilateral and ipsilateral mu activation patterns from (right) hand movements were only found in patients with left-sided hemisphere lesions. Contralateral mu band activity associated with foot movements was found in 50 % of the trials for left or right foot movements.

Table 2. Topography of mu and beta band event-related desynchronization (ERD) associated with laterality of hand and foot movements (left versus right) and laterality of tumor (affected versus unaffected hemisphere).

TOPOGRAPHY contralat. (%)

bilat. (%)

mu

12 (85.7)

2 (14.3)

beta

12 (63.2)

mu

9 (69.2)

beta

HAND

BAND

left right

AH UH

ipsilat. (%) 0

Total

(0)

14

6 (31.6)

1 (5.2)

19

2 (15.4)

2 (15.4)

13

12 (63.2)

5 (26.3)

2 (10.5)

19

mu

11 (78.6)

1 (7.1)

2 (14.3)

14

beta

12 (62.2)

5 (26.3)

2 (10.5)

19

mu

10 (76.9)

3 (23.1)

0

(0)

13

beta

12 (63.2)

6 (31.6)

1 (5.2)

19

FOOT

BAND

left

mu

5 (50)

4 (40)

1 (10)

10

beta

6 (46.1)

5 (38.5)

2 (15.4)

13

mu

5 (50)

3 (30)

2 (20)

10

beta

5 (41.7)

5 (41.7)

2 (16.6)

12

mu

6 (54.5)

3 (27.3)

2 (18.2)

11

beta

5 (38.5)

5 (38.5)

3 (23.0)

13

mu

4 (44.4)

4 (44.4)

1 (11.1)

9

beta

6 (50.0)

5 (41.7)

1 (8.3)

12

right

AH UH

AH, affected hemisphere; UH, unaffected hemisphere.

44

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

3

Figure 1. Differential source power (SAM pseudoT images) correspon­ding to beta and mu eventrelated desynchronization (ERD) during left (L) and right (R) hand movements in a 57-year old female (patient 1) with a right-sided premotor oligoden­ droglioma (WHO grade II) and slight motor weakness of the left arm.

Contralateral homotopic beta ERD is seen on the sensorimotor cortex and on the primary somatosensory cortex for mu ERD.

Movement-related beta band activity was found in the contralateral hemisphere in 63.2 % of the trials for both left and right hand movements and was not influenced by tumor laterality. The total of 14 ipsi- and bilateral beta band responses of the hand were equally distributed between left- and right sided lesions.

Bilateral beta band activity of the foot (Figure 2) was found in ten patients, six had leftsided and four had right-sided lesions. The four patients with ipsilateral beta responses of the foot were evenly distributed between left- and right-sided lesions. From the four patients with clinical motor deficits, only one had ipsilateral mu and beta band activity.

In general however, ipsilateral mu and beta band activity was more frequently associated with a motor task of the affected hemisphere (Figure 3).

The spatial somatotopy of the most significant power changes for mu and beta band activity associated with hand and foot movements are summarized in Table 3.

Overall, the most significant areas of mu and beta band activity occur in S/M (73 %).

Fifteen percent had a clear localization in MI and 5 % in SI. Five percent had a prefrontal localisation and 1 % was found in the inferior parietal lobule. Two datasets had an occipital mu localisation probably due to alpha interference. Mu and beta band activity during hand movements with a S/M localization were always located in the area of the hand knob (homotopical distribution). From the 34 datasets with a S/M localisation of mu and beta band activity associated with foot movements, eight (23.5 %) demonstrate a heterotopic distribution mostly corresponding to hand movements (Figure 3).

45

CHAPTER 3

Figure 2. SAM pseudo-T images correspon­ ding to mu and beta ERD associated with left and right sided foot movements in a 42-year old male (patient 16) with a right-sided oligoden­droglioma (WHO grade II) in close relation to the precentral gyrus. The ERD areas (blue) show a bilateral and heterotopic distribution for mu and beta rhythms after leftsided (affected) foot movements corresponding to the area of hand representation.

Right-sided foot movements show a more diffuse ERD associated with hand and foot representation.

Figure 3. SAM pseudo-T images corresponding to mu and beta ERD associated with hand and foot movements from the left and right side in a 48-year old male (patient 2) with a diffuse large right-parietal astrocytoma (WHO grade II) presenting with seizures. Ipsilateral beta band activity is seen during left-sided (affected) hand and foot movements.

Note the heterotopic distribution of foot movements for both mu and beta rhythms.

46

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

Table 3. Areas with most significant power changes for mu and beta band during hand and foot movements contralateral to the affected and unaffected hemisphere.

HAND AH

FOOT UH

AH

UH

Location

mu

beta

mu

beta

mu

beta

mu

beta

Total (%)

S/M

10

13

10

13

9

10

8

7

80 (73)

MI

3

4

1

5

2

1

16 (15)

SI

1

2

1

1

1

6 (5)

Prefrontal

1

3

5 (5)

1

IPL

1

Occipital

1

Trials (N)

14

1 (1)

1 19

13

19

11

3

2 (2) 13

9

12

110 (100)

S/M, sensorimotor cortex; MI, primary motor cortex; SI, primary sensory cortex; IPL, inferior parietal lobule; AH, affected hemisphere; UH, unaffected hemisphere.

DISCUSSION Noninvasive imaging of motor cortical localization in patients with intracranial lesions, has been studied with different modalities, including positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). These modalities do not measure neural activity directly but mainly depict hemodynamic changes with poor temporal resolution and lack the possibility to differentiate between motor and somatosensory activation. In addition, functional MR imaging of the motor cortex can be influenced by adjacent tumor due to vasculature-induced signal changes (15,16,26).

SAM analysis of motor evoked magnetic activity recorded with MEG is a relatively new spatial filtering technique to identify cortical regions responsible for changes in spectral power associated with voluntary movements. This technique has been used for noninvasive localization of the motor cortex in neurosurgical patients and demonstrated to be an accurate method for functional localization compared to intraoperative mapping (13,27). The presence or location of the tumor did not affect the localization accuracy.

Although we did not perform intraoperative comparisons, the spatial filtering technique seems a reliable method to assess the topographical organization of cortical oscillatory activity in our patient group.

Cortical rhythms involved in motor control are the mu and beta oscillations, which have been studied previously with electroencephalography (EEG), electrocorticography (ECoG) and MEG in healthy subjects (2,8,10,28,31,37,42). Comparable studies in patients are limited and mostly relate to patients with epilepsy (2,8,10) and intracranial lesions (13,27,41). It is well known that the execution of a hand movement or stimulation of the hand can block or desynchronize the mu rhythm with a clear contralateral dominance in healthy subjects (7). Analogous to the mu rhythm, beta band activity can be modulated during voluntary movements as well and is supposed to have a precentral somatotopical

47

CHAPTER 3

origin (37,38). The results for foot movements however are much less clear (28,31).

In the present study, mu and beta band spatial characteristics associated with hand movements retain the expected functional-anatomical boundaries to a large extent, where foot movements have altered patterns of mu and beta ERD. With respect to hand movements, several studies from different modalities demonstrate movement-related asymmetry, with a predominant contralateral activation during right hand movements and bilateral activation during left hand movements [29-32] (3,21,24,33). Our data do not demonstrate this hemispheric asymmetry for left or right hand movements, which is in accordance with other SAM studies in patients with intracranial lesions (13,27).

Hemispheric asymmetry is likely to occur in the preparatory phase of a movement and is not comparable to our continuous motor task, which will contain a mixture of motor and somatosensory reafference, which could explain our findings.

Bilateral and ipsilateral responses in this study were observed for both frequency bands, however more frequently with foot movements. Bilateral activation of motor areas during unilateral voluntary movements has been described earlier with different methods in both normal subjects, stroke patients and patients with intracranial lesions (6,9,2123,27,39). Participation of the ipsilateral sensorimotor cortex in unilateral limb movements could be explained by the partially uncrossed fibers of the corticospinal tract in combination with interhemispheric interactions (22,44). The role of concomitant ipsilateral activation seems to be more frequent during a complex movement (1,33) but can also occur as a mirror response through crossed corticospinal pathways (5). Since we do not consider foot movements more complex than hand movements probably the mirror response could explain the more frequent bilateral activity. Bilateral activity during hand movements occurred more frequently in the beta band compared to mu band activity.

Whether this is due to a larger coupling of beta-band activity between bilateral primary motor cortices or interindividual variability, remains to be clarified.

Ipsilateral motor cortex involvement during unilateral voluntary movements has previously been reported with MEG studies (8,23,42) and was also demonstrated using SAM analysis in healthy subjects (20) and patients with intracranial lesions (27,41). Exclusive ipsilateral activation in our study was only found in a small proportion of datasets, which is in agreement with Nagarajan et al. (27). Only one patient with ipsilateral (mu and beta band) activation had slight motor weakness, the other trials with ipsilateral activity were found in patients with seizures and without neurological deficits. Although ipsilateral activation was more frequently found in the affected hemisphere, our data only partially support the hypothesis of Taniguchi et al. who suggested that (sub)clinical impairment of motor areas will result in recruitment of other, especially ipsilateral S/M cortex, which could explain their ipsilateral (beta band) activation patterns in the patient group (41). Since MEG analysis and thresholding are largely comparable, probably differences in tumor characteristics could explain the differences found. Since our patient group only consisted of one high-grade glioma (WHO grade III) it is possible that ipsilateral recruitment is related to tumor invasion of motor areas and therefore will occur more often in high-grade glioma, as studied by Taniguchi et al.

48

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

The topographical distribution of mu and beta ERD with respect to the central sulcus has been suggested to originate from the post- en precentral cortex respectively (32,38). In this study, we found no clear precentral origin of the beta band activity, which is supported by other EEG studies (10,40). These results support sensorimotor cortex involvement in the generators of both mu and beta oscillations. However, the lack of topographical segregation between mu and beta may be attributed to methodological differences.

In agreement with other neuromagnetic studies involving beta oscillations, we also found a somatotopical distribution to different movements (37,38). Hand movements had a predominant localization near the hand area representation and foot movements frequently had a central paramedian localisation. Most of the peak-value activations for both mu and beta patterns were found at the S/M cortex or central sulcus region which might support the hypothesis that motor control is strongly linked to sensory information (10). An alternative explanation could be poor functional specificity due to the selected beta frequency band as 13 to 30 Hz, mixing mu and beta band activity together. Hand movements in our study frequently had a distribution, which was in accordance with the anatomically expected hand representation. The cortical representation of foot movements however, was in 23.5 % of the trials for both mu and beta band activity, located at the representation area of the hand. This heterotopic distribution of oscillatory changes in the hand area during non-hand movements has been described earlier in EEG, MEG and fMRI of studies of healthy subjects (8,14,28-30,39). Recently, Stippich et al. described their experience with heterotopic ipsilateral coactivation on fMRI after unilateral voluntary movements in brain tumor patients, but found no relation with tumor characteristics (39). In combination with our results, heterotopic activation patterns are probably not due to the presence of intracranial lesions and must be explained by a more widely distributed cortical motor network than the classic somatotopical distribution of the different body parts. Future studies involving quantification of cortical connectivity using graph-analysis of the MEG data during the active and control state, will hopefully give more insight in the complex network involved in cerebral motor control.

CONCLUSION Spatially filtered MEG can be used to detect and localize changes in cortical activity associated with hand and foot movements in patients with brain lesions. Hand and foot movements predominantly show contralateral activity in the mu and beta band with a frequent localization at the S/M cortex. Bilateral and ipsilateral responses were observed for both frequency bands, however more frequently with foot movements. Mu and beta band spatial characteristics in the presence of perirolandic lesions support previous findings of mu and beta band activity associated with hand movements to a large extent.

No distinct topographical segregation between mu and beta band activity could be observed. The results may provide further insight in the generators of oscillatory activity involved in voluntary movements.

49

3

CHAPTER 3

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

50

Allison JD, Meador KJ, Loring DW, Figueroa RE & Wright JC. Functional MRI cerebral activation and deactivation during finger movement. Neurology 2000;54:135-42.

Arroyo S, Lesser RP, Gordon B, Uematsu S, Jackson D & Webber R. Functional significance of the mu rhythm of human cortex: an electrophysiologic study with subdural electrodes. Electroencephalogr Clin Neurophysiol 1993;87:76-87.

Bai O, Mari Z, Vorbach S & Hallett M. Asymmetric spatiotemporal patterns of event-related desynchronization preceding voluntary sequential finger movements: a high-resolution EEG study. Clin Neurophysiol 2005;116:1213-21.

Bittar RG, Olivier A, Sadikot AF, Andermann F & Reutens DC. Cortical motor and somatosensory representation: effect of cerebral lesions. J Neurosurg 2000;92:242-8.

Borroni P, Montagna M, Cerri G & Baldissera F. Bilateral motor resonance evoked by observation of a one-hand movement: role of the primary motor cortex. Eur J Neurosci 2008;28:1427-35.

Cao Y, Dâ&#x20AC;&#x2122;Olhaberriague L, Vikingstad EM, Levine SR & Welch KM. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112-22.

CHATRIAN GE, PETERSEN MC & LAZARTE JA. The blocking of the rolandic wicket rhythm and some central changes related to movement. Electroencephalogr Clin Neurophysiol 1959;11:497-510.

Cheyne D, Bakhtazad L & Gaetz W. Spatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming approach. Hum Brain Mapp 2006;27:213-29.

Cramer SC, Finklestein SP, Schaechter JD, Bush G & Rosen BR. Activation of distinct motor cortex regions during ipsilateral and contralateral finger movements. J Neurophysiol 1999;81:383-7.

Crone NE, Miglioretti DL, Gordon B, et al. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. I. Alpha and beta event-related desynchronization. Brain 1998;121:2271-99. de Munck JC, Verbunt JP, Vanâ&#x20AC;&#x2122;t Ent D & Van Dijk BW. The use of an MEG device as 3D digitizer and motion monitoring system. Phys Med Biol 2001;46:2041-52.

Fandino J, Kollias SS, Wieser HG, Valavanis A & Yonekawa Y. Intraoperative validation of functional magnetic resonance imaging and cortical reorganization patterns in patients with brain tumors involving the primary motor cortex. J Neurosurg 1999;91:238-50.

Gaetz W, Cheyne D, Rutka JT, et al. Presurgical localization of primary motor cortex in pediatric patients with brain lesions by the use of spatially filtered magnetoencephalography. Neurosurgery 2009;64:177-85; discussion 186.

Hanakawa T, Parikh S, Bruno MK & Hallett M. Finger and face representations in the ipsilateral precentral motor areas in humans. J Neurophysiol 2005;93:2950-8.

Holodny AI, Schulder M, Liu WC, Maldjian JA & Kalnin AJ. Decreased BOLD functional MR activation of the motor and sensory cortices adjacent to a glioblastoma multiforme: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 1999;20:609-12.

Holodny AI, Schulder M, Liu WC, Wolko J, Maldjian JA & Kalnin AJ. The effect of brain tumors on BOLD functional MR imaging activation in the adjacent motor cortex: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 2000;21:1415-22.

Ishii R, Shinosaki K, Ukai S, et al. Medial prefrontal cortex generates frontal midline theta rhythm. Neuroreport 1999;10:675-9.

Ishii R, Canuet L, Herdman A, et al. Cortical oscillatory power changes during auditory oddball task revealed by spatially filtered magnetoencephalography. Clin Neurophysiol 2009;120:497-504.

Jensen O, Goel P, Kopell N, Pohja M, Hari R & Ermentrout B. On the human sensorimotor-cortex beta rhythm: sources and modeling. Neuroimage 2005;26:347-55.

Jurkiewicz MT, Gaetz WC, Bostan AC & Cheyne D. Post-movement beta rebound is generated in motor cortex: evidence from neuromagnetic recordings. Neuroimage 2006;32:1281-9.

Kim SG, Ashe J, Georgopoulos AP, et al. Functional imaging of human motor cortex at high magnetic field. J Neurophysiol 1993;69:297-302.

Kobayashi M, Hutchinson S, Schlaug G & Pascual-Leone A. Ipsilateral motor cortex activation on functional magnetic resonance imaging during unilateral hand movements is related to interhemispheric interactions. Neuroimage 2003;20:2259-70.

Kristeva R, Cheyne D & Deecke L. Neuromagnetic fields accompanying unilateral and bilateral voluntary movements: topography and analysis of cortical sources. Electroencephalogr Clin Neurophysiol 1991;81:284-98.

Leocani L, Toro C, Manganotti P, Zhuang P & Hallett M. Event-related coherence and event-related desynchronization/ synchronization in the 10 Hz and 20 Hz EEG during self-paced movements. Electroencephalogr Clin Neurophysiol 1997;104:199-206.

MOVEMENT-RELATED OSCILLATORY ACTIVITY IN PATIENTS WITH INTRACRANIAL LESIONS

25. Lin PT, Berger MS & Nagarajan SS. Motor field sensitivity for preoperative localization of motor cortex. J Neurosurg 2006;105:588-94.

26. Liu WC, Feldman SC, Schulder M, et al. The effect of tumour type and distance on activation in the motor cortex.

Neuroradiology 2005;47:813-9.

27. Nagarajan S, Kirsch H, Lin P, Findlay A, Honma S & Berger MS. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109:228-37.

28. Neuper C & Pfurtscheller G. Event-related dynamics of cortical rhythms: frequency-specific features and functional correlates. Int J Psychophysiol 2001;43:41-58.

29. Pfurtscheller G, Brunner C, Schlögl A & Lopes da Silva FH. Mu rhythm (de)synchronization and EEG single-trial classification of different motor imagery tasks. Neuroimage 2006;31:153-9.

30. Pfurtscheller G & Neuper C. Event-related synchronization of mu rhythm in the EEG over the cortical hand area in man.

Neurosci Lett 1994;174:93-6.

31. Pfurtscheller G, Neuper C, Andrew C & Edlinger G. Foot and hand area mu rhythms. Int J Psychophysiol 1997;26:121-35.

32. Pfurtscheller G & Lopes da Silva FH. Event-related EEG/MEG synchronization and desynchronization: basic principles.

Clinical neurophysiology 1999;110:1842-57.

33. Rao SM, Binder JR, Bandettini PA, et al. Functional magnetic resonance imaging of complex human movements. Neurology 1993;43:2311-8.

34. Robinson SE. Localization of event-related activity by SAM(erf). Neurol Clin Neurophysiol 2004;2004:109.

35. Salenius S, Portin K, Kajola M, Salmelin R & Hari R. Cortical control of human motoneuron firing during isometric contraction.

J Neurophysiol 1997;77:3401-5.

36. Salenius S, Schnitzler A, Salmelin R, Jousmäki V & Hari R. Modulation of human cortical rolandic rhythms during natural sensorimotor tasks. Neuroimage 1997;5:221-8.

37. Salmelin R, Hämäläinen M, Kajola M & Hari R. Functional segregation of movement-related rhythmic activity in the human brain. Neuroimage 1995;2:237-43.

38. Salmelin R & Hari R. Spatiotemporal characteristics of sensorimotor neuromagnetic rhythms related to thumb movement.

Neuroscience 1994;60:537-50.

39. Stippich C, Blatow M, Durst A, Dreyhaupt J & Sartor K. Global activation of primary motor cortex during voluntary movements in man. Neuroimage 2007;34:1227-37.

40. Szurhaj W, Derambure P, Labyt E, et al. Basic mechanisms of central rhythms reactivity to preparation and execution of a voluntary movement: a stereoelectroencephalographic study. Clinical neurophysiology 2003;114:107-19.

41. Taniguchi M, Kato A, Ninomiya H, et al. Cerebral motor control in patients with gliomas around the central sulcus studied with spatially filtered magnetoencephalography. J Neurol Neurosurg Psychiatry 2004;75:466-71.

42. Taniguchi M, Kato A, Fujita N, et al. Movement-related desynchronization of the cerebral cortex studied with spatially filtered magnetoencephalography. Neuroimage 2000;12:298-306.

43. Vrba J & Robinson SE. Signal processing in magnetoencephalography. Methods 2001;25:249-71.

44. Ziemann U, Ishii K, Borgheresi A, et al. Dissociation of the pathways mediating ipsilateral and contralateral motor-evoked potentials in human hand and arm muscles. J Physiol 1999;518 ( Pt 3):895-906.

51

3

CHAPTER 4

Slowing of M1 oscillations in brain tumor patients in resting state and during movement Clin Neurophysiol 2012; 123(11): 2212-9

Bernadette C.M. van Wijk Ronald B. Willemse W. Peter Vandertop A. Daffertshofer

CHAPTER 4

ABSTRACT Objective Brain tumors may severely disrupt the structure and function of the brain. While abnormal low-frequency activity can be found around tumor borders, disrupted structural connectivity may also impinge on neural activity in distant brain regions and other frequency bands. We investigated how glioma in patients with normal motor functioning affects activity in primary motor areas (M1).

Methods Using magnetoencephalography in 12 patients with unilateral glioma located around the central sulcus, we studied activity in bilateral M1s in resting state and during movement with focus on motor-related mu (8-12 Hz) and beta rhythms (15-30 Hz). Principal component analysis served to test for differences in spectral content.

Results A shift was found towards lower frequencies for M1 in the tumor hemisphere compared to M1 in the healthy hemisphere, caused by an increase in mu and decrease in beta power.

This pattern was observed both in resting state and during movement.

Conclusions This ‘slowing’ of brain oscillations in M1 resembles findings in patients with monohemispheric stroke and Parkinson’s disease. A loss of intra-cortical connectivity may account for these findings, possibly supplemented by tumor-induced changes in neurotransmitter systems.

Significance Motor functioning may be unaffected by a spectral shift of mu and beta oscillations.

54

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

INTRODUCTION Gliomas, usually classified as low-grade (slow-growing) or high-grade (fast-growing), not only exhibit local tumor growth, but also widespread, even contralateral, invasion of white matter pathways (25). Characteristic EEG findings are prominent delta (and/or theta) band oscillations in the tumorâ&#x20AC;&#x2122;s vicinity, first described by Walter (1936) (55), that stem from sources close to the tumor border (3,16,22,36,44). It is currently believed that not the tumor bulk itself but structural damage to surrounding tissue underlies the neural slowing as the delta oscillations do not disappear after surgical removal of the tumor (17). Indeed, white matter lesions are known to induce delta activity in cat cerebral cortex (26).

Besides cortical deafferentation of surrounding gray matter, disruptions of white matter fibers by the tumor may impinge on large, distributed networks, affecting distant brain areas. Widespread changes in spectral power, including increased theta power, have been found in addition to a decrease in gamma power over frontal and central regions (11).

Communication between areas, manifested by neural synchronization, is also shown to be disturbed in the presence of a tumor (7,18) and changes in global functional network organization, as observed in tumor patients, even seems to correlate with diminished cognitive performance (12).

Here, we put particular attention to the motor network with its elaborate connections between (pre-)frontal, parietal, cerebellar, and subcortical areas. Following a monohemispheric stroke, an increase in power for frequencies up to the mu band and decreases in power for higher frequencies has been shown in Rolandic regions (49,50).

Likewise, we would expect to find a similar slowing of oscillations within the motor network if tumors induce a loss of structural connectivity. For that reason, we studied M1-activity in the tumor hemisphere in comparison to that of the contralateral, healthy hemisphere. Moreover, we tested whether a possible slowing in activity was task-related by looking at both resting state and movement.

METHODS Patients From the patients with intracranial lesions referred to the department of Neurosurgery of the VU University Medical Center (Amsterdam, The Netherlands), who had preoperative MEG and surgical treatment, we selected 12 consecutive patients (six women and six men, mean age 41.8 years; age range 29Â -Â 57) with unilateral intracranial gliomas around the central sulcus to enter the study. Patient characteristics are summarized in Table 1.

The neuropathological diagnosis was determined according to the WHO Classification of Tumors affecting the central nervous system (37). Nine patients had a low-grade glioma (one grade I and eight grade II) and three a high-grade glioma (two grade III and one grade IV). All patients were self-identified as right-handed. Neurological examination showed no (obvious) deficits. All patients had tumor-related epilepsy and used one or more

55

4

CHAPTER 4

antiepileptic drugs (AEDs) to control their seizures. Tumor-related therapy such as steroids or chemotherapy had not been applied. The study was approved by the Medical Ethics Committee of the VU University Medical Center and informed consent was obtained prior to inclusion.

Table 1. Patient characteristics.

Case

Gender

Age

Localization

Lateralization

Histology

1

F

35

Parietal

R

Astrocytoma, II

2

F

47

Parietal

L

Astrocytoma, II

3

F

40

Frontal

R

Oligo-astrocytoma, II

4

M

48

Frontal

R

Astrocytoma, II

5

F

36

Frontal

L

Oligo-astrocytoma, II

6

M

42

Parietal

R

Oligodendroglioma, II

7

M

50

Frontal

L

Oligodendroglioma, III

8

F

57

Frontal

R

Oligodendroglioma, II

9

F

50

Frontal

L

Oligodendroglioma, II

10

M

31

Parietal

L

Pilocytic astrocytoma, I

11

M

29

Frontal

R

Glioblastoma multiforme, IV

12

M

37

Frontal

L

Astrocytoma, III

Data acquisition Subjects performed voluntary hand movements consisting of self-paced, repetitive, nonclenching opening and closing of the right or left hand at a frequency of about 1Hz. These were performed for the right and left hand separately in 15 epochs of 10s alternated by 10s of rest. Epochs were initiated by a brief single tone (movement) or a noisy burst (rest).

During recordings subjects were either lying or sitting comfortably with their eyes closed.

Brain activity was recorded using a 151-channel whole-head MEG system (CTF Systems Inc., Vancouver, Canada) with 3rd-order synthetic gradiometers. Signals were low-pass filtered at 99 Hz and sampled with a frequency of 625 Hz. Head position was monitored before and after each block using coils on the nasion and pre-auricular points. Following MEG registration, structural MR images were acquired using a 1.5T MR scanner (Siemens Sonata, Erlangen, Germany; T1-weighted MPRAGE sequence) with MR markers at the position of the MEG coils enabling co-registration of both datasets.

Source localization Previous MEG studies have used spatial filtering techniques to identify the source location of primary motor areas (48,57). The reliability of these methods has been verified by comparing the source locations with intra-operative electrical stimulation (24,43). We employed beamformers (synthetic aperture magnetometry, SAM) to identify brain regions that were significantly activated during the task. Beamformers yield spatial

56

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

filters that allow for pinpointing spectral power of a pre-selected frequency band in a certain time window (14,28,54). We chose for the beta band (15-30 Hz) because of its movement-related power decrease in M1. To assure proper task performance and to avoid possible after-effects during rest, we selected for each epoch a 3-9 s time interval and estimated beta power for both movement and resting state. For the beamforming we further defined a forward model based on the individual structural MRIs using a local spheres approximation. To identify active sources, the movement condition was contrasted with the resting state, yielding for each voxel (at 2 x 2 x 2mm resolution) a pseudo-t value indicating whether spectral power between movement and resting state differed significantly. The peak pseudo-t values close to the hand knob were selected both contra- and ipsilateral to the moving hand. Source activity was estimated by extracting the time series corresponding to these locations using the beamformer weights (resulting in a â&#x20AC;&#x2DC;virtual sensorâ&#x20AC;&#x2122;). For this, the beamformer weights were recomputed using a covariance matrix based on a 5-45 Hz frequency band, which is the frequency range that we considered for further analyses. This re-computation was necessary to avoid leakage of power outside the beta band originating from other sources into the M1 time series. Separate SAM analyses were performed for left and right hand movements.

Spectral analysis After determining the signals in left/right M1s, their power spectral densities were computed per epoch (3-9 s time interval) using Welchâ&#x20AC;&#x2122;s method with 0.5 s Hamming tapers with 0.25 s overlap. Power spectra were averaged over all 15 epochs per condition. We focused on the frequency range between 5 and 45 Hz as it was expected to contain the effects of interest without being contaminated by the movement frequency. Power spectra were first log-transformed before averaging over trials and/or subjects. In addition, relative spectra were calculated by normalizing the total power for each spectrum to one. All spectra were categorized according to tumor (tumor versus healthy hemisphere), active (movement versus rest) and movement side (contra- versus ipsilateral hand), yielding eight spectra per subject.

Statistics Power spectra were assessed in three ways. First, median frequencies were calculated in order to test for a possible shift in overall frequency content. Differences between conditions were assessed using a 2 x 2 x 2 repeated measures ANOVA with factors tumor (tumor vs. healthy), movement side (contra- vs. ipsilateral) and active (movement vs. rest).

Second, more detailed changes in the spectral distributions were investigated via principal component analysis (PCA). For this we used the relative power spectra to reduce interand intra-subject variability. We combined all spectra into a single matrix (number of subjects * number of conditions x number of frequencies). Prior to PCA, per frequency the grand averaged power over subjects and conditions was calculated and subtracted from the individual values. This guaranteed that the modes expressed deviations from a

57

4

CHAPTER 4

common, grand averaged power spectrum, hence making differences between conditions more apparent (see also Boonstra et al., 2007) (8). Note that this is equivalent to a meancentering partial least squares (McIntosh and Lobaugh 2004) (41) with the only difference that power spectra are not merely averaged over subjects within conditions but individual weight factors are being obtained from the corresponding eigenvectors prior to averaging.

We tested for differences between conditions in the weight factors of the first four principal components using again a 2 x 2 x 2 design for an ANOVA with repeated measures (see above).

Finally, we tested directly for tumor effects on movement-related activity. The power spectrum of each subject during movement was divided by the power spectrum during rest. The resulting spectrum was divided into 41 bins of 1 Hz centered around the main frequencies from 5 to 45 Hz. For each bin, we compared the average power between the healthy and tumor hemisphere using paired-samples t-tests.

RESULTS The spectra for the tumor hemisphere of one subject (#4) contained considerably more power at low frequencies compared to all other subjects and conditions, due to a shift in peak mu frequency from 9 to 6 Hz. In comparison with the other subjects we considered this an outlier and excluded the data for this subject from further analyses. Hence all the reported results are based on 11 subjects.

Beamformers To illustrate the beamformer outcome, Figure 1 shows the SAM sources found in three patients. Despite the presence of a tumor, the movement-related beta activity was well pronounced and for all subjects we could identify contra- and ipsilateral sources in the vicinity of the hand knob. The M1 peak pseudo-t values ranged from 2.6 to 29.3. In one of the subjects (#8) the ipsilateral source showed an increase in beta power. In all other cases the expected movement-related decrease in power was found. The peak locations for sources reconstructed for movement of the left hand did not differ more than 6 mm in the x, y or z-location compared to those found for movement of the right hand.

Power spectra Figure 2 shows the grand-averaged (relative) power for the eight conditions. From these spectra, it can be seen that power is higher in the tumor hemisphere compared to the healthy hemisphere for frequencies in the mu band and lower for frequencies in the beta band. This was the case in resting state and during movement for both power and relative power. This pattern points at a shift in frequency content for the tumor hemisphere towards lower frequencies when idling but also when the areas become functionally activated. These observations were subsequently tested statistically and results are described in the next sessions.

58

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

4

Figure 1. Beamformer results for three subjects during right hand movement. The peak locations (having the most significant movement-related beta power decrease) in left and right motor areas are shown. Corresponding pseudo-t values are indicated in yellow. The green arrows point at the location of the tumor. A) subject #1; B) subject #11; C) subject #12.

Median frequencies Median frequencies are summarized in Figure 3 and Table 2. We note that a low median frequency indicates more relative power for low frequencies compared to a high median frequency. This was indeed the case for the tumor hemisphere compared to the healthy hemisphere as was indicated by a significant main effect of tumor (F(1,10) = 10.723, p = .008). No further significant differences were found. The fact that the main effect of action did not reach significance (F(1,10) = 0.925, p = .359) indicated that relative power contributions for different frequencies did not change during movement. Since the movement-related decrease in power appears proportional to the amount of power in resting state (see Figure 5), the absence of effects on median frequency is not unexpected.

Principal component analysis The first four modes collectively explained 91% of the total variance in the data and are shown in Figure 4, where for visualization purposes weight factors are averaged over

59

CHAPTER 4

subjects to obtain a single weight per condition. These weight factors multiplied with the projections indicate for each condition the contribution of that mode to the power spectrum. Subsequent modes each explained less than 5% of the data variance and were not further considered, as they are likely to merely reflect inter-individual differences.

The first mode separated the tumor and healthy hemisphere indicated by positive and negative weight factors, respectively. This found support by a significant main effect of tumor (F(1,10) = 10.485, p = .009). The projection showed a positive deviation from the grand-average for frequencies below and a negative deviation for frequencies above 17 Hz. The positive weight factors for the tumor hemisphere imply that low frequencies contained more power and high frequencies contained less power than average (inversely for the healthy hemisphere). No other significant effects were found for this mode.

Figure 2. Grand-averaged (relative) power spectra. Power for low frequencies for the tumor (blue lines) is increased compared to the healthy hemisphere (red lines) and is decreased for high frequencies. An equivalent pattern can be observed regardless of the task and moving hand (contra- or ipsilateral side) and is found for both the power and the relative power.

60

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

Table 2. Average median frequencies. Corrected standard deviations are indicated between brackets (cf. Figure 3).

Rest

Movement

Healthy

Tumor

Healthy

Tumor

Contralateral

16.50 (1.20)

15.17 (1.55)

16.74 (2.22)

15.74 (1.28)

Ipsilateral

16.78 (0.79)

15.07 (1.81)

17.29 (1.19)

16.02 (1.11)

4

Figure 3. Average median frequencies. The bars are grouped by the factors active (rest or active) and movement side (contra- or ipsilateral to the moving hand) to make the significant main effect of tumor more apparent. Error bars indicate standard errors and have been corrected for a within-subject design by setting between-subject variance to zero (38).

The second mode reflected changes in the power spectra due to movement. A significant main effect of active was found (F(1,10) = 10.060, p = .009) as well as a significant interaction for movement side x active (F(1,10) = 31.476, p< .001). The difference between movement and rest was caused by a broad power decrease from about 9 to 22 Hz, as well as an increase at low frequencies that is only present in the relative power spectra and is likely due to normalization. The interaction effect revealed that the difference between the hemispheres contra- and ipsilateral to the moving hand in size of this modulation was larger during movement than in resting state. In addition, we found a trend for the tumor x movement side interaction (F(1,10) = 4.507, p = .060), which indicates that trials for the hand contralateral to the healthy hemisphere were accompanied by a larger beta decrease in both M1s than trials for the hand contralateral to the tumor hemisphere.

The third mode discriminated between frequencies in the low mu band (around 8 Hz) and beta band (around 18 Hz), with an opposite modulation around 12 Hz. At first glance, the weight factors seemed again to be related to movement but no significant effects were found.

61

CHAPTER 4

Figure 4. PCA results. The first four modes collectively explained 91% of the variability in the data. The shown weight factors are averages over subjects within a single condition obtained from the eigenvector of the corresponding mode.

The projections show the power deviations from a grand-averaged power spectrum. The weight factors multiplied by the projections indicate the modulations in power for the specific conditions. From the first mode an effect of tumor can be readily observed from the weight factors. An effect of movement becomes apparent in the second mode.

The fourth mode was characterized by a similar projection as the third mode but with peak frequencies shifted to higher frequencies. This time significant effects were found for active (F(1,10) = 6.214, p = .032) and the tumor x movement side x active interaction (F(1,10) = 9.500, p = .012). We note that the main effect of action disappears when applying a Bonferroni correction for multiple comparisons regarding the number of PCA components tested (corrected alpha-level = .0125). The three-way interaction suggests that the difference in weight factors between contralateral and ipsilateral hand were larger for the tumor hemisphere compared to the healthy hemisphere, but only when movements were performed. This was due to a larger difference in weight factors for the

62

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

contralateral hand between movement and rest for the tumor hemisphere. However, it is better to assess any interaction between tumor and action via spectra that have not been normalized for total power, as described in the next session.

Movement-related activity The spectral content that could be ascribed to movement execution did not reveal a shift towards lower frequencies (see Figure 5). On the contrary, the shape of the power spectra for the tumor and healthy hemisphere largely agreed. A few frequency bins however, did show significant differences, mostly for movement of the contralateral hand. In these latter cases, the power modulation was less pronounced for the tumor hemisphere in comparison to the healthy hemisphere. This trend could be observed across a broad frequency range. Since we did not record movement parameters, however, we cannot rule out that this effect may have been caused by differences in task performance like movement frequency. Importantly, the movement-related activity did not show a shift in frequency content as observed for the original spectra in rest and during movement.

Figure 5. Movement-related changes in spectral power. For each subject power during movement was divided by power in rest and spectra were subsequently grand-averaged. Grey patches indicate frequency bins that showed a significant difference between the tumor and healthy hemisphere (7 bins for movement of the contralateral hand, 2 bins for the ipsilateral hand). In contrast to the original power spectra in rest and during movement, a shift in spectral content towards lower frequencies cannot be observed.

DISCUSSION In glioma patients, a significant shift in spectral power towards lower frequencies for M1 in the tumor hemisphere is found compared to M1 in the contralateral, healthy hemisphere. We refer to this effect as â&#x20AC;&#x2DC;slowingâ&#x20AC;&#x2122; of brain oscillations but note that peak mu and beta frequencies were not shifted. A closer look reveals that relative power was increased for frequencies below approximately 17 Hz, whereas higher frequencies showed a decrease in power. This pattern was observed not only in resting state but prevailed during movement. Specific movement-related activity however, did not show such an

63

4

CHAPTER 4

effect. These findings indicate that glioma may cause alterations of M1 mu and beta oscillations that are task-independent.

The clear lateralization of the tumor bulk allowed for a within-subject comparison between healthy and tumor hemisphere. Although overall spectral power in these patients might be altered compared to healthy controls, we observed substantial differences between hemispheres. The use of AEDs in all patients can influence neurotransmitters and networks involved in cerebral motor control, however, this would be a generalized effect and is in our opinion not responsible for the mono-hemispheric findings in our study.

Significant SAM sources were found in contralateral and ipsilateral hemisphere in all subjects, regardless of the presence of a tumor. By contrast, Taniguchi et al. (2004) (48) reported that for brain tumor patients only ipsilateral M1 was significantly activated during movement. They found no sources in the contralateral hemisphere, which was activated in healthy controls. Since our movement tasks were very similar, only differences in tumor characteristics or location may explain the conflicting findings. We studied mainly lowgrade glioma patients with unaffected motor capabilities, whereas Taniguchi et al.’s patients had mainly high-grade gliomas and two of the six patients had motor deficits and one sensory disturbances. As such, the involvement of ipsilateral motor cortex is a common observation in healthy subjects during unimanual movements (5,27,29,30) and is related to task-complexity (33,39,45,53).

Slowing of neural oscillations in terms of increased power for low frequency bands and decreased power for higher frequency bands is a common phenomenon in neurodegenerative diseases like Alzheimer’s (15,20,34) and Parkinson’s disease (46,47), where it has been shown to correlate with disease severity (10,21,23,42). Using a movement paradigm comparable to the one employed in the present study, Vardy et al.

(2011) (52) found a slowing of brain oscillations in M1 for Parkinson’s patients compared to healthy controls by means of increased mu and decreased beta power. Remarkably, this slowing was correlated with disease severity in a task-specific way. The median frequency during movement correlated strongly with the UPDRS motor score, whereas the median frequency during rest correlated with the UPDRS scores on mental functioning.

These findings indicate that slowing of activity is associated with diminished functionality, although no direct causal relations could be inferred. A more direct investigation of the neural mechanisms underlying cortical slowing might possibly be obtained by examining patients with brain lesions.

Our findings resemble those seen over Rolandic regions of patients following monohemispheric stroke (49,50). Not only did they show higher power for frequencies up to the mu band and lower power for beta and gamma bands in the affected hemisphere compared to the unaffected hemisphere, also the overall power in both hemispheres was higher compared to healthy controls. This was ascribed to an increased neural synchronization by the formation of new intra-cortical connections via axonal sprouting and/or gap junctions and enhanced excitability of individual neurons, all as a result of acute deafferentiation. More importantly, the peak alpha frequency showed a correlation with clinical scores and was more reduced in patients with subcortical compared to cortical

64

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

lesions (49,50). Beta power was correlated with hand functionality. Despite the fact that stroke occurs abruptly and tumor growth is a slow and gradual process, the similarity of findings suggests that the manner of disruption is irrelevant for the occurrence of neural slowing in M1.

Looking at generating mechanisms for mu and beta rhythms may help unravel candidate causes of the here-observed changes in spectral content. Regarding the Rolandic mu, like the occipital alpha rhythm, there is accumulating evidence for the presence of thalamocortical circuits displaying oscillatory activity at the according frequencies (32).

The origin of Rolandic beta oscillations, however, seems less clear. Based on indications of the involvement of intra-cortical projections in the expression of beta oscillations, Jones et al. (2009) (35) modeled both mu and beta activity in the primary somatosensory cortex (SI) using a physiologically realistic model. In that model rhythmic 10 Hz feed-forward (FF) input from the lemniscal thalamus to the cortical SI column accounts for the mu rhythm, while the combination with additional 10 Hz feedback (FB) connections from (higherorder) cortical regions or non-lemniscal thalamic nuclei produces beta oscillations.

Specifics of the mu/beta ratio depend on four (types of) parameters: delay between FF and FB input, the variance of input (amount of synchrony), the amplitude of input (number of import bursts), and the post-synaptic conductance. Here most important, the amplitude of FB connections is likely to be diminished in glioma patients due to disrupted intracortical projections. Accordingly, the model predicts both a decrease in beta power and an increase in mu power with lower amplitude of FB input, which may indeed explain our findings, presuming this model is also valid for describing M1 activity.

Despite the slowing of M1 oscillations, patients did not experience movement deficits.

This suggests at least some flexibility in the (relative) strength of mu and beta power underlying normal motor functioning. Moreover, the neural slowing did also not prevent a clear movement-related desynchronization of mu and beta oscillations. We did find evidence for a stronger power decrease for the hand contralateral to the healthy hemisphere than the hand contralateral to the tumor hemisphere. However, we cannot rule out that this effect was caused by a difference in task performance that might have influenced the power spectrum, since we did not record movement parameters. Hence, our major finding is that the tumor causes a slowing of oscillatory activity both in resting state and during movement.

The infiltrative nature of gliomas with displacement and disruption of white matter tracts and cortical invasion can change intra- and interhemispheric neural synchronization (7,12,18), even in the absence of neurologic deficit. Even though these changes occur, specific task-related activity might be unaffected, as demonstrated by the absence of a slowing of spectral power in movement-related activity. A lack of obvious motor deficits could also be the result of the slow tumor growth in the majority of our patients and the functional reorganization (plasticity) that takes place: modulations of synaptic efficiency, unmasking of latent connections, recruitment of other brain areas (19). These factors may have contributed to the observed spectral changes as an adaptation mechanism to the damage caused by the tumor.

65

4

CHAPTER 4

Although gliomas are histopathologically heterogeneous tumors with varying genetic and chromosomal alterations (56), a common feature is their ability to change cerebral networks (structural) in combination with changes in neurotransmitter systems (functional). Regarding the latter, probably the most important observation in gliomas is an increased peritumoral glutamate release causing neuronal hyperexcitability inducing epileptic activity (13,40). All our patients had tumor-associated epilepsy, which is common in brain tumor patients (51), but no epileptic activity was recorded during MEG investigations. At a microscopic level, however, neuronal hyperexcitability due to glutamate may influence thalamocortical loops that may be involved in alpha rhythm generation (31). Another aspect of gliomas is the presence of cholinesterase activity found in tumor biopsies (6), which reduces cholinergic activity by breaking down acetylcholine. Cholinergic deficit is considered a key factor underlying the slowing of oscillatory activity observed in Alzheimerâ&#x20AC;&#x2122;s disease (34). In fact, treatment with cholinesterase inhibitors (promoting cholinergic activity) could reverse the shift towards lower frequencies seen in dementia (1,2,4,9). At the moment it is unclear to what extent these biochemical factors relate to our findings.

The findings of our study may have potential clinical diagnostic implications in the treatment of glioma patients. The slowing of M1 oscillations may prelude clinical motor deficits and could possibly be used to monitor the effects of glioma invasion in (central) motor pathways, before radiological or clinical progression becomes evident. This information might be helpful in determining the timing of surgery or to evaluate other therapeutic measures, e.g., chemotherapy in the treatment of glioma patients. To evaluate the diagnostic value of M1 oscillations in glioma patients, future studies should incorporate a longitudinal study of oscillatory activity in relation to quantitative measures of motor functioning.

CONCLUSION Peri-rolandic tumors decelerate M1 oscillations expressed by decreased beta and increased mu activity. This result was observed both in resting state and during movement along with a lack of task-specificity. Physiological modeling suggests that a loss of intracortical connectivity may account for our empirical findings. The presence of a tumor hence influences neural activity beyond local increases in delta and/or theta oscillations.

66

SLOWING OF M1 OSCILLATIONS IN BRAIN TUMOR PATIENTS

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Adler G, Brassen S, Chwalek K, Dieter B & Teufel M. Prediction of treatment response to rivastigmine in Alzheimer’s dementia. J Neurol Neurosurg Psychiatry 2004;75:292-4.

Adler G & Brassen S. Short-term rivastigmine treatment reduces EEG slow-wave power in Alzheimer patients.

Neuropsychobiology 2001;43:273-6.

Baayen JC, de Jongh A, Stam CJ, et al. Localization of slow wave activity in patients with tumor-associated epilepsy. Brain Topogr 2003;16:85-93.

Balkan S, Yaraş N, Mihçi E, Dora B, Ağar A & Yargiçoğlu P. Effect of donepezil on EEG spectral analysis in Alzheimer’s disease.

Acta Neurol Belg 2003;103:164-9.

Baraldi P, Porro CA, Serafini M, et al. Bilateral representation of sequential finger movements in human cortical areas.

Neurosci Lett 1999;269:95-8.

Barbosa M, Rios O, Velásquez M, Villalobos J & Ehrmanns J. Acetylcholinesterase and butyrylcholinesterase histochemical activities and tumor cell growth in several brain tumors. Surg Neurol 2001;55:106-12.

Bartolomei F, Bosma I, Klein M, et al. How do brain tumors alter functional connectivity? A magnetoencephalography study. Ann Neurol 2006;59:128-38.

Boonstra TW, Daffertshofer A, Breakspear M & Beek PJ. Multivariate time-frequency analysis of electromagnetic brain activity during bimanual motor learning. Neuroimage 2007;36:370-7.

Bosboom JL, Stoffers D, Stam CJ, Berendse HW & Wolters ECh. Cholinergic modulation of MEG resting-state oscillatory activity in Parkinson’s disease related dementia. Clin Neurophysiol 2009;120:910-5.

Bosboom JL, Stoffers D, Stam CJ, et al. Resting state oscillatory brain dynamics in Parkinson’s disease: an MEG study. Clin Neurophysiol 2006;117:2521-31.

Bosma I, Stam CJ, Douw L, et al. The influence of low-grade glioma on resting state oscillatory brain activity: a magnetoencephalography study. J Neurooncol 2008;88:77-85.

Bosma I, Reijneveld JC, Klein M, et al. Disturbed functional brain networks and neurocognitive function in low-grade glioma patients: a graph theoretical analysis of resting-state MEG. Nonlinear Biomed Phys 2009;3:9.

Buckingham SC, Campbell SL, Haas BR, et al. Glutamate release by primary brain tumors induces epileptic activity. Nat Med 2011;17:1269-74.

Cheyne D, Bakhtazad L & Gaetz W. Spatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming approach. Hum Brain Mapp 2006;27:213-29. de Haan W, Stam CJ, Jones BF, Zuiderwijk IM, van Dijk BW & Scheltens P. Resting-state oscillatory brain dynamics in Alzheimer disease. J Clin Neurophysiol 2008;25:187-93. de Jongh A, de Munck JC, Baayen JC, Jonkman EJ, Heethaar RM & van Dijk BW. The localization of spontaneous brain activity: first results in patients with cerebral tumors. Clin Neurophysiol 2001;112:378-85. de Jongh A, Baayen JC, de Munck JC, Heethaar RM, Vandertop WP & Stam CJ. The influence of brain tumor treatment on pathological delta activity in MEG. Neuroimage 2003;20:2291-301.

Douw L, Baayen H, Bosma I, et al. Treatment-related changes in functional connectivity in brain tumor patients: a magnetoencephalography study. Exp Neurol 2008;212:285-90.

Duffau H. Lessons from brain mapping in surgery for low-grade glioma: insights into associations between tumour and brain plasticity. Lancet Neurol 2005;4:476-86.

Fernández A, Maestú F, Amo C, et al. Focal temporoparietal slow activity in Alzheimer’s disease revealed by magnetoencephalography. Biol Psychiatry 2002;52:764-70.

Fernández A, Hornero R, Mayo A, Poza J, Gil-Gregorio P & Ortiz T. MEG spectral profile in Alzheimer’s disease and mild cognitive impairment. Clin Neurophysiol 2006;117:306-14.

Fernández-Bouzas A, Harmony T, Bosch J, et al. Sources of abnormal EEG activity in the presence of brain lesions. Clin Electroencephalogr 1999;30:46-52.

Franciotti R, Iacono D, Della Penna S, et al. Cortical rhythms reactivity in AD, LBD and normal subjects: a quantitative MEG study. Neurobiol Aging 2006;27:1100-9.

Gaetz W, Cheyne D, Rutka JT, et al. Presurgical localization of primary motor cortex in pediatric patients with brain lesions by the use of spatially filtered magnetoencephalography. Neurosurgery 2009;64:177-85; discussion 186.

Giese A, Bjerkvig R, Berens ME & Westphal M. Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol 2003;21:1624-36.

Gloor P, Ball G & Schaul N. Brain lesions that produce delta waves in the EEG. Neurology 1977;27:326-33.

Gross J, Pollok B, Dirks M, Timmermann L, Butz M & Schnitzler A. Task-dependent oscillations during unimanual and bimanual movements in the human primary motor cortex and SMA studied with magnetoencephalography. Neuroimage 2005;26:91-8.

67

4

CHAPTER 4

28. Hillebrand A & Barnes GR. Beamformer analysis of MEG data. Int Rev Neurobiol 2005;68:149-71.

29. Houweling S, Daffertshofer A, van Dijk BW & Beek PJ. Neural changes induced by learning a challenging perceptual-motor task. Neuroimage 2008;41:1395-407.

30. Huang MX, Harrington DL, Paulson KM, Weisend MP & Lee RR. Temporal dynamics of ipsilateral and contralateral motor activity during voluntary finger movement. Hum Brain Mapp 2004;23:26-39.

31. Hughes SW, Errington A, Lorincz ML, et al. Novel modes of rhythmic burst firing at cognitively-relevant frequencies in thalamocortical neurons. Brain Res 2008;1235:12-20.

32. Hughes SW & Crunelli V. Thalamic mechanisms of EEG alpha rhythms and their pathological implications. Neuroscientist 2005;11:357-72.

33. Hummel F, Kirsammer R & Gerloff C. Ipsilateral cortical activation during finger sequences of increasing complexity: representation of movement difficulty or memory load? Clin Neurophysiol 2003;114:605-13.

34. Jeong J. EEG dynamics in patients with Alzheimer’s disease. Clin Neurophysiol 2004;115:1490-505.

35. Jones SR, Pritchett DL, Sikora MA, Stufflebeam SM, Hämäläinen M & Moore CI. Quantitative analysis and biophysically realistic neural modeling of the MEG mu rhythm: rhythmogenesis and modulation of sensory-evoked responses. J Neurophysiol 2009;102:3554-72.

36. Kamada K, Möller M, Saguer M, et al. A combined study of tumor-related brain lesions using MEG and proton MR spectroscopic imaging. J Neurol Sci 2001;186:13-21.

37. Kleihues P, Louis DN, Scheithauer BW, et al. The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 2002;61:215-25; discussion 226-9.

38. Loftus G & Masson M. Using confidence-intervals in within-subject designs. Psychon B Rev 1994;1:476-90.

39. Manganotti P, Gerloff C, Toro C, et al. Task-related coherence and task-related spectral power changes during sequential finger movements. Electroencephalogr Clin Neurophysiol 1998;109:50-62.

40. Marcus HJ, Carpenter KL, Price SJ & Hutchinson PJ. In vivo assessment of high-grade glioma biochemistry using microdialysis: a study of energy-related molecules, growth factors and cytokines. J Neurooncol 2010;97:11-23.

41. McIntosh AR & Lobaugh NJ. Partial least squares analysis of neuroimaging data: applications and advances. Neuroimage 2004;23 Suppl 1:S250-63.

42. Morita A, Kamei S, Serizawa K & Mizutani T. The relationship between slowing EEGs and the progression of Parkinson’s disease. J Clin Neurophysiol 2009;26:426-9.

43. Nagarajan S, Kirsch H, Lin P, Findlay A, Honma S & Berger MS. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109:228-37.

44. Oshino S, Kato A, Wakayama A, Taniguchi M, Hirata M & Yoshimine T. Magnetoencephalographic analysis of cortical oscillatory activity in patients with brain tumors: Synthetic aperture magnetometry (SAM) functional imaging of delta band activity. Neuroimage 2007;34:957-64.

45. Salmelin R, Forss N, Knuutila J & Hari R. Bilateral activation of the human somatomotor cortex by distal hand movements.

Electroencephalogr Clin Neurophysiol 1995;95:444-52.

46. Soikkeli R, Partanen J, Soininen H, Pääkkönen A & Riekkinen P. Slowing of EEG in Parkinson’s disease. Electroencephalogr Clin Neurophysiol 1991;79:159-65.

47. Stoffers D, Bosboom JL, Deijen JB, Wolters EC, Berendse HW & Stam CJ. Slowing of oscillatory brain activity is a stable characteristic of Parkinson’s disease without dementia. Brain 2007;130:1847-60.

48. Taniguchi M, Kato A, Ninomiya H, et al. Cerebral motor control in patients with gliomas around the central sulcus studied with spatially filtered magnetoencephalography. J Neurol Neurosurg Psychiatry 2004;75:466-71.

49. Tecchio F, Zappasodi F, Pasqualetti P, et al. Long-term effects of stroke on neuronal rest activity in rolandic cortical areas.

J Neurosci Res 2006;83:1077-87.

50. Tecchio F, Zappasodi F, Pasqualetti P, et al. Rhythmic brain activity at rest from rolandic areas in acute mono-hemispheric stroke: a magnetoencephalographic study. Neuroimage 2005;28:72-83.

51. van Breemen MS, Wilms EB & Vecht CJ. Epilepsy in patients with brain tumours: epidemiology, mechanisms, and management. Lancet Neurol 2007;6:421-30.

52. Vardy AN, van Wegen EE, Kwakkel G, Berendse HW, Beek PJ & Daffertshofer A. Slowing of M1 activity in Parkinson’s disease during rest and movement--an MEG study. Clin Neurophysiol 2011;122:789-95.

53. Verstynen T, Diedrichsen J, Albert N, Aparicio P & Ivry RB. Ipsilateral motor cortex activity during unimanual hand movements relates to task complexity. J Neurophysiol 2005;93:1209-22.

54. Vrba J & Robinson SE. Signal processing in magnetoencephalography. Methods 2001;25:249-71.

55. Walter W. The location of cerebral tumours by electro-encephalography. Lancet 1936;2:305-8.

56. Wen PY & Kesari S. Malignant gliomas in adults. N Engl J Med 2008;359:492-507.

57. Willemse RB, de Munck JC, Verbunt JP, et al. Topographical organization of mu and Beta band activity associated with hand and foot movements in patients with perirolandic lesions. Open Neuroimag J 2010;4:93-9.

68

69

CHAPTER 5

Magnetoencephalographic study of hand and foot sensorimotor organization in 325 consecutive patients evaluated for tumor or epilepsy surgery Neuroimage: Clinical 10 (2016), 46 - 53

Ronald B. Willemse Arjan Hillebrand Hanneke E. Ronner W. Peter Vandertop Cornelis J. Stam

CHAPTER 5

ABSTRACT Objectives The presence of intracranial lesions or epilepsy may lead to functional reorganization and hemispheric lateralization. We applied a clinical magnetoencephalography (MEG) protocol for the localization of the contralateral and ipsilateral S1 and M1 of the foot and hand in patients with non-lesional epilepsy, stroke, developmental brain injury, traumatic brain injury and brain tumors. We investigated whether differences in activation patterns could be related to underlying pathology.

Methods Using dipole fitting, we localized the sources underlying sensory and motor evoked magnetic fields (SEFs and MEFs) of both hands and feet following unilateral stimulation of the median nerve (MN) and posterior tibial nerve (PTN) in 325 consecutive patients.

The primary motor cortex was localized using beamforming following a self-paced repetitive motor task for each hand and foot.

Results The success rate for motor and sensory localization for the feet was significantly lower than for the hands (motor_hand 94.6 % versus motor_feet 81.8 %, p < 0.001; sensory_hand 95.3 % versus sensory_feet 76.0 %, p < 0.001). MN and PTN stimulation activated 86.6 % in the contralateral S1, with ipsilateral activation < 0.5 %. Motor cortex activation localized contralaterally in 76.1 % (5.2 % ipsilateral, 7.6 % bilateral and 11.1 % failures) of all motor MEG recordings. The ipsilateral motor responses were found in 43 (14 %) out of 308 patients with motor recordings (range: 8.3 â&#x20AC;&#x201C; 50 %, depending on the underlying pathology), and had a higher occurrence in the foot than in the hand (motor_foot 44.8 % versus motor_hand 29.6 %, p = 0.031). Ipsilateral motor responses tended to be more frequent in patients with a history of stroke, traumatic brain injury (TBI) or developmental brain lesions (p = 0.063).

Conclusions MEG localization of sensorimotor cortex activation was more successful for the hand compared to the foot. In patients with neural lesions, there were signs of brain reorganization as measured by more frequent ipsilateral motor cortical activation of the foot in addition to the traditional sensory and motor activation patterns in the contralateral hemisphere. The presence of ipsilateral neural reorganization, especially around the foot motor area, suggests that careful mapping of the hand and foot in both contralateral and ipsilateral hemispheres prior to surgery, might minimize postoperative deficits.

72

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

INTRODUCTION Magnetoencephalography (MEG) in combination with Magnetic Resonance Imaging (MRI) has developed from a research tool into a useful and accepted clinical modality in the management of patients with epilepsy and brain tumors. (2,8,14,29) Using information obtained from MEG in the pre-surgical evaluation of epilepsy increases the success rate of epilepsy surgery (29), and MEG identification of the sensorimotor cortex has been validated by several groups using intraoperative measurements as a support to neurosurgical planning and intraoperative guidance of resection. (8,14,31,49,60) Localization of somatosensory cortex is typically achieved using dipole fitting applied to the 1st main peak of the somatosensory evoked field (SEF) following electrical simulation of the median (MN) or posterior tibial nerve (PTN). (16,17) The changes in oscillatory power in the beta band and mu rhythm following limb movement are typically localized using beam-former approaches (10,18), and have been shown to provide reliable preoperative localization of motor hand cortex in patients with epilepsy and brain tumors. (42) Localization of the hand primary motor and sensory cortex has been studied extensively using MEG, but less is known about the reliability of somatosensory and motor responses of the foot in a clinical setting, especially in the presence of intracranial lesions.

(17,40,43,67,68) The clinical utility of MEG, to map the sensorimotor cortex in surgical candidates depends on the ability to accurately and reliably lateralize and/or localize the primary sensorimotor cortex. In healthy subjects, the strongest activation is typically found contralateral to the side of stimulation or executed movement. (26,54) However, patients with brain lesions may have altered topographic organization of cortical functions, which can affect the results of non-invasive pre-surgical functional mapping (35,53); the occurrence of such reorganization for patients with epilepsy is less clear, and may be related to underlying pathology. It is conceivable that different lesions affect the somatosensory network in different ways. Therefore, knowledge about the structural, as well as functional, changes in the network in the presence of intracranial lesions or epilepsy has clinical significance for pre-surgical planning.

In this paper, we retrospectively evaluated the results of our clinical MEG protocol in a large group of patients, eligible for epilepsy or tumor surgery, with respect to the success rate in locating the contralateral foot primary sensorimotor cortex in comparison to the hand. In addition, we studied whether differences between sensorimotor responses of the hand and foot could be related to underlying pathology.

METHODS The procedures with respect to recording and analysis of responses following electrical median nerve stimulation and hand movements has been described previously by Hillebrand et al. (2013). (20)

73

5

CHAPTER 5

Patients Patients were referred from the VU University Medical Center but also externally from the University Medical Center Utrecht, Utrecht; Kempenhaeghe, Academic Center for Epileptology, Sleep Medicine and Neurological Learning & Development Disability, Heeze and SEIN, Dutch Epilepsy Clinics Foundation, the Netherlands. All 407 consecutive patients referred for clinical MEG from April 2010 until March 2014 were evaluated. All patients had at least MEG recordings with at least analysis of spontaneous activity. The majority of patients also had an additional motor and/or sensory paradigm tested as part of the routine clinical workup. Exclusion of 82 patients who had no sensory or motor paradigm tested, resulted in 325 patients for further analysis. The patientsâ&#x20AC;&#x2122; diagnosis is summarized in Table 1.

Table 1. Diagnosis for all patients.

Diagnosis

Nall (%)

Nincluded (%)

Non-lesional epilepsy

168 (41.3)

134 (41.2)

Focal Cortical Dysplasia

50 (12.3)

45 (13.8)

Low Grade Glioma

50 (12.3)

39 (12)

Mesiotemporal Gliosis

40 (9.8)

26 (8)

Stroke

19 (4.7)

18 (5.5)

DNET

11 (2.7)

9 (2.8)

Cavernoma

11 (2.7)

7 (2.2)

Traumatic Brain Injury

5 (1.2)

4 (1.2)

Developmental Disorder

5 (1.2)

4 (1.2)

Tuberous Sclerosis

5 (1.2)

2 (0.6)

High Grade Glioma

4 (1.0)

1 (0.3)

Cyst

4 (1.0)

4 (1.2)

Other

35 (8.6)

32 (9.8)

Total

407 (100)

325 (100)

DNET: dysembryoplastic neo-epithelial tumor

As the patients were not subjected to procedures and were not required to follow rules of behavior other than routine clinical care, approval of the study by the institutional review board (Medical Ethical Research Committee, VU University Medical Center, Amsterdam, The Netherlands) and informed consent was not required according to the Dutch health law of February 26, 1998 (amended March 1, 2006), i.e. Wet medischwetenschappelijk onderzoek met mensen (WMO; Medical Research Involving Human Subjects Act), Division 1, Section 1.2.

74

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

MEG recordings MEG recordings were obtained using a 306-channel whole-head neuro-magnetometer (Elekta Neuromag Oy, Helsinki, Finland) with subjects lying inside a magnetically shielded room during MEG recordings (Vacuumschmelze GmbH, Hanau, Germany). The head position relative to the MEG sensors was recorded continuously using the signals from four or five head-localization coils. The positions of the coils, as well as the outline of the participants scalp (~500 points), were digitized using a 3D digitizer (3Space Fast-Track, Polhemus, Colchester, VT, USA). This scalp surface was used for co-registration with the patients anatomical MRI. Anatomical MRI and co-registration Structural MR-images were available from previous studies or otherwise acquired with a 1.5 or 3.0 T MR scanner, where the axial slice distance varied from 1.5 to 3 mm. Coregistration of these T1-weighted MRIs with the MEG data was achieved by using surface matching software developed by one of the authors (AH), resulting in an estimated coregistration accuracy of approximately 4 mm. (65) A single best fitting sphere was fitted to the outline of the scalp as obtained from the co-registered MRI, which was used as a volume conductor model for the dipole fitting and beam-former analysis described below.

Somatosensory stimulation MEG responses to electrical stimulation of the left and right median nerve (MN) and the left and right posterior tibial nerve (PTN) were recorded. Constant current square wave pulses (2 Hz, 0.2 ms duration, 500 epochs) were delivered trans-cutaneous at the wrist (MN) and the ankle (PTN) just above motor threshold.

Motor task Subjects performed voluntary hand movements consisting of slow, unilateral, self-paced repetitive non-clenching opening and closing of the hand at about 1 Hz. The movements were performed for 15 repeats of 10 s movement followed by 10 s without movement.

With foot movements patients were instructed to alternate flexion and extension at the ankle at about 1 Hz. Movement instructions were presented to the subject using a brief tone (movement) or brief burst of white noise (no movement). Movements were monitored on camera. Left and right movements of the hand and foot were performed in separate runs.

Analysis The MEG recordings were analyzed according to standard clinical procedures for presurgical mapping of somatosensory and motor cortex by an experienced MEG/EEG technician, and evaluated by a team consisting of two experienced clinical neurophysiologists (HR and CJS), MEG/EEG technicians and physicists (AH).

75

5

CHAPTER 5

Dipole fitting Somatosensory evoked fields were offline-averaged (from -100 msec to + 150 msec with respect to stimulus onset). In accordance with our standard procedure for localization of early SEF responses, the evoked response was low-pass filtered at 70 Hz and high-pass filtered at 0.5 Hz, after baseline correction based on the window from -50 to 0 msec. A single equivalent current dipole (ECD) was fitted to the peak of the SEF of the contralateral hemisphere during the first prominent deflection (using xfit, version 5.5.18, Elekta Neuromag, Oy, Helsinki, Finland).

Beamformer analysis The MEG data for the motor task were band-pass filtered from 13 – 30 Hz (beta band) before sources were reconstructed using a dual-state beamformer (beamformer, Elekta Neuromag, Oy, Helsinki, Finland). A grid covering the entire brain, using a spacing of 5 mm, was used for the localization of changes in neuronal power. (22) A scalar beamformer implementation was used, which determines the optimal current orientation for each voxel. (51)The 10 seconds preceding each auditory cue to start (hand or foot) movements were used for the control period and the 10 s following the auditory cue were used for the active period. Approximately 10 s of data, taken from the beginning of the recording (before the task was started), were used to estimate the noise covariance. Taking this noise covariance and the data co-variances from the active and passive periods, the pseudo-t metric was computed for each voxel in the source grid. (64) Statistical analysis Results are presented for each of the (maximum four) sensory and (maximum four) motor paradigms as a contra- (C), ipsi- (I) or bilateral (B) result and failures were defined as MEG recordings with un-interpretable results. The success rate of the MEG recordings for each paradigm was defined as the number of MEG recordings with an interpretative result (C-, I- or B) divided by the total number of performed MEG recordings, i.e. excluding the MEG recordings with failures. Paired proportions were analysed by means of McNemar’s test, significance was set at p < 0.05 (IBM SPSS Statistics, version 22).

RESULTS Of the 325 patients, 168 (51.7 %) were male (mean age: 29.3 yrs, range 3.7 – 65.5) and 74 (22.8%) were under the age of 17 (43 male (58.1 %); mean age: 11.4; range 3.7 – 16.8).

The distribution of the MEG recordings in all patients is shown in Table 2. The majority of patients (N = 291; 89.5%) had a sensory and motor paradigm recorded. Seventeen (5.2%) patients only had data with sensory stimulation and 17 (5.2%) patients only had motor data available. In total, 1025 sensory and 1042 motor MEG recordings were performed.

The localization and lateralization results for all sensory and motor recordings are shown in Table 2, including the number of failures and success rates.

76

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

Table 2. Number of MEG recordings, localization results, failures and success rate (%) after sensory stimulation and during motor tasks of the extremities.

Sensory MEG C (%)

I (%)

B (%)

Failure (%)

Total (%)

Succes %

MN-L

282 (95.6)

1 (0.3)

0

12 (4.1)

295 (100)

95.9

MN-R

282 (94.6)

0 (0.0)

0

16 (5.4)

298 (100)

94.6

PTN-L

163 (75.8)

3 (1.4)

0

49 (22.8)

215 (100)

77.2

PTN-R

161 (74.2)

1 (0.5)

0

55 (25.3)

217 (100)

74.7

Total

888 (86.6)

5 (0.5)

0 (0.0)

132 (12.9)

1025 (100)

87.1

Motor MEG C (%)

I (%)

B (%)

Failure (%)

Total (%)

Succes %

HL

238 (82.1)

12 (4.1)

27 (9.3)

13 (4.5)

290 (100)

95.5

HR

239 (80.5)

12 (4.0)

26 (8.8)

20 (6.7)

297 (100)

93.3

FL

154 (68.1)

18 (8.0)

15 (6.6)

39 (17.3)

226 (100)

82.7

FR

162 (70.7)

12 (5.2)

11 (4.8)

44 (19.2)

229 (100)

80.8

Total

793 (76.1)

54 (5.2)

79 (7.6)

116 (11.1)

1042 (100)

88.9

5

C: contralateral; I: ipsilateral; B: bilateral; MN: median nerve; PTN: posterior tibial nerve; L: left; R: right; HL: left hand; HR: right hand; FL: left foot; FR: right foot;

Somatosensory MEG Somatosensory MEG recordings localized the contralateral primary sensory cortex in 86.6 % of all recordings (mean MN 95.1 % versus mean PTN 75.0 %). Ipsilateral sensory responses were rare (0.5 %); four patients were identified with five ipsilateral responses.

One patient had a lobar hemi-microencephaly (see Figure 1), one had non-lesional epilepsy (NLE), one had mesial temporal sclerosis (MTS) and one had a left frontal oligodendroglioma WHO grade II. Failures occurred in 132 (12.9 %) of the sensory MEG recordings. MEG recordings after MN stimulation had a significantly (p < 0.001) higher success rate (95.3 %) compared to PTN stimulation (76.0 %).

Motor MEG Localization following hand movement was more successful (94.6 %) than for foot movement (81.8 %; p < 0.001). Motor tasks resulted in contralateral localization in 76.1 % of the recordings; 5.2 % of the MEG recordings had ipsilateral motor results, 7.6 % bilateral activation patterns and 11.1 % failures. Hand movements resulted in a mean of 4.1 % ipsilateral and 9.1 % bilateral responses, and foot movements had a mean of 6.6 % ipsi- and 5.7 % bilateral responses.

Ipsilateral motor recordings Of the 308 patients who had motor recordings, 43 (14 %) patients (28 male; mean age: 32.6 yrs; range 7.8 â&#x20AC;&#x201C; 64.7) had one or more ipsilateral responses. The distribution of the different localization results is shown in Table 3, where it can be seen that the incidence

77

CHAPTER 5

of ipsilateral motor responses varied between 8.3 â&#x20AC;&#x201C; 50% for different patient groups.

Patients with a history of stroke, TBI or a developmental disorder had a relatively high occurrence of ipsilateral results. However, the comparison of this combined patient group (stroke, TBI and developmental disorder) versus the group with slow- or non - (growing) lesions (NLE, FCD, MTS and LGG) did not reach statistical significance (p = 0.063). The occurrence of ipsilateral results in patients with NLE is similar to patients with lesional epilepsy (MTS, DNET, FCD). Table 4 shows the distribution of the MEG results between the hand and foot in the patients with ipsilateral motor responses.

Motor tasks of the foot showed significantly (p = 0.031) more ipsilateral responses than for the hand. Ipsilateral responses of the hand were equally distributed between left- and right hand motor tasks. Left foot movement resulted in significantly more ipsilateral responses (9.6 %) versus right foot movement (6.5 %, p = 0.046). Examples of ipsilateral hand and foot responses are shown in Figure 2.

Figure. 1. Axial (left), coronal (middle) and sagittal (right) MR images corresponding to a 35-year-old male with a leftsided motor weakness since the age of six months with symptomatic therapy-resistant epilepsy with right-sided lobar hemi-microencephaly of the frontal lobe and insular region, in addition to polymicrogyric pachygyria. The only ipsilateral median nerve (MN) result in the patient group was found in this patient, who also had ipsilateral activation after posterior tibial nerve (PTN) stimulation and for both hand and foot motor responses.

Top panels: results of MN and PTN stimulation on both sides showing ipsilateral MN and PTN responses.

Lower panels: task-related power decreases in the beta band during self-paced hand and foot movements with localization in the ipsilateral hemisphere for both hand and foot movements. L: left; R: right; A: anterior; P: posterior;

78

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

Table 3. Number of patients (N) with ipsilateral motor responses (N ipsi).

Patients Diagnosis

N

%

N_ipsi

% of N

NLE

128

41.6

14

10.9

FCD

42

13.6

5

11.9

LGG

36

11.7

4

11.1

MTS

24

7.8

2

8.3

Stroke

18

5.8

5

27.8

DNET

8

2.6

1

12.5

Cavernoma

7

2.3

1

14.3

TBI

4

1.3

2

50.0

Develop Disord

4

1.3

2

50.0

Cyst

4

1.3

0

0.0

Tuberous Sclerosis

2

0.6

0

0.0

HGG

1

0.3

0

0.0

Other

30

9.7

7

23.3

308

100

43

14.0

5

NLE: nonlesional epilepsy; LGG: low grade glioma; FCD: focal cortical dysplasia; MTS: mesiotemporal sclerosis; DNET: dysembryoplastic neo-epithelial tumor; TBI: traumatic brain injury; Develop. Disorder: developmental disorder; HGG: high-grade glioma

Figure. 2. Axial (left) and coronal (right) MR images showing two examples of task-related power decreases in the beta band, demonstrating ipsilateral motor responses in Case 2 and 3 during hand or foot movements.

Case 2: 9-year-old female with intractable and non-lesional epilepsy showing pronounced ipsilateral cortical responses for both hands with localization in the hand area. Foot movements, despite good performance, show no cortical response (not shown).

Case 3: 36-year-old female with focal cortical dysplasia in the depth of the central sulcus of the left hemisphere, shows a contralateral response of both hands and the left foot. Right foot movements show an ipsilateral response at the medial wall of the primary motor cortex. L: left; R: right; A: anterior; P: posterior

79

CHAPTER 5

Table 4. Distribution of MEG results in the patients with ipsilateral motor recordings.

MEG recording

N

N_ipsi (%)

Hand

81

24 (29.6)

Foot

67

30 (44.8)

DISCUSSION In the present study we used MEG to assess the functional organization of the hand and foot sensorimotor cortex in a heterogeneous group of patients evaluated for epilepsy or tumor surgery. Somatosensory responses after PTN stimulation are less successful than MN responses and ipsilateral somatosensory results are rare. MEG motor recordings localize the contralateral M1 in the majority of cases but also show ipsi- and bilateral M1 activation with differential occurrence in patient groups, especially with foot movements.

Ipsilateral somatosensory cortex localization The most reliably elicited somatosensory responses are the early responses (N20m and P40m for MN and PTN stimulation respectively), which usually cause a contralateral response at S1. MN stimulation gives robust results and has been validated with structural and intraoperative cortical stimulation mapping. (14,49) In our study, we also found robust contralateral results after MN and PTN stimulation.

However, few patients (and only 0.5 % of the MEG recordings) showed ipsilateral somatosensory responses. In adults, ipsilateral somatosensory responses have been described with MEG in normal subjects (32,33), as well as in patients. (28) In a large group of 482 heterogeneous patients, 2.9 % of the patients showed an ipsilateral MN response, but no relation could be established between the underlying disease and the presence of an ipsilateral response, which is in accordance with our results. Ipsilateral MN responses have also been described in cerebral palsy (15,66), as in one of our patients (Figure 1).

It has been hypothesized that an ipsilateral response is a normal variant in the population (28) but another possible explanation for this rare occurrence is a lack of transcallosal inhibition of the ipsilateral S1 area after unilateral somatosensory stimulation. (23) Others, using either a different source model (MEG) or imaging modality (fMRI), found that unilateral MN stimulation can activate both the left and right S1 in healthy subjects.

(33,56) However, as usually the single ECD model is used for clinical MEG applications, as in our study, we can only make comparisons with previous studies using the same model.

There are only a few studies with PTN-evoked magnetic responses in the presence of intracranial pathology (49,67), and ipsilateral PTN responses have not been described previously with MEG. We found four ipsilateral PTN responses in three patients with different pathology and the only ipsilateral MN response was also found in one of these patients (Figure 1). It is unclear whether this lateralization reflects functional reorganization or whether other factors may contribute. For MN SEFs, we know that tactile interference

80

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

can reduce the contralateral response and can increase the ipsilateral response, possibly via excitatory transcallosal pathways. (50) We do not know whether this is also true for the foot and whether patients induced this involuntarily. Others however, have concluded that ipsilateral responses after PTN stimulation might also be explained by its location, adjacent to the inter-hemispheric fissure. (27) Our data confirm that ipsilateral somatosensory cortical activation is rare, and that somatosensory cortical activation remains in the lesioned hemisphere, even in the presence of lesions, in contrast to motor activation. (66) Ipsilateral hand responses Unilateral hand movements usually give the strongest activation patterns in contralateral M1 (cM1). Ipsilateral motor cortex involvement during unilateral voluntary movements has been reported with MEG studies (10,34,59) and has also been demonstrated using beamformer analysis in healthy subjects (25) and patients with intracranial lesions.

(42,58,68) Exclusive ipsilateral activity was found in the sensorimotor cortex, the premotor cortex (PMC) and the inferior parietal lobule (IPL) during movement of the affected hand in brain tumor patients with mostly high-grade gliomas around the central sulcus, and it was hypothesized that recruitment of ipsilateral motor areas was needed to maintain effective movement of the affected hand. (58) Using fMRI, Tozakidou (2013) found an increased occurrence of ipsilateral M1 activation in a large group of patients with tumors of the central region, especially in fast-growing lesions such as grade IV gliomas and metastases. (61) In our study, we only had one patient with a high-grade glioma and hence we were not able to confirm these findings. We only found exclusive ipsilateral activation in a small proportion of the datasets, which is in accordance with the findings of others. (42,68) Enhanced ipsilateral acitivity has been described in the presence of pathology. (7,55) Carpentier et al. showed with fMRI in a heterogeneous group of patients with different types of tumors, arteriovenous malformations and epileptogenic cortical malformations, that ipsilateral activation was more pronounced in the latter group. (7) It is generally accepted that such lesions, acquired in the pre- and perinatal period can be compensated easier by the immature brain than the adult brain, with ipsilateral takeover of motor functions. (55) This could explain the findings from Carpentier et al., but also the increased ipsilateral responses in our patients with developmental disorders. Patients with a history of stroke and TBI have also shown frequent ipsilateral responses, which could be explained by a disruption of inter-hemispheric inhibition as described in stroke patients using the recovered hand (6) or TBI patients with corpus callosum lesions. (57) The majority of our patients had lesional (MTS, FCD), or non-lesional epilepsy and data about ipsilateral motor responses in these patient groups are rare, but could be potentially interesting for resective epilepsy surgery around motor areas. We only found one report of an ipsilateral response in MTS (11) and one case report with FCD. (38) Recently, Mäkelä et al. described two patients from a group of 19 patients with intractable epilepsy, with unexpected motor cortex localization, of whom one had a history of a large perinatal

81

5

CHAPTER 5

vascular infarction in the left hemisphere and ipsilateral hand motor cortex representation in the right hemisphere. (39) Motor cortex plasticity in the presence of epilepsy could be the result of functional network alterations in lesional, perilesional but also remote neocortical areas as has been described in MTS and FCD. (5) It is conceivable that patients with NLE, which is a heterogeneous group with different epileptogenic mechanisms, also show functional reorganization comparable to epileptic patients with structural lesions.

In healthy adults, there is increasing evidence that upper limb function relies on the balanced control of cM1 and ipsilateral M1 (iM1) and it is assumed that iM1 assists cM1 by modulating the extent of transcallosal inhibition. (1,13,30,62,70) Before activation of both primary motor cortices in unilateral movements there is evidence that ipsilateral PMC activity precedes activity in cM1 (24,36), which may explain bilateral activity found in fMRI studies. It is possible that the ipsilateral responses in our study are related to PMC activity instead of M1 activity, because they are in close anatomical relationship with overlap in temporal dynamics, or that the result is an average of both, which makes it difficult to disentangle these sources. Finally, participation of the ipsilateral sensorimotor cortex in unilateral limb movements by the partially uncrossed descending fibers of the corticospinal tract recently has been reconsidered as a compensatory pathway in stroke patients. (3) Ipsilateral foot responses We found an increased occurrence of ipsilateral responses after foot movements. This could not be explained by previous M/EEG studies on lower limb movements which indicated bilateral activity over the sensorimotor areas after voluntary movements (41,45,47) or the contribution of the PMC and supplementary motor area (SMA). (46) Comparable fMRI studies however, have shown that active ankle dorsiflexion was associated with a greater relative contribution of iM1 and PMC than finger movements (12,48), suggesting a more significant role of iM1 in the motor planning of lower limb movements than for similar hand movements. These findings may explain the increased ipsilateral foot responses in our study and previous findings in patients with perirolandic lesions. (68) Another possible explanation could be related to task complexity, where ipsilateral activation, especially in M1, has been considered to reflect the degree of task complexity for the upper limb. (63) A comparable study with respect to the lower limb found significant differences between ipsilateral sensorimotor activity during largeamplitude (40°) dorsiflexion at 0.5 Hz, compared with small amplitude (15°) dorsiflexion, suggesting that larger amplitude dorsiflexion is a more difficult task. (37) Our patients were instructed to perform ankle dorsiflexion and plantar flexion in the most comfortable way at about 1 Hz, however we do not consider foot movements more complex than hand movements as a possible explanation for the increased ipsilateral foot responses.

The previous studies were all performed in healthy subjects and our results may indicate functional plasticity in the presence of disease. However, since EMG was not performed in our study, we cannot rule out the possibility of subtle mirror movements of the contralateral limb as another explanation. Future investigation of foot motor function is required to elucidate the clinical value of ipsilateral motor responses.

82

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

Our results demonstrate that the motor network in a large heterogeneous population of patients with lesions or epilepsy (or both) mostly show expected contralateral responses, but also that ipsilateral motor responses may occur. This may have implications for surgical planning in order to avoid post-operative motor deficits (see example in Figure 3).

5

Figure 3. Axial (left), and sagittal (right) MR images of a 40-year-old patient with recurrent motor seizures (epilepsia partialis continua) of the left hand since the age of two, due to cortical dysplasia in the hand area of the right motor cortex. The MEG motor findings suggested functional reorganization with lateralization of the left hand to the left hemisphere (task-related power decreases in the beta band during self-paced movements of the left hand (red) and right hand (green). Identification of the epileptogenic focus was necessary with invasive techniques. The MEG results supported additional functional mapping as well and therefore, subdural grid monitoring of the right hemisphere was performed, showing an extensive and scattered area for motor hand function. Surgery was performed with premotor removal of tissue and multiple subpial transections of the right hand motor cortex. Postoperatively, the patient had no neurological deficits of the left arm and a significant seizure reduction. L: left; R: right; A: anterior; P: posterior

Limitations of the study The results of motor mapping with a beamformer approach depend on thresholding the pseudo-t value at each voxel location. Usually motor activity of the hand shows a strong contralateral activity peak and a weak ipsilateral activity peak, depending on the threshold. Inversely, the occurrence of ipsilateral responses in our group, does not rule out the possibility of a concurrent contralateral responses as well. For practical purposes, we choose the side of the strongest activity and the results of bilateral responses, i.e. nearly equal responses were not evaluated as a separate group. Furthermore, the clinical imaging protocol only evaluated decreases in beta band spectral power, information on a possible increase in beta band power is therefore lacking in the analyses presented here.

83

CHAPTER 5

An increase of beta band power can be found in sensorimotor areas following voluntary movement and somatosensory stimulation (44), but the clinical value with respect to localization of motor execution is unclear. The spatial resolution of MEG with respect to source localization of midline structures such as the foot sensorimotor cortex is an important issue. Detectability of sources in foot sensorimotor cortex by MEG, may be compromised by their depth, yet these sources have a favorable orientation (19).

Beamformer analysis offers further improvements in spatial resolution compared to dipole fitting (see Appendix in Hillebrand and Barnes 2003) (21), and the spatial resolution was optimized through the use of a large number of channels, a long covariance window, and the use of a bandwidth that matched the frequency distribution of the signals of interest (4). However, despite these efforts and the fact that the majority of patients showed foot activation at the expected contralateral foot motor cortex, we cannot rule out the possibility of false foot-lateralization in some cases.

Even in this large group of 325 patients the number of patients with signs of functional reorganization is too small for reliable statistical inferences. Future studies in larger patient groups with stroke, TBI and developmental disorders will have to show whether the observed trend towards increased ipsilateral responses in these patient groups is a consistent finding.

The failures of the MEG recordings could be attributed to some technical constraints (e.g. the presence of a vagal nerve stimulator or stimulation artifacts) or patient-related factors, such as fear for electrical stimulation, restlessness or mirror movements.

Data about handedness were only partially available and not analyzed. Although handedness is known to affect hemispheric asymmetry (52,63), this is usually related to task complexity. Simple motor tasks, as in our study, elicit similar responses in both hands, which were also found in our study as well as in other studies with intracranial lesions. (10,42,68) With respect to the results of sensory stimulation, sensory dominance has not been established with electrical median nerve or pneumatically driven finger stimulation. (9,69)

CONCLUSION A clinical imaging protocol using MEG with respect to sensorimotor cortex activation has a high success rate with respect to identification of the contralateral sensorimotor cortex.

Functional reorganization in the primary somatosensory cortex is rare, but can occur in the primary motor cortex in patients with intracranial lesions and non-lesional epilepsy, especially during foot movements. The presence of ipsilateral neural reorganization, especially around the foot motor area, may support clinicians to perform careful mapping of the hand and foot in both hemispheres prior to surgery, to minimize postoperative deficits.

84

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Allison JD, Meador KJ, Loring DW, Figueroa RE & Wright JC. Functional MRI cerebral activation and deactivation during finger movement. Neurology 2000;54:135-42.

Anderson CT, Carlson CE, Li Z & Raghavan M. Magnetoencephalography in the preoperative evaluation for epilepsy surgery. Curr Neurol Neurosci Rep 2014;14:446.

Bradnam LV, Stinear CM & Byblow WD. Ipsilateral motor pathways after stroke: implications for non-invasive brain stimulation. Front Hum Neurosci 2013;7:184.

Brookes MJ, Vrba J, Robinson SE, et al. Optimising experimental design for MEG beamformer imaging. Neuroimage 2008;39:1788-802.

Caciagli L, Bernhardt BC, Hong SJ, Bernasconi A & Bernasconi N. Functional network alterations and their structural substrate in drug-resistant epilepsy. Front Neurosci 2014;8:411.

Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR & Welch KM. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112-22.

Carpentier AC, Constable RT, Schlosser MJ, et al. Patterns of functional magnetic resonance imaging activation in association with structural lesions in the rolandic region: a classification system. J Neurosurg 2001;94:946-54.

Castillo EM, Simos PG, Wheless JW, et al. Integrating sensory and motor mapping in a comprehensive MEG protocol: clinical validity and replicability. Neuroimage 2004;21:973-83.

Chen AC, Theuvenet PJ, de Munck JC, Peters MJ, van Ree JM & Lopes da Silva FL. Sensory handedness is not reflected in cortical responses after basic nerve stimulation: a MEG study. Brain Topogr 2012;25:228-40.

Cheyne D, Bakhtazad L & Gaetz W. Spatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming approach. Hum Brain Mapp 2006;27:213-29.

Chlebus P, Brázdil M, Hlustík P, Mikl M, Pazourková M & Krupa P. Handedness shift as a consequence of motor cortex reorganization after early functional impairment in left temporal lobe epilepsy--an fMRI case report. Neurocase 2004;10:326-9.

Ciccarelli O, Toosy AT, Marsden JF, et al. Identifying brain regions for integrative sensorimotor processing with ankle movements. Exp Brain Res 2005;166:31-42.

Ferbert A, Priori A, Rothwell JC, Day BL, Colebatch JG & Marsden CD. Interhemispheric inhibition of the human motor cortex. J Physiol 1992;453:525-46.

Ganslandt O, Fahlbusch R, Nimsky C, et al. Functional neuronavigation with magnetoencephalography: outcome in 50 patients with lesions around the motor cortex. Neurosurg Focus 1999;6:e3.

Guzzetta A, Bonanni P, Biagi L, et al. Reorganisation of the somatosensory system after early brain damage. Clin Neurophysiol 2007;118:1110-21.

Hari R & Forss N. Magnetoencephalography in the study of human somatosensory cortical processing. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 1999;354:1145-54.

Hari R, Nagamine T, Nishitani N, et al. Time-varying activation of different cytoarchitectonic areas of the human SI cortex after tibial nerve stimulation. Neuroimage 1996;4:111-8.

Hillebrand A & Barnes GR. Beamformer analysis of MEG data. Int Rev Neurobiol 2005;68:149-71.

Hillebrand A & Barnes GR. A quantitative assessment of the sensitivity of whole-head MEG to activity in the adult human cortex. Neuroimage 2002;16:638-50.

Hillebrand A, Fazio P, de Munck JC & van Dijk BW. Feasibility of clinical magnetoencephalography (MEG) functional mapping in the presence of dental artefacts. Clin Neurophysiol 2013;124:107-13.

Hillebrand A & Barnes GR. The use of anatomical constraints with MEG beamformers. Neuroimage 2003;20:2302-13.

Hillebrand A, Singh KD, Holliday IE, Furlong PL & Barnes GR. A new approach to neuroimaging with magnetoencephalography. Hum Brain Mapp 2005;25:199-211.

Hlushchuk Y & Hari R. Transient suppression of ipsilateral primary somatosensory cortex during tactile finger stimulation. J Neurosci 2006;26:5819-24.

Huang MX, Harrington DL, Paulson KM, Weisend MP & Lee RR. Temporal dynamics of ipsilateral and contralateral motor activity during voluntary finger movement. Hum Brain Mapp 2004;23:26-39.

Jurkiewicz MT, Gaetz WC, Bostan AC & Cheyne D. Post-movement beta rebound is generated in motor cortex: evidence from neuromagnetic recordings. Neuroimage 2006;32:1281-9.

Kakigi R, Hoshiyama M, Shimojo M, et al. The somatosensory evoked magnetic fields. Prog Neurobiol 2000;61:495523.

85

5

CHAPTER 5

27. Kakigi R, Koyama S, Hoshiyama M, Shimojo M, Kitamura Y & Watanabe S. Topography of somatosensory evoked magnetic fields following posterior tibial nerve stimulation. Electroencephalogr Clin Neurophysiol 1995;95:127-34.

28. Kanno A, Nakasato N, Hatanaka K & Yoshimoto T. Ipsilateral area 3b responses to median nerve somatosensory stimulation. Neuroimage 2003;18:169-77.

29. Knowlton RC. Can magnetoencephalography aid epilepsy surgery? Epilepsy Curr 2008;8:1-5.

30. Kobayashi M, Hutchinson S, Schlaug G & Pascual-Leone A. Ipsilateral motor cortex activation on functional magnetic resonance imaging during unilateral hand movements is related to interhemispheric interactions. Neuroimage 2003;20:2259-70.

31. Korvenoja A, Kirveskari E, Aronen HJ, et al. Sensorimotor cortex localization: comparison of magnetoencephalography, functional MR imaging, and intraoperative cortical mapping. Radiology 2006;241:213-22.

32. Korvenoja A, Wikstrom H, Huttunen J, et al. Activation of ipsilateral primary sensorimotor cortex by median nerve stimulation. Neuroreport 1995;6:2589-93.

33. Korvenoja A, Huttunen J, Salli E, et al. Activation of multiple cortical areas in response to somatosensory stimulation: combined magnetoencephalographic and functional magnetic resonance imaging. Hum Brain Mapp 1999;8:13-27.

34. Kristeva R, Cheyne D & Deecke L. Neuromagnetic fields accompanying unilateral and bilateral voluntary movements: topography and analysis of cortical sources. Electroencephalogr Clin Neurophysiol 1991;81:284-98.

35. Lee HW, Shin JS, Webber WR, Crone NE, Gingis L & Lesser RP. Reorganisation of cortical motor and language distribution in human brain. J Neurol Neurosurg Psychiatry 2009;80:285-90.

36. Liuzzi G, Hörniss V, Hoppe J, et al. Distinct temporospatial interhemispheric interactions in the human primary and premotor cortex during movement preparation. Cereb Cortex 2010;20:1323-31.

37. MacIntosh BJ, Mraz R, Baker N, Tam F, Staines WR & Graham SJ. Optimizing the experimental design for ankle dorsiflexion fMRI. Neuroimage 2004;22:1619-27.

38. Maegaki Y, Yamamoto T & Takeshita K. Plasticity of central motor and sensory pathways in a case of unilateral extensive cortical dysplasia: investigation of magnetic resonance imaging, transcranial magnetic stimulation, and short-latency somatosensory evoked potentials. Neurology 1995;45:2255-61.

39. Mäkelä JP, Vitikainen AM, Lioumis P, et al. Functional plasticity of the motor cortical structures demonstrated by navigated TMS in two patients with epilepsy. Brain Stimul 2013;6:286-91.

40. Mäkelä JP, Kirveskari E, Seppä M, et al. Three-dimensional integration of brain anatomy and function to facilitate intraoperative navigation around the sensorimotor strip. Hum Brain Mapp 2001;12:180-92.

41. Müller-Putz GR, Zimmermann D, Graimann B, Nestinger K, Korisek G & Pfurtscheller G. Event-related beta EEGchanges during passive and attempted foot movements in paraplegic patients. Brain Res 2007;1137:84-91.

42. Nagarajan S, Kirsch H, Lin P, Findlay A, Honma S & Berger MS. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109:228-37.

43. Nakamura A, Yamada T, Goto A, et al. Somatosensory homunculus as drawn by MEG. Neuroimage 1998;7:377-86.

44. Neuper C, Wörtz M & Pfurtscheller G. ERD/ERS patterns reflecting sensorimotor activation and deactivation. Prog Brain Res 2006;159:211-22.

45. Pfurtscheller G, Neuper C, Andrew C & Edlinger G. Foot and hand area mu rhythms. Int J Psychophysiol 1997;26:121-35.

46. Pittaccio S, Zappasodi F, Tamburro G, et al. Passive ankle dorsiflexion by an automated device and the reactivity of the motor cortical network. Conf Proc IEEE Eng Med Biol Soc 2013;2013:6353-6.

47. Pittaccio S, Zappasodi F, Viscuso S, et al. Primary sensory and motor cortex activities during voluntary and passive ankle mobilization by the SHADE orthosis. Hum Brain Mapp 2011;32:60-70.

48. Sahyoun C, Floyer-Lea A, Johansen-Berg H & Matthews PM. Towards an understanding of gait control: brain activation during the anticipation, preparation and execution of foot movements. Neuroimage 2004;21:568-75.

49. Schiffbauer H, Berger MS, Ferrari P, Freudenstein D, Rowley HA & Roberts TP. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97:1333-42.

50. Schnitzler A, Salmelin R, Salenius S, Jousmäki V & Hari R. Tactile information from the human hand reaches the ipsilateral primary somatosensory cortex. Neurosci Lett 1995;200:25-8.

51. Sekihara K, Nagarajan SS, Poeppel D & Marantz A. Asymptotic SNR of scalar and vector minimum-variance beamformers for neuromagnetic source reconstruction. IEEE Trans Biomed Eng 2004;51:1726-34.

52. Solodkin A, Hlustik P, Noll DC & Small SL. Lateralization of motor circuits and handedness during finger movements.

Eur J Neurol 2001;8:425-34.

53. Staudt M. Reorganization after pre- and perinatal brain lesions. J Anat 2010;217:469-74.

54. Stippich C, Blatow M, Durst A, Dreyhaupt J & Sartor K. Global activation of primary motor cortex during voluntary movements in man. Neuroimage 2007;34:1227-37.

86

SENSORIMOTOR MAPPING OF THE HAND AND FOOT IN PATIENTS STUDIED WITH MEG

55. Stoeckel MC & Binkofski F. The role of ipsilateral primary motor cortex in movement control and recovery from brain damage. Exp Neurol 2010;221:13-7.

56. Sutherland MT & Tang AC. Reliable detection of bilateral activation in human primary somatosensory cortex by unilateral median nerve stimulation. Neuroimage 2006;33:1042-54.

57. Takeuchi N, Oouchida Y & Izumi S. Motor control and neural plasticity through interhemispheric interactions. Neural Plast 2012;2012:823285.

58. Taniguchi M, Kato A, Ninomiya H, et al. Cerebral motor control in patients with gliomas around the central sulcus studied with spatially filtered magnetoencephalography. J Neurol Neurosurg Psychiatry 2004;75:466-71.

59. Taniguchi M, Kato A, Fujita N, et al. Movement-related desynchronization of the cerebral cortex studied with spatially filtered magnetoencephalography. Neuroimage 2000;12:298-306.

60. Tarapore PE, Tate MC, Findlay AM, et al. Preoperative multimodal motor mapping: a comparison of magnetoencephalography imaging, navigated transcranial magnetic stimulation, and direct cortical stimulation. J Neurosurg 2012;117:354-62.

61. Tozakidou M, Wenz H, Reinhardt J, et al. Primary motor cortex activation and lateralization in patients with tumors of the central region. Neuroimage (Amst) 2013;2:221-8.

62. van Wijk BC, Beek PJ & Daffertshofer A. Differential modulations of ipsilateral and contralateral beta (de) synchronization during unimanual force production. Eur J Neurosci 2012;36:2088-97.

63. Verstynen T, Diedrichsen J, Albert N, Aparicio P & Ivry RB. Ipsilateral motor cortex activity during unimanual hand movements relates to task complexity. J Neurophysiol 2005;93:1209-22.

64. Vrba J & Robinson SE. Signal processing in magnetoencephalography. Methods 2001;25:249-71.

65. Whalen C, Maclin EL, Fabiani M & Gratton G. Validation of a method for coregistering scalp recording locations with 3D structural MR images. Hum Brain Mapp 2008;29:1288-301.

66. Wilke M, Staudt M, Juenger H, Grodd W, Braun C & Krägeloh-Mann I. Somatosensory system in two types of motor reorganization in congenital hemiparesis: topography and function. Hum Brain Mapp 2009;30:776-88.

67. Willemse RB, de Munck JC, vanâ&#x20AC;&#x2122;t Ent D, et al. Magnetoencephalographic study of posterior tibial nerve stimulation in patients with intracranial lesions around the central sulcus. Neurosurgery 2007;61:1209-17; discussion 1217-8.

68. Willemse RB, de Munck JC, Verbunt JP, et al. Topographical organization of mu and Beta band activity associated with hand and foot movements in patients with perirolandic lesions. Open Neuroimag J 2010;4:93-9.

69. Zhu Z, Disbrow EA, Zumer JM, McGonigle DJ & Nagarajan SS. Spatiotemporal integration of tactile information in human somatosensory cortex. BMC Neurosci 2007;8:21.

70. Ziemann U, Ishii K, Borgheresi A, et al. Dissociation of the pathways mediating ipsilateral and contralateral motorevoked potentials in human hand and arm muscles. J Physiol 1999;518 ( Pt 3):895-906.

87

5

CHAPTER 7

Localisation of the central sulcus region in glioma patients with three-dimensional fluid-attenuated inversion recovery and volume rendering: comparison with functional and conventional magnetic resonance Br J Neurosurg 2011; 25(2):210-7

Ronald B. Willemse Petra J.W. Pouwels Frederik Barkhof W. Peter Vandertop

CHAPTER 7

ABSTRACT Purpose Volume rendering (VR) of three-dimensional (3D) fluid-attenuated inversion recovery (FLAIR) magnetic resonance (MR) images shows regional intensity differences, reflecting the central sulcus (CS) region and occipital cortex. The purpose of this study was to determine whether 3D FLAIR with VR could be used as an alternative method to localize the CS region in comparison with functional and conventional MR-imaging in patients with perirolandic glioma.

Methods Eleven patients with intracranial gliomas were studied with single-slab 3D FLAIR including VR and conventional T1-weighted imaging. In all patients preoperative fMRI was performed with a motor paradigm of the hand. The hypo-intense central gyri on 3D FLAIR with VR were interpreted as the CS area. Localization of the motor hand knob on anatomical images and fMRI results were used for identification of the primary motor cortex.

Results Anatomical localization of the motor hand knob on T1-weighted images was possible in 91 % of both hemispheres. In 73 % of the affected (AH) and 91 % of the unaffected (UH) hemispheres the hand knob and CS region could be identified on 3D FLAIR axial and VR images respectively. With one exception, fMRI activation confirmed the CS region as observed with 3D FLAIR with VR. Conclusions Volume rendering of 3D FLAIR MR images shows central hypo-intensities frequently corresponding with the CS region. Two-dimensional localization of the CS region on conventional T1-weighted images and fMRI seems favourable compared to 3D FLAIR. However, in selected cases, especially where fMRI is not possible or feasible, volume rendering with 3D FLAIR may enhance the 3D visualization of gliomas in relation to the CS region which can be used as an alternative method in the presurgical structural and functional evaluation of neurosurgical patients.

112

3D FLAIR VOLUME RENDERING IN GLIOMA

INTRODUCTION Structural magnetic resonance (MR) imaging is essential in the evaluation of patients with intracranial lesions, in whom a neurosurgical procedure is considered. Successful presurgical planning requires knowledge of the spatial relation between the lesion and eloquent cortex. Large interindividual differences between structural and functional neuroanatomy make it difficult to accurately assess this relation, especially in the presence of intracranial lesions with edema and brain shift (1; 2). For lesions near the central sulcus (CS) neuro-anatomical landmarks, such as gyral morphology on MR-imaging, can be used to locate the motor cortex of the hand (3), but can be unreliable in the presence of intracranial pathology (4). Direct cortical stimulation is widely accepted as the best means for identifying the primary motor cortex in humans but requires an operation and lacks the possibility of presurgical planning (5). Noninvasive preoperative functional information can be obtained with functional magnetic resonance imaging (fMRI) (6; 7), magnetoencephalography (MEG)(8; 9) or positron-emission tomography (PET) (10-12) , to locate quite reliably the primary motor and sensory cortex, which can be subsequently used for preoperative planning and intra-operative navigation. However, these methods have methodological and logistical drawbacks, are time-consuming, costly and mostly require adequate patient cooperation.

An alternative identification of the CS region is possible due to the lower signal intensities (SI) as observed on turbo fluid-attenuated inversion recovery (FLAIR) MR images in the normal population (13; 14). The FLAIR sequence is useful in detecting white matter abnormalities or subtle lesions in structures situated in the interface between the cerebrospinal fluid (CSF) and the cerebral parenchyma, because of suppression of the CSF signal. Recently, single-slab methods have been developed with T2 and FLAIR contrasts (15; 16). Using this sequence, 3D FLAIR images can be obtained with isotropic voxels. This allows reconstruction of the brain in all orientations as well as volume rendering (VR).

Volume rendering of the 3D FLAIR dataset show hypo-intense cortical regions corresponding to the CS region and occipital cortex. The purpose of our study was twofold.

First, we wanted to determine if the observed hypo-intense areas on 3D FLAIR is the result of lowered SIs or could be related to the volume-rendering process itself. Our second goal was to determine whether this single-slab 3D FLAIR MR imaging technique with VR could be used as an alternative method to localize the CS region in comparison with functional and conventional MR-imaging techniques in patients with perirolandic glioma. In case of reliable CS localization with the volume-rendered 3D FLAIR MR images, this imaging technique could be used as an alternative or complimentary method of CS localization, especially in cases where other functional imaging modalities are either not available or feasible due to patient cooperation or neurologic deficit.

113

7

CHAPTER 7

MATERIALS AND METHODS Patients From the patients with intracranial lesions, referred to the department of Neurosurgery of the VU University Medical Centre (Amsterdam, The Netherlands), we retrospectively selected eleven consecutive patients with perirolandic gliomas who had had surgery (four female and seven male, age range 35 - 63 years; mean: 45.0 years), a pre-operative Karnofsky Performance Status score â&#x2030;Ľ 70, 3D FLAIR and fMRI data available.

Table 1 summarizes the clinical data of the patient group. Seven patients had an astrocytoma (four WHO grade II, three WHO grade III), one had an oligo-astrocytoma (WHO grade II) and three had an oligodendroglioma (one WHO grade II, two WHO grade III). The lesions were located in the right hemisphere in seven cases and in the left hemisphere in four cases. Ten patients had seizures and one (patient 6) had a neglect preoperatively. All patients had conventional T1-weighted imaging and fMRI data had been acquired in the context of a prospective functional imaging study. Functional MRI was performed with a motor paradigm of the contra- and ipsilateral hand in six cases. In five cases motor activation was only performed with the hand contralateral to the lesion.

Intraoperative mapping of the CS was performed in 8 out of 11 patients. Postoperatively, two patients had a transient dysphasia and paresis, one patient had persistent dysphasia.

The study was approved by the Medical Ethics Committee of the VU University Medical Centre and informed consent from the participants was obtained prior to inclusion.

Table 1. Summary of clinical characteristics.

Patient

Age / Gender

Diagnosis

Location, Lateralization

1

48 / M

astrocytoma, II

parietal, R

2

40 / F

oligodendroglioma, III

frontotemporal, L

3

43 / M

astrocytoma, III

frontotemporal, R

L

4

63 / M

astrocytoma, III

frontal, L

R

5

61 / F

astrocytoma,II

frontal, R

L

6

43 / M

oligodendroglioma, III

parietal, R

L+R

7

41 / M

oligo-astrocytoma, II

parietal, R

L+R

8

35 / M

astrocytoma, II

parietal, R

L+R

9

48 / F

oligodendroglioma, II

parietal, L

L+R

10

38 / M

astrocytoma, III

temporal, L

R

11

35 / F

astrocytoma, II

parietal, R

L+R

M, male; F, female; L, left; R, right

114

Motor paradigm hand L+R R

3D FLAIR VOLUME RENDERING IN GLIOMA

MR acquisition Imaging was performed with a 1.5 T MR scanner (Siemens Sonata, Erlangen, Germany).

A single-slab 3D FLAIR MR sequence was used(15; 16), with slightly different parameters to obtain nearly isotropic voxels (17). Repetition time TR = 6500 ms, inversion time TI = 2200 ms, 191 echoes, with effective echotime TE = 355 ms. Whole brain coverage was obtained with a sagittal 3D slab of 156 â&#x20AC;&#x201C; 166 mm, consisting of 120 â&#x20AC;&#x201C; 128 partitions of 1.3 mm. Using a 190 x 256 matrix and 230 x 310 mm field-of-view, the in-plane resolution was 1.21 x 1.21 mm. No interpolation was used either in-plane or in the slab direction.

Employing 75% partial Fourier in the slab direction, the acquisition time varied between 9 min 47 s and 10 min 15 s.

Functional MR imaging was performed using the blood oxygenation level dependent (BOLD) technique (echo-planar imaging, TR = 3000 ms, TE = 60 ms, 25 axial slices were obtained with slice thickness of 3 mm and 10 % gap, a 64 x 64 matrix and 211 x 211 mm field-of-view resulting in isotropic 3.3 x 3.3 x 3.3 mm3 voxels). The motor paradigm consisted of self-paced clenching and opening of the hand during 15 seconds alternated with rest periods of 15 sec, for a total of 10 epochs. Conventional 3D T1 (inversion recovery fast gradient echo) images (TR = 2700 ms, TE = 5 ms, TI = 950 ms, 176 coronal 1.5 mm slices with 1.0 x 1.0 mm2 in-plane resolution) were obtained for anatomical labelling.

Post processing Post processing and visualisation software (Siemens, Erlangen, Germany) was used to make VR images of the 3D FLAIR dataset. The visual properties of the volume images can be changed by optimizing the transparency, brightness and signal intensity. The result is a freely rotatable 3D volume image with three additional orthogonal cutting planes available. Modified parameters can be stored to be used for subsequent evaluations and individual adjustments of contrast.

After motion-correction of the fMRI volumes, activated areas were calculated with a t-test taking into account the delay of the hemodynamic response function. Voxels exceeding a t-value of 3.0 were colour-coded and considered as activated areas. The functional data-set was then transferred to the VR images to examine the activated areas in their anatomical context (Figure 1a).

Signal Intensities The signal-to-noise ratio (SNR) of the 3D FLAIR images of all subjects was calculated in three cortical regions of interest (ROI), the CS-region, the parietal and occipital cortex.

The ROIs were placed on a paramedian sagittal image with a visible marginal ramus of the cingulate fissure and parieto-occipital fissure in the unaffected hemisphere. The cortex anterior of the fissure was taken for the CS-region, the parietal cortex was taken between the cingulate fissure and the parieto-occipital fissure and the occipital cortex posterior from the parieto-occipital fissure. The SI of the cortex was obtained from the three locations described above. The standard deviation (SD) of the noise was obtained from a ROI outside and superior of the skull. The SNR was calculated as SIROI / 0.654 SD noise 115

7

CHAPTER 7

taking into account a correction for underestimation of the noise in magnitude images (18). The SNR of the three regions were compared and tested statistically, using the paired t-test. Statistical significance was determined at an alpha level of 0.05 and data are presented as means Âą SD. Identification of the central sulcus region The anatomical hand knob localization of the CS was identified on reformatted axial slices of the 3D T1 (Figure 1b) and the 3D FLAIR scan. Without the functional data available, an experienced neuroradiologist (F.B.) identified the presence of the hand knob on the posterior border of the precentral gyrus in both hemispheres. The hand knob was only identified in case of a typical inverted omega- or horizontal epsilon shaped cortex. Failure to identify the hand knob could be related to the presence of tumour or variations in the course of one of the fissures. Functional MR localization of the motor hand area was judged positive when a clear area of activation was found in the central region contralateral to the hand movement. The CS region was identified on 3D FLAIR with VR in case of regional hypo-intensity in the central area. Figure 2 demonstrates an example of 3D FLAIR with VR from a superior view in two patients. Figure 2a shows bilateral lowered SI in the CS region, whereas Figure 2b demonstrates only unilateral lowered SI around the CS region. The affected hemisphere lacks the presence of a visible CS region.

Figure 1. A: Right-sided postero-lateral view of 3D FLAIR MR imaging with volume rendering of patient 8. Low signal intensities are visible in the area of the CS and the occipital cortex (arrowheads). Superimposed fMRI data (red) showing left hand motor activation in the CS region with two orthogonal planes and a right-sided parietal astrocytoma, WHO grade II (white). B: axial reformatted 3D T1 gadolinium-enhanced MR image with arrowheads pointing to the hand knob (L = left; R = right) in the same patient.

116

3D FLAIR VOLUME RENDERING IN GLIOMA

7 Figure 2. Superior view of 3D FLAIR MR images with volume rendering of patient 4 (A) and 9 (B) with arrowheads pointing at the CS. A: Lowered signal intensities are bilaterally visible at the CS region. B: unilateral visible CS region at the unaffected side due to lower signal intensities, while the left hemisphere with a postcentrally located low-grade glioma lacks the presence of a hypo-intense CS region (P = posterior; L = left; R = right).

RESULTS In 3D FLAIR images, the SNR of both the CS region (64.2 ± 9.9) and the occipital cortex (60.1 ± 8.0) was significantly lower (p < 0.01) than the SNR of the parietal cortex (72.9 ± 10.1). No significant difference in SNR was found between the CS region and the occipital cortex. The correspondence between the presence of the motor hand knob on anatomical axial slices, fMRI activation and the presence of a visible CS region on 3D FLAIR is shown in Table 2. Anatomical identification of the hand knob on axial reformatted 3D T1 images was possible in 10 out of 11 (91 %) of both the affected and unaffected hemisphere. The two failures (patient 2 and 9) were due to anatomical variance and not related to tumour presence. Hand knob identification on axial reformatted 3D FLAIR images was possible in 8 out of 11 (73 %) of the affected and in 10 out of 11 (91 %) of the unaffected hemispheres. Two failures were due to the presence of tumour (patient 6 and 11), the other two were identical to the failures of the 3D T1 scan.

The fMRI results were successful in 10 out of 11 (91 %) of the affected hemispheres and in 5 out of 6 (83 %) of the unaffected hemispheres. The two failures were in the same patient (patient 6) due to large tumour mass on the affected side and inadequate

117

CHAPTER 7

performance of the motor task for the unaffected hemisphere. The 3D FLAIR VR show that the CS region could be visualized in 8 out of 11 (73 %) of the affected and 10 out of 11 (91 %) of the unaffected hemispheres. Failures in affected hemispheres were due to tumour presence.

In all twelve hemipheres with a visible CS region and corresponding fMRI data available, the fMRI results were located in the hypo-intense CS region, except for one fMRI due to inadequate patient performance. In three of the AHs and one of the UH the CS region was not visible on VR images, while fMRI was successful in three of the corresponding hemispheres. The absence of CS hypo-intensity in combination with posterior tumour invasion and succesful fMRI is demonstrated in the VR image of Figure 3a with corresponding sagittal projection of the FLAIR image. The unaffected hemisphere shows a clear hypo-intense CS in combination with successful fMRI (Figure 3b).

In five cases bilateral anatomical (hand knob) and functional localization (fMRI) was possible in combination with a visible CS region on 3D FLAIR with VR.

Table 2. Correspondence of motor hand knob presence on axial MR images (3D T1 and 3D FLAIR), fMRI activation in motor hand area and hypo-intense CS region on 3D FLAIR with volume rendering (VR).

Unaffected Hemisphere Patient

Affected Hemisphere

hand knob 3D T1

hand knob FLAIR

fMRI

3D FLAIR VR

hand knob 3D T1

hand knob FLAIR

fMRI

3D FLAIR VR

Symptoms

Intraoperative Monitoring CS

1

+

+

+

+

+

+

+

+

seizures

+

2

+

+

not done

+

-*

-*

+

+

seizures

not performed

3

+

+

not done

+

+

+

+

+

seizures

+

4

+

+

not done

+

+

+

+

+

seizures

not performed

5

+

+

not done

+

+

+

+

+

seizures

+

6

+

+

-Âś

+

+

-#

-#

-#

neglect

+

7

+

+

+

-$

+

+

+

+

seizures

+ +

8

+

+

+

+

+

+

+

-#

seizures

9

-*

-*

+

+

+

+

+

-#

seizures

+

10

+

+

not done

+

+

+

+

+

seizures

not performed

11

+

+

+

+

+

-#

+

+

seizures

+

10/11

10/11

10/11

10/11

TOTAL

5/6

8/11

10/11

8/11

FLAIR, fluid attenuated inversion recovery; +, presence; -, absence; CS, central sulcus. Cause of failure due to: * anatomical variance; # tumour mass; Âś inadequate performance; $ presumably structural cortical characteristics.

118

3D FLAIR VOLUME RENDERING IN GLIOMA

Figure 3. A: 3D FLAIR MR image with volume rendering from patient 9, viewed from the left posterolateral side with fMRI overlay (red) , showing no signal intensity differences in the CS-region, but with successful motor hand localization using fMRI. The projection of the sagittal FLAIR image shows posterior tumour invasion of the CS. B: View from the right posterolateral side shows a clear hypo-intense CS region in combination with corresponding fMRI overlay (red).

7 DISCUSSION We demonstrated an isotropic single-slab 3D FLAIR MR imaging technique of the brain, which is suitable for VR. The images have no signal from CSF and low signal from skull and skin, which allows a reconstruction of the cortical surface without pre-processing steps like skull-stripping. We have shown that VR of the 3D FLAIR dataset has lowered SIs in the CS region and the occipital cortex and can be used to visualize the CS region in the large majority of cases. The data from fMRI confirmed the visible area of the CS in most cases, which offers new possibilities to use structural MR imaging with functional characteristics.

Hypo-intensity in 3D FLAIR volume rendering The regional intensity differences in the VR images reflect differences in T1 relaxation time and have been described earlier for two-dimensional (2D) FLAIR, showing the lowest signal intensities along the CS (13; 14; 19) . In our study, we also found significant lower SIs in the CS region and the occipital cortex as measured on sagittal 3D FLAIR images.

Therefore, the regional hypo-intensity as seen on the VR images is an objective finding and seems not related to the VR process itself. The regional variations are possibly related to structural differences in cortical areas (20). Classic post-mortem studies have shown regional variability in cortical thickness across the cortex, with thickness varying between 1 and 4.5 mm (21). The most variable areas of the cortex are the pre- and post-central gyri, the primary visual areas and the anterior medial temporal lobes (22), reflecting

119

CHAPTER 7

different proportions of grey matter. Fischl et al. have used an automated method to accurately measure cortical thickness across the entire brain and demonstrated that the CS region and the visual cortex have the lowest average cortical thickness, which could be an explanation for the observed hypo-intensities on VR images (23). Others have used measurements of cortical thickness ratios on T2-weighted images in patients with brain tumours and vasogenic edema and showed a significant difference between the CS versus both frontal and parietal sulci (24). However, probably cortical thickness properties are not the only explanation for lower signal intensities. Cytoarchitectural factors in the different cortical layers influence relaxation time and other MR properties as well (25). In three hemispheres at the affected side the VR images failed to show a local hypo-intense CS region, possibly related to nearby tumour location with edema or tumour infiltration (as demonstrated on the sagittal FLAIR projection in Figure 3a). Remarkably, in one hemisphere the VR image did not show a rolandic hypo-intensity at the unaffected side, while it was present at the affected side. This finding seems not related to motion artefacts or other pathology and might be explained by cortical properties.

Functional localization with 3D FLAIR volume rendering Intra-operative cortical stimulation is still considered the standard method to localize language and motor function in neurosurgical procedures. To date, an increasing number of non-invasive imaging modalities are used for preoperative functional localization.

Gyral morphology on structural MR images can be used to localize the CS region. Yousry et al. described the motor cortex hand knob as a potential neuro-anatomical landmark (3). The motor hand knob has a characteristic inverted omega or horizontal epsilon shape on axial anatomical images. Despite the presence of intracranial lesions, identification of the motor hand knob on axial reformatted 3D T1 images was still possible in the majority of cases and failures were only due to anatomical variance in the course of one of the fissures. Hand knob identification on axial reformatted FLAIR images was similar for the UHs. In the AHs two extra cases could not be identified due to lower resolution of the FLAIR scan in combination with tumour presence. In our series, identification of the CS region on VR images was possible in 73 % of the AHs, which is comparable with motor hand knob identification on axial FLAIR images. While hand knob identification on the 3D T1 sequence seems to be favourable compared to 3D FLAIR, the latter sequence has the advantage of 3D whole-brain visualization not only showing the tumour but also its relation to the CS in a freely rotatable manner.

Others have used a brain surface reformatted imaging technique based on isotropic T1weighted 3D data as an accurate and reliable technique to visualize superficial brain lesions in relation to the precentral gyrus. This technique may also provide useful additional information for preoperative surgical planning, but it only shows a part of the brain surface. In addition, anatomical variations in the CS region may lead to false localization (26).

Progress in non-invasive functional imaging modalities like fMRI, MEG and PET has made it possible to localize functions in the human brain. Functional MRI is the most widely

120

3D FLAIR VOLUME RENDERING IN GLIOMA

used modality and has shown successful localization of the motor hand area in more than 80 % of the cases (27-30). Others have shown that cerebral structure-function correlation is possible with special structural MR techniques to localize different functional areas of the cortex (20; 24; 31) , but these techniques are not very useful in daily neurosurgical practice.

Advantages of 3D FLAIR volume rendering Conventional 2D imaging is not always sufficient to perceive the true spatial relationships between anatomical structures. Therefore, 3D datasets can be used to select images from any plane and visualize the lesion from any direction for better understanding. However, true 3D visualization can only be acquired with VR methods, showing realistic-looking brain surface images. Volume rendering of 3D FLAIR images can be performed without pre-processing steps like skull stripping which increases the possible use in daily routine.

The VR image gives a good view of the localization of the tumour in the case of cortical extension, which can be viewed in all directions. In most cases the spatial relationship to the CS or occipital area can be viewed as well and by adding three orthogonal planes, the extension of the lesion in the white matter can easily be visualized simultaneously. Most unenhancing lesions are better visualized on T2-weighted or FLAIR sequences than on T1 sequences. Therefore, this technique is especially suitable for low-grade gliomas to assess the relation to the CS. Identification of the sensorimotor cortex with 3D FLAIR renderings seems to be comparable with hand knob identification but has a lower success rate than the results of fMRI in our patient group. However, functional MRI with a motor paradigm, requires good patient cooperation and no severe motor deficits, conditions not necessary for 3D FLAIR imaging.

Furthermore, the 3D FLAIR images can easily be imported into a neuronavigational system and automatically be fused with other modalities. In this way a registration accuracy as well as high-resolution reconstruction in any direction is possible.

Limitations of 3D FLAIR volume rendering Because of differences in signal intensity between sessions and subjects, there is no single solution for optimal parameter settings to achieve the best VR. On an individual basis, the stored parameters sometimes need slight modifications in opacity and brightness. In the presence of edema, a strong contrast-enhancing lesion or a very large tumour with compression of the CS region the results can be difficult, making additional functional imaging necessary.

Because of signal intensity differences between the parietal lobe and the CS region, the CS region is frequently visible on 3D FLAIR, however delineation of the CS can be difficult because of lack of contrast between the pre- and post-central gyrus. Furthermore, in some cases there is a prefrontal extension of the hypo-intense area, which is not apparent in the parietal area. Therefore, a better delineation between the parietal lobe and the CS region is seen than between the premotor area and the CS region. This makes 3D FLAIR renderings more useful for parietal than premotor lesions.

121

7

CHAPTER 7

Using a qualitative dependant variable (visibility of the CS region) requires further analysis with inter-observer agreement. However, our first objective was to evaluate the presence of the hypo-intense rolandic zone in AH and UH, found in 3D FLAIR VR for comparison with anatomical and functional imaging techniques.

Finally, the concordance between VR images and fMRI was only studied for motor hand function, suggesting that the visible area on these images corresponds to the CS region.

While this may be true for the motor cortex of the hand, we did not investigate the localization of other functions in the CS region such as sensory function or foot and tongue movements.

Implications for the future Information from 3D FLAIR VR can be useful in the preoperative evaluation of patients with intracranial gliomas to assess the spatial relationship of the lesion to the CS region and may support additional functional imaging investigations. However, due to the preliminary character of this study, additional prospective studies should be performed to show the usefulness of 3D FLAIR with VR in the presurgical evaluation of neurosurgical patients. At the same time, the use of multi-channel phased-array coils and parallel imaging techniques will lead to an increased spatial resolution of 3D FLAIR images in similar or shorter acquisition times. This will presumably improve localization of the CS region.

CONCLUSIONS We demonstrated an isotropic 3D FLAIR imaging sequence with VR to visualize important cortical areas in relation to intracranial gliomas. With this new non-invasive structural MR technique it seems possible to identify functional areas corresponding with the pre- and post-central cortex and possibly the primary visual cortex. Two-dimensional localization of the CS region on conventional axial T1-weighted images and fMRI seem favourable compared to 3D FLAIR with VR. In selected cases, especially where fMRI is not possible or feasible, VR with 3D FLAIR may enhance the 3D visualization of gliomas in relation to the CS region which can be used as an alternative method in the presurgical structural and functional evaluation of neurosurgical patients.

122

3D FLAIR VOLUME RENDERING IN GLIOMA

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Rademacher J, Bürgel U, Geyer S, Schormann T, Schleicher A, Freund HJ, and Zilles K. Variability and asymmetry in the human precentral motor system. A cytoarchitectonic and myeloarchitectonic brain mapping study. Brain. 2001, Nov;124(Pt 11):2232-58.

Branco DM, Coelho TM, Branco BM, Schmidt L, Calcagnotto ME, Portuguez M, et al. Functional variability of the human cortical motor map: electrical stimulation findings in perirolandic epilepsy surgery. J Clin Neurophysiol. 2003, Feb;20(1):17-25.

Yousry TA, Schmid UD, Alkadhi H, Schmidt D, Peraud A, Buettner A, and Winkler P. Localization of the motor hand area to a knob on the precentral gyrus. A new landmark. Brain. 1997, Jan;120 ( Pt 1)141-57.

Sobel DF, Gallen CC, Schwartz BJ, Waltz TA, Copeland B, Yamada S, et al. Locating the central sulcus: comparison of MR anatomic and magnetoencephalographic functional methods. AJNR Am J Neuroradiol. 1993;14(4):915-25.

Penfield W., and Boldrey E. Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain. 1937;60389-443.

Hammeke TA, Yetkin FZ, Mueller WM, Morris GL, Haughton VM, Rao SM, and Binder JR. Functional magnetic resonance imaging of somatosensory stimulation. Neurosurgery. 1994, Oct;35(4):677-81.

Rao SM, Binder JR, Hammeke TA, Bandettini PA, Bobholz JA, Frost JA, et al. Somatotopic mapping of the human primary motor cortex with functional magnetic resonance imaging. Neurology. 1995, May;45(5):919-24.

Hari, R. Human cortical functions revealed by magnetoencephalography. Progr Brain. 1994;100163-168.

Gallen CC, Schwartz BJ, Bucholz RD, Malik G, Barkley GL, Smith J, et al. Presurgical localization of functional cortex using magnetic source imaging. J Neurosurg. 1995, Jun;82(6):988-94.

Fox PT, Burton H, and Raichle ME. Mapping human somatosensory cortex with positron emission tomography. J Neurosurg.

1987, Jul;67(1):34-43.

Bittar RG, Olivier A, Sadikot AF, Andermann F, Pike GB, and Reutens DC. Presurgical motor and somatosensory cortex mapping with functional magnetic resonance imaging and positron emission tomography. J Neurosurg. 1999, Dec;91(6):915-21.

Nyberg G, Andersson J, Antoni G, Lilja A, Pellettieri L, Valind S, and Långström B. Activation PET scanning in pretreatment evaluation of patients with cerebral tumours or vascular lesions in or close to the sensorimotor cortex. Acta Neurochir (Wien). 1996;138(6):684-94.

Karaarslan E, and Arslan A. Perirolandic cortex of the normal brain: low signal intensity on turbo FLAIR MR images.

Radiology. 2003;227(2)538-541.

Karaarslan E, and Arslan A. ROI measurement of the signal intensity of precentral cortex in the normal brain. Eur J Radiol.

2004, Dec;52(3):221-3.

Kallmes DF, Hui FK, and Mugler JP. Suppression of cerebrospinal fluid and blood flow artifacts in FLAIR MR imaging with a single-slab three-dimensional pulse sequence: initial experience. Radiology. 2001, Oct;221(1):251-5.

Zhao L, Mugler J, Wei X, Chen N, Mulkern R, and Panych L. A high-resolution clinical whole-brain scan using single-slab three dimensional T1W, T2W and FLAIR fast spin-echo sequences. Proc. 10th ISMRM. 2002;1294.

Moraal B, Roosendaal SD, Pouwels PJ, Vrenken H, van Schijndel RA, Meier DS, et al. Multi-contrast, isotropic, single-slab 3D MR imaging in multiple sclerosis. Eur Radiol. 2008, Oct;18(10):2311-20.

Henkelman RM. Measurement of signal intensities in the presence of noise in MR images. Med Phys. 1985;12(2):232-3.

Hirai T, Korogi Y, Yoshizumi K, Shigematsu Y, Sugahara T, and Takahashi M. Limbic lobe of the human brain: evaluation with turbo fluid-attenuated inversion-recovery MR imaging. Radiology. 2000, May;215(2):470-5.

Bendersky M, Rugilo C, Kochen S, Schuster G, and Sica RE. Magnetic resonance imaging identifies cytoarchitectonic subtypes of the normal human cerebral cortex. J Neurol Sci. 2003, Jul 15;211(1-2):75-80. von Economo C, and Koskinas G. Die Cytoarchitektonik der Hirnrinde des erwachsenen Menschen. Berlin: Springer Verlag; 1925.

Lerch JP, and Evans AC. Cortical thickness analysis examined through power analysis and a population simulation.

Neuroimage. 2005, Jan 1;24(1):163-73.

Fischl B, and Dale AM. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A. 2000, Sep 26;97(20):11050-5.

Biega TJ, Lonser RR, and Butman JA. Differential cortical thickness across the central sulcus: a method for identifying the central sulcus in the presence of mass effect and vasogenic edema. AJNR Am J Neuroradiol. 2006, Aug;27(7):1450-3.

Luders E, Narr KL, Thompson PM, Woods RP, Rex DE, Jancke L, et al. Mapping cortical gray matter in the young adult brain: effects of gender. Neuroimage. 2005, Jun;26(2):493-501.

Hattingen E, Good C, Weidauer S, Herminghaus S, Raab P, Marquardt G, et al. Brain surface reformatted images for fast and easy localization of perirolandic lesions. J Neurosurg. 2005, Feb;102(2):302-10.

123

7

CHAPTER 7

27. Pujol J, Conesa G, Deus J, Lรณpez-Obarrio L, Isamat F, and Capdevila A. Clinical application of functional magnetic resonance imaging in presurgical identification of the central sulcus. J Neurosurg. 1998, May;88(5):863-9.

28. Fandino J, Kollias SS, Wieser HG, Valavanis A, and Yonekawa Y. Intraoperative validation of functional magnetic resonance imaging and cortical reorganization patterns in patients with brain tumors involving the primary motor cortex. J Neurosurg.

1999, Aug;91(2):238-50.

29. Krings T, Reinges MH, Erberich S, Kemeny S, Rohde V, Spetzger U, et al. Functional MRI for presurgical planning: problems, artefacts, and solution strategies. J Neurol Neurosurg Psychiatry. 2001, Jun;70(6):749-60.

30. Vlieger EJ, Majoie CB, Leenstra S, and Den Heeten GJ. Functional magnetic resonance imaging for neurosurgical planning in neurooncology. Eur Radiol. 2004, Jul;14(7):1143-53.

31. Walters NB, Egan GF, Kril JJ, Kean M, Waley P, Jenkinson M, and Watson JD. In vivo identification of human cortical areas using high-resolution MRI: an approach to cerebral structure-function correlation. Proc Natl Acad Sci U S A. 2003, Mar 4;100(5):2981-6.

124

125

CHAPTER 8

General discussion and future perspectives

CHAPTER 8

Originally, this project started with evaluating the correlation of presurgical localization procedures, using MEG and fMRI, with intra-operative findings using electrical cortical stimulation (ECS), the gold standard. However, because the incorporation of functional data into the neuronavigation system initially failed, our research focus shifted towards the analysis of the influence of structural brain lesions on the activity and localization of the sensory and motor cortex, with special emphasis on the foot area. The main part of this discussion will therefore focus on these results. Multimodal imaging and structural imaging aspects are shortly discussed afterwards.

Intracranial lesions and sensorimotor MEG Somatosensory cortex localization The main indication for somatosensory mapping using SEFs is localization of S1 and its spatial relationship to a mass lesion or epileptogenic zone (8,9,36). Usually, the information from the early-latency (electrical stimulation) and middle-latency (tactile stimulation) SEF response is used. The validity of the SEF source locations has been compared with intraoperative SEP recordings and/or electrical stimulation mapping and was found to be very reliable in brain tumor patients (9,37,40). Most of these studies used the early-component (N20m) of the median nerve SEF to localize the contralateral S1. Comparable to previous studies in brain tumor patients, we also found robust S1 localization after MN stimulation in a large patient group with different types of pathology.

The results of lower limb SEFs are limited and controversial, especially in the presence of intracranial lesions. The first recordings of SEFs after lower limb stimulation (10) confirmed the source location in the medial wall of the hemisphere with field patterns rotating as a function of time (11,16) . The origins of the evoked potentials and fields were controversial (47), only recently it was confirmed in healthy subjects that the postero-anterior component after PTN stimulation is analogous to N20m in the median nerve using MEG (33).

We used PTN SEFs to investigate changes in the cortical response to sensory stimulation in patients with unilateral intracranial lesions (Chapter 2). The early-component (N37m) was not identified, which could be related to an insufficient dipolar pattern around 37 ms to calculate a dipole with a residual error < 25 % or differences in the experimental settings. However, the middle- and late components were identified and not affected by the presence of intracranial lesions and we advocate the use of these components as well to study the relationship between the localization of the lower limb S1 and the lesion.

Motor cortex localization Motor mapping using MEG can be performed with spatial filtering techniques, which has been successfully used to detect time-locked increases and decreases in the mu (8 â&#x20AC;&#x201C; 12 Hz) and beta (15 â&#x20AC;&#x201C; 30 Hz) frequency bands in the vicinity of the central sulcus during voluntary hand movements (17,45). This technique has also been compared with intraoperative mapping and was found to be an accurate method for functional localization (7,32). Displacement of the sensorimotor cortex due to a lesion has no influence on correct identification of the motor cortex (7). We also found hand motor function in the expected

128

GENERAL DISCUSSION AND FUTURE PERSPECTIVES

contralateral motor cortex, not only in brain tumor patients but also in patients with different kinds of pathology (Chapter 3 and 5).

While motor hand localization using spatial filtering techniques seems reliable and unaffected by the presence of lesions, we did find significant changes in the spectral content of activity in resting state and during movement, in the presence of glioma (Chapter 4). The shift in spectral power towards lower frequencies for M1 in the affected hemisphere, refers to so-called ‘slowing’ of brain oscillations. Others found slowing of M1 oscillations related to impairment of motor function in Parkinson’s disease (49).

However, our patients had no motor deficit and mostly were slow-growing low-grade gliomas. Possibly, neural plasticity due to tumor invasion may cause the observed changes without affecting motor functioning yet (5).

Most motor MEG studies were performed to evaluate upper limb function, however localization of the foot area in the pre-operative setting in neurosurgical patients is uniquely important. First, iatrogenic damage to the foot motor cortex and the resultant paresis of the leg can render a patient wheelchair- or bed-bound, which may be more debilitating than paresis of the non-dominant hand or arm. Second, the localization of the foot motor cortex lacks a discernible anatomic landmark such as the “motor hand knob” for the hand motor area. Third, the foot motor cortex is located under the sagittal sinus along the medial wall of the hemisphere, making its localization difficult to confirm by intraoperative direct cortical stimulation. Due to the importance of foot motor function preservation and the difficulties with identifying the foot area in combination with a lack of brain mapping results from the lower limb in neurosurgical candidates, we described the results of lower limb stimulation with MEG (Chapter 2, 3 and 4). Repetitive foot movements can be used as a surrogate for gait, as the neural processes associated with these movements can resemble those during walking (31). Both cortical and subcortical structures are involved in cerebral motor control of rhythmic foot movements as shown by studies with fMRI (3,38). Foot movements studied with MEG showed altered patterns of cortical activity in the presence of pathology. Bilateral and ipsilateral activation was observed with foot movements. Bilateral activation of motor areas has been described earlier with different modalities in both normal subjects, stroke patients and patients with intracranial lesions (2,19,20,24,32,48) . Participation of the ipsilateral motor cortex could be explained by the partially uncrossed fibers of the corticospinal tract in combination with interhemispheric interactions (35,51).

In a small fraction of patients we found exclusive ipsilateral motor responses, especially with foot movements, which could be the result of functional reorganization in the presence of intracranial pathology. The use of MEG in documenting brain plasticity for sensorimotor functions in patients with intracranial lesions has been demonstrated in few studies (22,30,32,44). Some authors suggest the recruitment of the ipsilateral motor cortex in the presence of glioma (44), while others suggested that it represents a normal variant in the population (7) .

The majority of patients with ipsilateral motor responses had a history of stroke, traumatic brain injury (TBI) or a developmental disorder. In patients with stroke ipsilateral responses

129

8

CHAPTER 8

could be explained by disruption of interhemispheric inhibition and in TBI patients by damage to the corpus callosum (2,43). Patients with developmental disorders have a well-known capacity to compensate for lesions acquired in the pre- and perinatal period, including ipsilateral takeover of motor functions (42). Yet, the exact mechanisms underlying functional plasticity in brain tumors remain unknown.

Multimodal imaging using MEG and fMRI – the more the better?

The tremendous advances in functional neuroimaging make detection and localization of almost every brain activity possible, however, in neurosurgical practice the awake-craniotomy with ECS is still considered the gold standard for localization of eloquent brain areas.

It is clear that MEG and fMRI differ in many ways with respect to the detection of evoked responses. Despite these (complementary) differences, it may be useful to combine both techniques for the most complete and reliable characterization of functional anatomy.

A few studies compared activation of the sensorimotor cortex using fMRI and MEG (4,21,22,41) with average differences in spatial localization between 10 – 15 mm., comparable to our study (Chapter 6). Despite methodological and interpretational limitations, the main problem is that the presence of these spatial differences does not help in the decision which method is best suited for localization of the sensorimotor cortex. At this moment, S1 localization is ideal for MEG. The reliability has been demonstrated in many studies, the SEFs have a good signal-to-noise ratio and the first cortical component (N20m) has a clearly defined dipolar field pattern not overlapping with activity in other areas (9,18,22). As a consequence, with reliable S1 identification, M1 can easily be inferred.

Functional MRI has shown good agreement with intraoperative ECS mapping, but has some methodological limitations with respect to M1 cortex mapping in patients with brain lesions (6,23,25). Alterations in neurovascular coupling in the presence of brain lesions and discrimination difficulties between different brain areas due to the limited temporal resolution are the two most important factors to take into account (14,22). Despite the limitations of fMRI, its widespread availability still makes it the method of choice for functional localization.

However, whatever preoperative functional information is used and transferred to a neuronavigation system, there is still uncertainty during surgery about the exact localization of eloquent brain areas, since it does not account for brain deformation or ‘brain shift’ during surgery. After the craniotomy and opening of the dura, various combined factors contribute to brain-shift, such as: cerebrospinal fluid (CSF) leakage, gravity, edema, tissue resection, retraction and the administration of osmotic diuretics (34). Intra-operative MRI (i-MRI) could be used to correct for brain shift, but lacks functional information and usually has reduced image-quality. Recently, the possibility of an awake intraoperative functional MRI (ai-fMRI) has been described. The main advantage would be to avoid the problem of seizure induction due to electrical cortical stimulation and extensive cortical exposure during conventional awake-craniotomies (27).

130

GENERAL DISCUSSION AND FUTURE PERSPECTIVES

However, there are still many practical and technical problems to make ai-fMRI a useful procedure in everyday practice. Despite the intra-operative limitations of pre-operative functional imaging data, the functional data can be used to reduce the number of locations to be tested during ECS and speed up the procedure during awake-craniotomies.

Structural imaging â&#x20AC;&#x201C; is functional imaging necessary?

Anatomical landmarks such as gyral morphology can be used to determine the motor cortex of the hand (50). Others used brain surface reformatting techniques for reliable identification of the CS (12). However, in the presence of intracranial pathology, especially with compression of the sensorimotor cortex, identification can be difficult or even impossible. In Chapter 7 we described an alternative structural imaging technique based on the lower signal intensities (SI) as observed on turbo fluid-attenuated inversion recovery (FLAIR) images. Especially, the volume-rendered 3D FLAIR MR images show regional hypo-intensities at the CS region. Although the results are comparable with hand knob identification, it is less successful than fMRI. Still, the method could be useful, especially in cases where functional imaging is not possible or feasible. The technique could be improved by the use of multi-channel phased-array coils and parallel imaging techniques to increase spatial resolution. Probably, ultra-high field (7T) MR imaging with high spatial resolution in combination with a high signal-to-noise ratio (SNR) may better delineate cortical structural differences (39).

Diffusion tensor imaging (DTI) is also a structural MR-imaging technique, able to visualize fiber tracts in the brain with the potential to identify functional tracts such e.g. the corticospinal tract (CST). Unfortunately, DTI has a poor sensitivity to delineate motor pathways, especially in the presence of pathology and therefore DTI must be used with caution and only as adjunctive data to established methods of motor mapping (28).

However, despite advances in structural imaging, considerable inter-individual variation in the exact properties of the sensorimotor cortex and the possible influence of brain pathology on functional organization, renders functional imaging still necessary in the preoperative evaluation of patients with lesions around the central sulcus.

Methodological considerations Functional brain mapping using MEG and fMRI has several methodological challenges.

Especially in MEG, the process of source localization, i.e. the determination of the spatial location of brain sources (or â&#x20AC;&#x2DC;generatorsâ&#x20AC;&#x2122;) responsible for the measured magnetic signals is complex. Estimating the generators that account for the recorded magnetic field is mathematically known as the inverse problem, meaning that there is no unique solution because more quantities are unknown than known. This can be handled by choosing modeling assumptions, to reduce the number of unknowns. However, different models therefore give different results. For clinical MEG, specific protocols for data acquisition, data processing and data analysis must be used which have been shown to produce reliable (and reproducible) results.

131

8

CHAPTER 8

The beamformer approach used in Chapter 3, 4 and 5 is another source model with special considerations. The spatial resolution varies across the brain, depending on location of the neuronal activity, orientation of the cortex and signal-to-noise ratio (1,13). The orientation and location of the sensorimotor foot area at the medial wall probably cause a lower spatial resolution, which has to be taken into account in the analysis of foot data.

Functional MRI, despite major advantages has some disadvantages as well. The most important one is that the BOLD signal is an indirect measure of neuronal activity with limited temporal resolution due to the â&#x20AC;&#x2DC;slowâ&#x20AC;&#x2122; hemodynamic response. Secondly, the BOLD signals can be affected due to aberrant tumor neovascularity (15).

Future perspectives This thesis combined structural and functional aspects of the primary sensorimotor cortex using MEG and fMRI. The worldwide application of fMRI and to a (much) lesser extent MEG in the preoperative localization of sensorimotor cortex has been shown to accurately identify the sensorimotor cortex, as has been demonstrated in several studies (21,22,26,32,44). Despite FDA approval of both techniques, the published evidence on the clinical utility of MEG and fMRI is suboptimal with respect to sensorimotor cortex identification. At this moment there are no clinical trials demonstrating utility. The literature on diagnostic accuracy has methodological limitations, mostly selection bias, which makes it impossible to draw firm conclusions whether pre-operative functional imaging may lead to better post-operative outcomes. This would require a control group of patients having surgery without pre-operative functional data available. It is questionable whether this will ever happen.

An important feature of MEG and fMRI is its value in documenting brain plasticity for sensorimotor functions. This has already been done in stroke patients (46) but can also be valuable in the pre-surgical workup of brain tumor patients or to monitor recovery following surgery with postoperative deficits.

We demonstrated the use of MEG in a large cohort of patients, showing brain plasticity or functional reorganization in the presence of lesions and/or epilepsy. An interesting phenomenon was the presence of ipsilateral motor activity. Ipsilateral motor activity cannot be evaluated under surgical conditions, however navigated transcranial magnetic stimulation (nTMS) may be considered as an alternative. In essence, it is the reverse of MEG, instead of measuring tiny magnetic signals, strong magnetic pulses (about 2 T) are used to modify cortical activity. Since TMS can now be performed under navigated conditions, it is possible to induce magnetic activity at the areas where for instance ipsilateral motor activity was measured with MEG. Few clinical applications have been described yet, but the combination of these techniques may be valuable in the understanding of brain networks involved in cerebral motor control under pathologic conditions (29).

132

GENERAL DISCUSSION AND FUTURE PERSPECTIVES

Finally, alterations in the sensorimotor network due to intracranial lesions should preferably be evaluated in the context of global network changes of the brain, which makes it clear that we as neurosurgeons not only have to worry about the classical functional areas, but also have to consider network features, to prevent postoperative neurological deficits.

Conclusion In this thesis we describe aspects of sensorimotor localization and lateralization using functional and structural imaging in patients with brain lesions and healthy subjects. We showed that the results of functional mapping using MEG, can be affected by the presence of intracranial lesions and epilepsy. Functional reorganization is a complex process and sometimes occurs due to lesions and may have implications for neurosurgical practice.

The combination of imaging modalities may offer additional value and insights in sensorimotor functioning. For clinical practice however, new technologies still have to be found to replace the gold standard procedure of the awake craniotomy.

8

133

CHAPTER 8

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

134

Barnes GR, Hillebrand A, Fawcett IP & Singh KD. Realistic spatial sampling for MEG beamformer images. Hum Brain Mapp 2004;23:120-7.

Cao Y, Dâ&#x20AC;&#x2122;Olhaberriague L, Vikingstad EM, Levine SR & Welch KM. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112-22.

Ciccarelli O, Toosy AT, Marsden JF, et al. Identifying brain regions for integrative sensorimotor processing with ankle movements. Exp Brain Res 2005;166:31-42.

Del Gratta C, Della Penna S, Ferretti A, et al. Topographic organization of the human primary and secondary somatosensory cortices: comparison of fMRI and MEG findings. Neuroimage 2002;17:1373-83.

Desmurget M, Bonnetblanc F & Duffau H. Contrasting acute and slow-growing lesions: a new door to brain plasticity. Brain 2007;130:898-914.

Fandino J, Kollias SS, Wieser HG, Valavanis A & Yonekawa Y. Intraoperative validation of functional magnetic resonance imaging and cortical reorganization patterns in patients with brain tumors involving the primary motor cortex. J Neurosurg 1999;91:238-50.

Gaetz W, Cheyne D, Rutka JT, et al. Presurgical localization of primary motor cortex in pediatric patients with brain lesions by the use of spatially filtered magnetoencephalography. Neurosurgery 2009;64:177-85; discussion 186.

Gallen CC, Schwartz BJ, Bucholz RD, et al. Presurgical localization of functional cortex using magnetic source imaging. J Neurosurg 1995;82:988-94.

Ganslandt O, Fahlbusch R, Nimsky C, et al. Functional neuronavigation with magnetoencephalography: outcome in 50 patients with lesions around the motor cortex. Neurosurg Focus 1999;6:e3.

Hari R, Reinikainen K, Kaukoranta E, et al. Somatosensory evoked cerebral magnetic fields from SI and SII in man.

Electroencephalogr Clin Neurophysiol 1984;57:254-63.

Hari R, Nagamine T, Nishitani N, et al. Time-varying activation of different cytoarchitectonic areas of the human SI cortex after tibial nerve stimulation. Neuroimage 1996;4:111-8.

Hattingen E, Good C, Weidauer S, et al. Brain surface reformatted images for fast and easy localization of perirolandic lesions. J Neurosurg 2005;102:302-10.

Hillebrand A & Barnes GR. A quantitative assessment of the sensitivity of whole-head MEG to activity in the adult human cortex. Neuroimage 2002;16:638-50.

Holodny AI, Schulder M, Liu WC, Wolko J, Maldjian JA & Kalnin AJ. The effect of brain tumors on BOLD functional MR imaging activation in the adjacent motor cortex: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 2000;21:1415-22.

Hou BL, Bradbury M, Peck KK, Petrovich NM, Gutin PH & Holodny AI. Effect of brain tumor neovasculature defined by rCBV on BOLD fMRI activation volume in the primary motor cortex. Neuroimage 2006;32:489-97.

Huttunen J, Kaukoranta E & Hari R. Cerebral magnetic responses to stimulation of tibial and sural nerves. J Neurol Sci 1987;79:43-54.

Jurkiewicz MT, Gaetz WC, Bostan AC & Cheyne D. Post-movement beta rebound is generated in motor cortex: evidence from neuromagnetic recordings. Neuroimage 2006;32:1281-9.

Kakigi R, Hoshiyama M, Shimojo M, et al. The somatosensory evoked magnetic fields. Prog Neurobiol 2000;61:495-523.

Kim SG, Ashe J, Georgopoulos AP, et al. Functional imaging of human motor cortex at high magnetic field. J Neurophysiol 1993;69:297-302.

Kobayashi M, Hutchinson S, Schlaug G & Pascual-Leone A. Ipsilateral motor cortex activation on functional magnetic resonance imaging during unilateral hand movements is related to interhemispheric interactions. Neuroimage 2003;20:2259-70.

Kober H, Nimsky C, MĂśller M, Hastreiter P, Fahlbusch R & Ganslandt O. Correlation of sensorimotor activation with functional magnetic resonance imaging and magnetoencephalography in presurgical functional imaging: a spatial analysis.

Neuroimage 2001;14:1214-28.

Korvenoja A, Kirveskari E, Aronen HJ, et al. Sensorimotor cortex localization: comparison of magnetoencephalography, functional MR imaging, and intraoperative cortical mapping. Radiology 2006;241:213-22.

Krings T, Reinges MH, Erberich S, et al. Functional MRI for presurgical planning: problems, artefacts, and solution strategies.

J Neurol Neurosurg Psychiatry 2001;70:749-60.

Kristeva R, Cheyne D & Deecke L. Neuromagnetic fields accompanying unilateral and bilateral voluntary movements: topography and analysis of cortical sources. Electroencephalogr Clin Neurophysiol 1991;81:284-98.

GENERAL DISCUSSION AND FUTURE PERSPECTIVES

25. Lehéricy S, Duffau H, Cornu P, et al. Correspondence between functional magnetic resonance imaging somatotopy and individual brain anatomy of the central region: comparison with intraoperative stimulation in patients with brain tumors.

J Neurosurg 2000;92:589-98.

26. Lin PT, Berger MS & Nagarajan SS. Motor field sensitivity for preoperative localization of motor cortex. J Neurosurg 2006;105:588-94.

27. Lu JF, Zhang H, Wu JS, et al. “Awake” intraoperative functional MRI (ai-fMRI) for mapping the eloquent cortex: Is it possible in awake craniotomy? Neuroimage Clin 2012;2:132-42.

28. Mandelli ML, Berger MS, Bucci M, Berman JI, Amirbekian B & Henry RG. Quantifying accuracy and precision of diffusion MR tractography of the corticospinal tract in brain tumors. J Neurosurg 2014;121:349-58.

29. Mäkelä JP, Vitikainen AM, Lioumis P, et al. Functional plasticity of the motor cortical structures demonstrated by navigated TMS in two patients with epilepsy. Brain Stimul 2013;6:286-91.

30. Mäkelä JP, Kirveskari E, Seppä M, et al. Three-dimensional integration of brain anatomy and function to facilitate intraoperative navigation around the sensorimotor strip. Hum Brain Mapp 2001;12:180-92.

31. Miyai I, Tanabe HC, Sase I, et al. Cortical mapping of gait in humans: a near-infrared spectroscopic topography study.

Neuroimage 2001;14:1186-92.

32. Nagarajan S, Kirsch H, Lin P, Findlay A, Honma S & Berger MS. Preoperative localization of hand motor cortex by adaptive spatial filtering of magnetoencephalography data. J Neurosurg 2008;109:228-37.

33. Nakanishi K, Inoue K, Hadoush H, Sunagawa T & Ochi M. Dipole orientation of receptive fields in the somatosensory cortex after stimulation of the posterior tibial nerve in humans. J Clin Neurophysiol 2014;31:236-40.

34. Nimsky C, Ganslandt O, Hastreiter P & Fahlbusch R. Intraoperative compensation for brain shift. Surg Neurol 2001;56:35764; discussion 364-5.

35. Niu C, Zhang M, Min Z, et al. Motor network plasticity and low-frequency oscillations abnormalities in patients with brain gliomas: a functional MRI study. PLoS One 2014;9:e96850.

36. Ossenblok P, Leijten FS, de Munck JC, Huiskamp GJ, Barkhof F & Boon P. Magnetic source imaging contributes to the presurgical identification of sensorimotor cortex in patients with frontal lobe epilepsy. Clin Neurophysiol 2003;114:221-32.

37. Romstöck J, Fahlbusch R, Ganslandt O, Nimsky C & Strauss C. Localisation of the sensorimotor cortex during surgery for brain tumours: feasibility and waveform patterns of somatosensory evoked potentials. J Neurol Neurosurg Psychiatry 2002;72:221-9.

38. Sahyoun C, Floyer-Lea A, Johansen-Berg H & Matthews PM. Towards an understanding of gait control: brain activation during the anticipation, preparation and execution of foot movements. Neuroimage 2004;21:568-75.

39. Sanchez-Panchuelo RM, Francis S, Bowtell R & Schluppeck D. Mapping human somatosensory cortex in individual subjects with 7T functional MRI. J Neurophysiol 2010;103:2544-56.

40. Schiffbauer H, Berger MS, Ferrari P, Freudenstein D, Rowley HA & Roberts TP. Preoperative magnetic source imaging for brain tumor surgery: a quantitative comparison with intraoperative sensory and motor mapping. J Neurosurg 2002;97:133342.

41. Stippich C, Freitag P, Kassubek J, et al. Motor, somatosensory and auditory cortex localization by fMRI and MEG. Neuroreport 1998;9:1953-7.

42. Stoeckel MC & Binkofski F. The role of ipsilateral primary motor cortex in movement control and recovery from brain damage. Exp Neurol 2010;221:13-7.

43. Takeuchi N, Oouchida Y & Izumi S. Motor control and neural plasticity through interhemispheric interactions. Neural Plast 2012;2012:823285.

44. Taniguchi M, Kato A, Ninomiya H, et al. Cerebral motor control in patients with gliomas around the central sulcus studied with spatially filtered magnetoencephalography. J Neurol Neurosurg Psychiatry 2004;75:466-71.

45. Taniguchi M, Kato A, Fujita N, et al. Movement-related desynchronization of the cerebral cortex studied with spatially filtered magnetoencephalography. Neuroimage 2000;12:298-306.

46. Tecchio F, Zappasodi F, Tombini M, et al. Brain plasticity in recovery from stroke: an MEG assessment. Neuroimage 2006;32:1326-34.

47. Terada K, Umeoka S, Baba K, et al. Generators of tibial nerve somatosensory evoked potential: recorded from the mesial surface of the human brain using subdural electrodes. J Clin Neurophysiol 2009;26:13-6.

48. Tozakidou M, Wenz H, Reinhardt J, et al. Primary motor cortex activation and lateralization in patients with tumors of the central region. Neuroimage (Amst) 2013;2:221-8.

49. Vardy AN, van Wegen EE, Kwakkel G, Berendse HW, Beek PJ & Daffertshofer A. Slowing of M1 activity in Parkinson’s disease during rest and movement--an MEG study. Clin Neurophysiol 2011;122:789-95.

50. Yousry TA, Schmid UD, Alkadhi H, et al. Localization of the motor hand area to a knob on the precentral gyrus. A new landmark. Brain 1997;120 ( Pt 1):141-57.

51. Ziemann U, Ishii K, Borgheresi A, et al. Dissociation of the pathways mediating ipsilateral and contralateral motor-evoked potentials in human hand and arm muscles. J Physiol 1999;518 ( Pt 3):895-906.

135

8

APPENDICES

List of abbreviations List of publications Acknowledgements | Dankwoord About the author

APPENDICES

LIST OF ABBREVIATIONS USED IN THIS THESIS AED AH BOLD CS CSF CST DNET DTI ECD ECS EEG ERD ERS FCD FLAIR FMRI HGG ISI LGG M1 MPRAGE MEG MN MRI MTS NLE PET PMC PPC PTN ROI S1 S2 SAM SEF SI SMA SNR TBI UH VR WHO 150

antiepileptic drugs affected hemisphere blood oxygenation level dependent central sulcus cerebrospinal fluid corticospinal tract dysembryoplastic neuroepithelial tumor diffusion tensor imaging equivalent current dipole electrical cortical stimulation electroencephalography event-related desynchronisation event-related synchronization focal cortical dysplasia fluid-attenuated inversion recovery functional magnetic resonance imaging high-grade glioma inter-stimulus interval low-grade glioma primary motor cortex magnetization prepared rapid acquisition gradient magnetoencephalography median nerve magnetic resonance imaging mesial temporal sclerosis non-lesional epilepsy positron emission tomography premotor cortex posterior parietal cortex posterior tibial nerve region of interest primary somatosensory cortex secondary somatosensory cortex synthetic aperture magnetometry somatosensory evoked field signal intensity supplementary motor area signal-to-noise ratio traumatic brain injury unaffected hemisphere volume rendering World Health Organization

APPENDICES

LIST OF PUBLICATIONS Willemse RB, van ’t Ent D, Pouwels PJW, de Munck JC, Vandertop WP Spatiotemporal imaging of somatosensory cortical activity with identical paradigms: comparison of fMRI and MEG (submitted) Willemse RB, Hillebrand A, Ronner HE, Vandertop WP, Stam CJ. Magnetoencephalographic study of hand and foot sensorimotor organization in 325 consecutive patients evaluated for tumor or epilepsy surgery. Neuroimage Clin 2016; 10: 46 -53 Willemse RB, Georgalas C, van Furth WR, Fokkens W. Endoscopic approach to the sella.

In: Rhinology and Skull Base Surgery: from the Lab to the Operating Room. Eds: Georgalas C and Fokkens W. Thieme 2012: 740-59.

Van Wijk BC, Willemse RB, Vandert WP, Daffertshofer A. Slowing of primary motor cortex oscillations in brain tumor patients in resting state and during movement.

Clin Neurophsyiol 2012, 123(11): 2212-9 Willemse RB, Pouwels PJW, Barkhof F, Vandertop WP. Localisation of the central sulcus region in glioma patients with three-dimensional fluid-attenuated inversion recovery and volume rendering: comparison with functional and conventional magnetic resonance.

Br J Neurosurg 2011; 25(2): 210-7 Willemse RB, de Munck JC, Verbunt JPA, van ’t Ent D, Ris PJ, Baayen JC, Stam CJ, Vandertop WP. Topographical organization of mu and beta band activity associated with hand and foot movements in patients with perirolandic lesions.

Open Neuroimag J 2010, 4: 93-99.

Brouwer MC, de Gans J, Willemse RB, van de Beek D. Neurological pictures. Sarcoidosis presenting with hydrocephalus. J Neurol Neurosurg Psychiatry 2009; 80(5): 550-1 Willemse RB, de Munck JC, van ’t Ent D, Ris PJ, Baayen JC, Stam CJ, Vandertop WP. Magnetoencephalographic study of posterior tibial nerve stimulation in patients with intracranial lesions around the central sulcus. Neurosurgery 2007 Dec;61(6): 1209-17.

Baayen JC, Willemse RB. Navigatie in het brein. Medisch Contact, 2004; 22: 890-893.

Willemse RB, Westermann CJJ, Vandertop WP. Cerebrovasculaire malformaties bij hereditaire hemorrhagische telangiëctasieën. Ned Tijdschr Neurol 2002; 6 : 472 – 476.

Willemse RB, Mager JJ, Westermann CJ, Overtoom TT, Mauser H, Wolbers JG. Bleeding risk of cerebrovascular malformations in hereditary hemorrhagic telangiectasia.

J Neurosurg 2000 May; 92(5): 779-84.

151

A

APPENDICES

Willemse RB, Egeler-Peerdeman SM. External lumbar drainage in uncontrollable intracranial pressure in adults with severe head injury: a report of 7 cases. Acta Neurochir Suppl. 1998 ; 71 : 37 –9.

Koelman JH, Willemse RB, Bour LJ, Hilgevoord AA, Speelman JD, Ongerboer de Visser BW. Soleus H-reflex tests in dystonia. Mov Disord. 1995 Jan; 10 (1): 44 – 50.

Willemse RB, Koelman JH, Bour LJ, Ongerboer de Visser BW, Independence of soleus H-reflex tests in control and spastic subjects shown by principal components analysis. Electroencephalogr Clin Neurophysiol. 1994 Dec ; 93(6) : 440-3

152

APPENDICES

DANKWOORD Als je tien jaar lang als staflid bij ieder jaargesprek gevraagd wordt wanneer je denkt dat proefschrift te gaan afronden, omdat je anders niet serieus genomen zal worden, dan komt er een moment dat je twee dingen kunt doen. Ofwel er helemaal mee stoppen, of er toch voor te gaan. Ik moet eerlijk zeggen dat ik neigde naar de eerste optie, maar dankzij de hulp van mijn promotoren heb ik toch besloten om ermee door te gaan. Dat ik dit dankwoord mag schrijven heb ik derhalve vooral aan hen te danken.

Prof. Dr. Vandertop, beste Peter, in de eerste plaats ben ik je zeer dankbaar dat jij in een cruciale fase van mijn opleiding in het VUmc kwam en dat ik veel met je mocht opereren.

Ik heb daar veel van geleerd en ook na de opleiding was en ben je altijd bereid bij moeilijke operaties hulp te bieden indien nodig, ongeacht het tijdstip! Ik vind je een fenomenale neurochirurg en uitstekende clinicus. Dat je ondanks je zeer drukke werkzaamheden nog tijd ziet om manuscripten in korte tijd te becommentariëren vind ik ongelooflijk, jammer dat Kees Stam soms toch nog net even sneller was….

Prof. Dr. Stam, beste Kees, je was al eerder betrokken bij het onderzoek, maar pas in de laatste fase gaf je aan om ook promotor te willen zijn, waarbij je opmerkte dat er nog nooit een promotie van jou niet was doorgegaan. Ik heb even gedacht dat ik misschien wel je eerste uitzondering zou zijn, maar besloot al snel dat ik je hulp met beide handen moest aanpakken. Ik bewonder je gave om de heel ingewikkelde materie waar jij mee bezig bent eenvoudig te verwoorden, evenals je subtiele opmerkingen bij manuscripten die je verplichten na te denken over een andere (betere..) manier van analyseren of noteren. Zeer veel dank voor al je hulp de afgelopen jaren.

Dr. Ir. Hillebrand, beste Arjan, zeer veel dank voor de uitgebreide hulp bij het laatste artikel.

Dankzij jouw doortastendheid m.b.t. de eisen van reviewers is publicatie toch een feit geworden en bleek promoveren opeens binnen handbereik.

Dr. Bernadette van Wijk en Prof. Dr. Andreas Daffertshofer van de Faculteit Bewegingswetenschappen, het was voor mij indrukwekkend om te zien wat jullie met data kunnen…het is geen eenvoudige statistiek, maar dan heb je ook wat. Dank voor de samenwerking, ik vond het erg leerzaam om te zien hoe jullie research bedrijven. Veel respect daarvoor!

De leden van de leescommissie, Prof. Dr. Ir. Natasha Maurits, Prof. Dr. Bernard Uitdehaag, Prof. Dr. FRDRK Barkhof, Prof. Dr. Jeroen Geurts en Prof. Dr. Clemens Dirven ben ik dankbaar voor het beoordelen van het manuscript.

De mede-auteurs Petra Pouwels, Jan de Munck, Dennis van ‘t Ent, Jeroen Verbunt, Frederik Barkhof, Hanneke Ronner, Hans Baayen (ook wel Baaijen) en Peter-Jan Ris wil ik danken voor alle hulp tijdens de beginjaren van dit lange traject.

153

A

APPENDICES

Dr. Linda Douw ben ik dank verschuldigd voor de hulp bij een jaar lang netwerk-analyse, helaas kwam er niks uit voort maar ik kan nu tenminste wel macro’s schrijven in Excel.

Inez Branco ben ik speciale dank verschuldigd voor de fMRI-MEG analyse. Infelizmente você desapareceu de vista ...

De stafleden van het Neurochirurgisch Centrum Amsterdam, de arts-assistenten en het secretariaat dank ik voor de prettige samenwerking.

Mijn goede vriend en paranimf Rob Gons, begenadigd pianist, top-clinicus en inmiddels ook hobby-fotograaf… ik bewonder je talenten in combinatie met bescheidenheid. Het doet me deugd dat we al die jaren ondanks de afstand contact hebben gehouden en dat jij nu ook mijn paranimf kunt zijn. Dank voor alle extramurale activiteiten van de afgelopen jaren, ik hoop op nog veel gezellige avonden met 23 buren en het zero-spel.

Mijn lieve zus en paranimf Sandra, ik bewonder je enorme doorzettingsvermogen en humor. Om samen met Pieter een bedrijf en een gezin te runnen valt niet mee, maar toch weet je dit allemaal ook nog te combineren met het onderhouden van al je vriendschappen.

Ik ben trots op je en heel erg blij dat je mij wilt bijstaan tijdens de promotie.

Mijn lieve schoonouders, Els en William, jullie hulp en steun in ons drukke leven maakt het soms net even wat makkelijker en bovendien het huis ook steeds wat mooier!

Mijn ouders, lieve Leni en Ben, ik ben jullie zeer dankbaar voor alle kansen die ik heb gekregen om dit mooie leven te leiden.

Mijn lieve Ellen, zonder jou was dit proefschrift waarschijnlijk veel eerder afgerond. Maar ondanks je tegenwerking cq. afleiding is het me toch gelukt….!

Dank voor je liefde, relativeringsvermogen en intercollegiaal overleg.

Lieve Eva, Friso en Kik, mijn dagelijkse portie vreugde en geluk!

154

APPENDICES

ABOUT THE AUTHOR Ronald Willemse was born on July 8th, 1968 in Haarlem. After finishing secondary school at the “Christelijk Atheneum Adriaen Pauw” in Heemstede, he started medical school at the University of Amsterdam in 1986. In between, he studied Medical Informatics at the same university and obtained his doctoral degree in 1992. His medical doctor degree was obtained in 1994. In 1994 he started as a resident in the department of Neurosurgery of the VU University Medical Center (VUmc) where he started his neurosurgical training in 1997 (head: prof dr. H.A.M. van Alphen † & prof. dr. W.P. Vandertop). Since 2003 he is working as staff member at the Neurosurgical Center Amsterdam. He lives with Ellen Mandl, their daughter Eva and their two sons Friso and Kik.

A

155

APPENDICES

DISSERTATION SERIES BRAIN TUMOR CENTER AMSTERDAM Datum

Naam

Titel

1

10-06-05

Carla Verstappen

Cancer therapy related neurotoxicity

2

28-09-05

Maaike Vos

Evaluation of reponse, toxicity and outcome in glioma therapy

3

20-12-05

Birgit Georger

Conditionally replicative adenoviruses for the treatment of malignant glioma and neuroblastoma

4

20-12-05

Jacques Grill

Functional molecular imaging of cancer development and stem cell regeneration in the nervous system

5

19-06-09

Fonnet Bleeker

Mutational profiling of glioblastoma

6

24-11-09

Philip de Witt Hamer

Glioblastoma: between bed and bench

7

07-05-10

Ingeborg Bosma

Cognitive dysfunction in glioma; underlying mechanisms and consequences

8

23-09-10

Christian Badr

Bioluminescence imaging in glioblastom: monitoring of biological processes and novel therapeutics

9

08-11-10

Linda Douw

Neural networks in brain tumors; the interplay between tumor, cognition, and epilepsy

10

10-06-11

Sander Idema

Improving oncolytic viral therapy for glioma

11

05-10-11

Anneke Niers

Novel biosensors for preclinical brain tumor analysis

12

03-07-12

Viola Caretti

Pioneering preclinical research in diffuse intrinsic pontine glioma: towards new treatment strategies

13

29-10-12

Leonora Balaj

Exosomes: the biological messengers

14

08-02-13

Marjolein de Groot

Epilepsy in brain tumor patients; towards improved and personalized treatment

15

04-06-13

Edwin van Dellen

Lesions in the connected brain; a network perspective on brain tumors and lesional epilepsy

16

04-12-13

Michiel Smits

Micro-RNA and epigenetic signaling in glioma angiogenesis

17

11-12-13

Eefje Sizoo

The end-of-life phase of high-grade glioma patients; towards a dignified death

18

17-06-14

Dannis van Vuurden

Innovative treatment targets in pediatric high-grade brain tumors

19

07-01-15

Lotte Hidding

Treatment sensitizers for high-grade tumors

20

11-05-15

Florien Boele

Towards improving health-related quality of life in glioma patients and their informal caregivers

21

30-09-15

Johan Koekkoek

Epilepsy in glioma patients; optimizing treatment until the end of life

22

19-01-16

Sjoerd van Rijn

Functional molecular imaging of cancer development and stem cell regeneration in the nervous system

23

24-03-16

Hinke van Thuijl

Molecular characterization of low-grade glial neoplasms

24

08-06-16

Ronald Willemse

Functional mapping of the sensorimotor cortex: Clinical studies with MEG and fMRI

156