Volume 2 Issue 1 Fall 2012 neurogenesisjournal.com
Neurogenesis gether o T s e c e i P
Putting
the
The Journal of Undergraduate Neuroscience
Autism Creativity Spatial Memory
Feature Articles
Learning and non-learning effects of ginkgo biloba extract Decoding methods in brain-machine interface systems
Volume 2 Issue 1 Fall 2012
Copyright Š 2012
NEURO Genesis
Editorial Board Editors-in-Chief Rory Lubner Class of 2013 rory.lubner@duke.edu
Kelly Ryan Murphy Class of 2013 kelly.murphy@duke.edu
Publishing Editors Christine Lee Class of 2014 christine.lee@duke.edu Biqi Zhang Class of 2014 biqi.zhang@duke.edu
Managing Editors Banafsheh Sharif-Askary Class of 2013 bs118@duke.edu Ha Tran Class of 2015 ha.tran@duke.edu
Design Editor Tiffany Chien Class of 2015 tiffany.chien@duke.edu
ii | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Faculty Advisors Leonard White, Ph.D. Duke University School of Medicine Director of Education Duke Institute for Brain Sciences len.white@duke.edu Christina Williams, Ph.D. Professor Director of Undergradute Studies Duke Institute for Brain Sciences williams@psych.duke.edu
Craig Roberts, Ph.D. Assistant Director of Education Duke Institute for Brain Sciences craig.roberts@duke.edu Ann Motten, Ph.D. Department of Chemistry ann.motten@duke.edu
Online Editor Kyle Rand Class of 2014 kyle.rand@duke.edu
NEURO Genesis
Letter from the Editors Putting the pieces together A mind is much like a puzzle—at the risk of sounding cliché. But as we research, study, and explore, we fill in the gaps to create a whole picture of the brain. We piece together its intricacies and mysteries with the help of professors, textbooks, and our own research. To figure out more, to discover earlier. The topics featured in this issue—learning, memory, and creativity—place focus on cognitive neuroscience. From this scope, we look at the biological substrates that contribute to higher mental function. These topics are highly relevant to undergraduate studies, where our minds work to establish our creative bounds. We also feature the work of Olga Mutter and Mary Petrosko, exemplifying the extended research experience. Petrosko of Dominican University explores the Ginkgo biloba extract, an ingredient featured in the news for its positive effects on memory, while Mutter explores brain-interface systems that control our movement reaction times. We congratulate them for their persistence in the lab and dedication to discovery. Please enjoy the issue, and fill-in some pieces for yourself. Next issue, we have plans to feature articles from all over the country and world, but for now, we keep the lens primarily on the Durham area. Talents from many people came together to create this issue. Thank you to all those people, and especially to the research mentors and professors of undergraduates who are so willing to have students work with them. Sincerely,
Rory Lubner & Kelly Ryan Murphy Editors-in-Chief
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | iii
NEURO Genesis
Table of Contents 2
Neurogenesis: Evidence, function and influences
6
The joining of spatial episodic memories in the rat hippocampus
7
Young genes and their regulation as a window to human brain evolution
Banafsheh Sharif-Askary
Dennis Kwon
Adrienne Niederriter
11
Creativity and the default network: Is the brain at work while the mind is wandering? Elizabeth Beam
25
Examining the “dys”-ordered schizophrenic brain
40
Autism as a model system for examination of Rushworth’s reinforcement-guided decision-making model and the functional neuroanatomy of the Anterior Cingulate Cortex
Barrington Quarrie
Colin Martz
43
Society’s stake in biomedical cognitive enhancement
47
The a4b2 neuronal nicotinic acetylcholine receptor: Developing new therapeutic agents for smoking cessation based on cytisine and bupropion
Akash Shah
Sheetal Hegde
F eature
A rticles
14
Learning and non-learning effects of Ginkgo biloba extract EGb 761 in Aplysia californica
27
Improved decoding methods to reduce reaction time in brain-machine interface systems
Mary Petrosko
Olga Mutter
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 1
NEURO
Review
Genesis
Neurogenesis: Evidence, function, and influences Banafsheh Sharif-Askary1 1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Banafsheh Sharif-Askari (banafsheh.sharif-askary@duke.edu)
First observed in 1962 by Joseph Altman, neurogenesis is defined as the birth and development of neurons that integrate into existing neural circuitry under natural, physiological, and neurological conditions (Altman, 1962). There is ample evidence for neurogenesis in the adult hippocampal dentate gyrus, subventricular zone and substantia nigra in response to increasing cognitive complexity, biological demands and neurodegeneration. Neurogenesis in the dentate gyrus was originally refuted, but more recent work with bromodeoxyuridine (BrdU) labeling gives evidence for cellular growth in both rodents and nonhuman primates (Eldridge, Tilbury, Goldsworthy, & Butterworth, 1990; Esposito et al., 2005; Ngwenya, Peters, & Rosene, 2006). Hippocampal neurogenesis is spurred by new memory formation and functions to create space for growing input (Cameron & McKay, 2001; Tulving & Markowitsch, 1998). Neurogenesis in the subventricular zone has been linked to the growth of olfactory bulb cells which serve to strengthen distinction among odors, a function necessitated by evolutionary needs (Bédard, Lévesque, Bernier, & Parent, 2002; Bédard & Parent, 2004; Cecchi, Petreanu, Alvarez-Buylla, & Magnasco, 2001; Gheusi et al., 2000; Lois & Alvarez-Buylla, 1994). Lastly, while neurogenesis in the substantia nigra functions to replace lost dopamine (DA) neurons, DA itself actively regulates the rate of cell proliferation (Diaz et al., 1997, Shan et al. 2006; Zhao et al., 2003). The various mechanisms of neurogenesis in the dentate gyrus, subventricular area and substantia nigra have been elucidated through innovative labelling techniques within a wide variety of model organisms. Cell proliferation of this nature gives rise to a variety of cognitive enhancements such as memory formation, olfactory differentiation. Beyond cognitive enhancements, neurogenesis has been shown to reverse the effect of neuronal loss through negative feedback mechanisms in the substantia nigra. Recent evidence for neurogenesis in the primate hippocampal dentate gyrus elucidates the potential for neuronal regeneration in a part of the brain previously thought to undergo no such growth. Pasko Rakic’s previous work disproving the lack of neurogenesis in the adult rhesus monkey, Macaca mulatta, had several shortcomings that prevented a full understanding of the potential for neurogenesis. Rakic chose to examine pregnant monkeys which have been shown to exhibit increased levels of circulating 2 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
glucocorticoids (i.e. estradiol). Since estradiol has been found to inhibit neurogenesis, Rakic may have unintentionally underestimated the presence of new cells (Ormerod, 2003). Another source of underestimation may have been the 3[H] thymidine, a mitotic S-phase label, which infiltrates only 1 to 3 millimeters (mm) of tissue. This superficial labelling leaves deeper tissue unanalyzed, therefore making it impossible to see potential neuronal proliferation. In fact, many current studies have abandoned the 3[H] thymidine marking technique and have opted instead for the use of bromodeoxyuridine (BrdU), which directly integrates itself into the nuclei of cells undergoing mitosis, and retroviral fluorescence labeling (Dawley, Fingerlin, Hwang, John, & Stankiewicz, 2000; Eldridge, Tilbury, Goldsworthy, & Butterworth, 1990; Lois & Alvarez-Buylla, 1994; Ngwenya, Peters, & Rosene, 2006; Zhao et al., 2003). These labeling techniques avoid contamination and cell apoptosis while producing more efficient results than 3[H] thymidine (Eldridge, Tilbury, Goldsworthy, & Butterworth, 1990). More accurate forms of subject selection and labeling have identified neurogenesis in the dentate gyrus of mice. Using a retroviral fluorescence labeling technique, researchers monitored the various stages of division and maturation in progenitor cells. After four weeks, dendrites had branched all the way to the molecular layer leading to more glutamatergic connections. In addition to showing that neurogenesis is possible within the mouse dentate gyrus, this study also demonstrates that the development of new neurons in adulthood follows the same sequence of growth as early hippocampal cells developed during birth (Esposito et al., 2005). To see if there was parallel dentate gyrus neurogenesis in primates, researchers utilized BrdU to track neuronal growth in 11 adult rhesus monkeys. Electron microscopy demonstrated that these mature neurons did indeed form branching synapses within the dentate gyrus, integrating themselves into the hippocampal circuitry. Although these results are similar to findings in rodents, primate neuronal maturation took at least 5 weeks whereas rodent neuronal maturation took only 8 days (Ngwenya, Peters, & Rosene, 2006). This difference in maturation time may be attributed to the markedly lower threshold for long-term potentiation (LTP), which is known to enhance the extent of neurogenesis. (Bruel-Jungerman, Davis, Rampon, & Laroche, 2006; Kempermann, Gast, Kronenberg, Yamaguchi,
Review & Gage, 2003). This delayed maturation in primates may have negative effects for treatments that stimulate neuron proliferation. Rather than instant improvements, drugs for neurodegenerative diseases may take months or years to have visible effects. Neurogenesis in the adult dentate gyrus allows organizational and morphological alterations in proliferating cells in response to the formation of new memories (Tulving & Markowitsch, 1998). Studies have shown that elimination of adult-born hippocampal cells affects spatial memory, indicating that neurogenesis in this area functions to aid complex spatial-relational memory. Researchers gave doses of reverse tetracycline-controlled transactivator (rtTA) transcription factor to overexpress the apoptosis-inducing Bax protein. The results reveal that prevention of neurogenesis does not affect simple spatial tasks like habituating to new environments or remembering previous environments. However, more complex flexibility-dependent learning tasks, such as encoding spatial-relational memories and creating spatial maps with cues were disrupted by inhibiting the proliferation of neurons (Dupret et al., 2008). While many new neurons serve as replacements for non-functioning or dead neurons, others neurons derived from learning experiences may serve entirely new roles within the brain (Cameron & McKay, 2001). Neurons generated in the dentate gyrus in adulthood often demonstrate distinctive electrophysiological characteristics from mature neurons. For example, immature neurons were found to have extended long-term potentiation as well as smaller responses to GABA inhibition (Liu & Martin, 2003). This neuronal distinctiveness may serve as a mechanism to avoid catastrophic interference, the phenomenon wherein novel training disrupts prior training, in response to the acquisition of new information (French, 2003). This flexibility, provoked by the creation of new memories, is the neural basis for new memory encoding. While newly born neurons in the adult dentate gyrus extend their processes into existing hippocampal circuitry, neurons generated in the subventricular zone (SVZ) of the lateral ventricles must migrate to the olfactory bulb via the rostral migratory stream. In 1994, researchers utilized injected thymidine and dye to track clusters of SVZ cells as they migrated through the mouse brain. SVZ cells from the lateral ventricle travel to the olfactory bulb where they differentiate (Lois & Alvarez-Buylla, 1994). This experiment was later repeated with adult squirrel monkeys using BrdU tracking. While BrdU+ cells were found in the olfactory bulb, in accordance with the results of Lois and Alvarez-Buylla (1994), there were also a significant number of BrdU+ cells that strayed from the rostral migratory stream and rested instead in the olfactory tubercle (Bédard, Lévesque, Bernier, & Parent, 2002). These results are notable because they imply that neurogenesis within new world primates, such as the squirrel monkey, has persisted and may pose an evolutionary advantage. Furthermore, the population of the olfactory tubercle suggests that immature neurons are involved in the dopaminergic
NEURO Genesis
pathways present there (Wesson & Wilson, 2010). Unlike studies on rodents and nonhuman primates, Bédard and Parent (2004) used double immunofluorescence labeling to understand neuroblast migration in postmortem human olfactory bulbs. Markers that indicate cell proliferation and migration provided evidence that neurogenesis does occur in the postmortem human olfactory bulb. However, there are limitations to using the postmortem brain as a model. First, the olfactory bulb is rather small and fragile and, therefore, more prone to damage. Furthermore, data regarding the fate of new neurons, such as life span, procession to neural systems and death, are difficult to gather. This information is available only in living tissue that integrates markers into its own replicating DNA (Bédard & Parent, 2004). Subventricular neurogenesis and migration to the olfactory bulb have been linked to greater olfactory differentiation. Gheusi et al. (2000) demonstrated this by observing mice with a small number of SVZ cells that migrate to the olfactory bulb. While the ability to discern between various odors was weakened, the ability to sift out and identify one particular odor, or odor sensitivity, remained unaffected . The conservation of memory may be accounted for by the normal neurogenesis within the dentate gyrus, the region commonly associated with memory. However impairment in smell discrimination suggests that new granule cells allow the olfactory bulb to redistribute neural representations of different odors. This reallocation of neural resources creates greater differences between related odors and ultimately improves olfactory discrimination (Cecchi, Petreanu, Alvarez-Buylla, & Magnasco, 2001). These findings can be linked to biological and evolutionary behaviors in nature. For example, prairie voles have a significantly higher number of proliferating cells in the rostral migratory stream during the estrous cycle (Smith, Pencea, Wang, Luskin, & Insel, 2001). Mating behaviors in mammals are often associated with heightened olfaction due to the secretion of pheromones by potential mates. Thus, an increase in olfactory cells would allow for greater discrimination between different mates and prove useful in a stage of life heavily dependent on olfactory senses. Temporal influences on neurogenesis are also evident in the work of Dawley, Fingerlin, Hwang, John, and Stankiewicz (2000) with red-backed salamanders. Salamanders observed in the month of May displayed significantly higher BrdU+ nuclei than salamanders observed in any other month. This seasonal neurogenesis could be an indicator of adjusting olfactory circuitry in response to novel environments and changing chemosensory priorities. Compared to neurogenesis in the dentate gyrus and subventricular zone, neuronal proliferation occurs on a much smaller scale in the adult substantia nigra (SN). Neurogenesis in the SN functions to replenish DA neurons and is invariably related to the levels of dopamine present. Work with rodents demonstrated that 21 days after BrdU injections in the SN, there were high levels of BrdU+ nuclei in the medial-rostral SN. The medial-rostral SN has Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 3
NEURO Genesis
the highest neuron density indicating that growth is most prevalent in areas already receiving the most information (Zhao et al., 2003). In a second study, Zhao (2003) gauged the connectivity of the newly generated neurons using the neurotropic psuedorabies virus (PRV). A large portion of these marked neurons had axon projections to both the striatum and the cortex. Though the rate of neurogenesis is rather low, further work over a longer period of time could investigate the potential of completely replenishing the SN with new dopamine neurons. However, findings from Zhao et al. are challenged by Frielingsdorf, Schwarz, Brundin and Mohapel (2006), who found no such proof of dopaminergic neurogenesis. These differing results may be attributed to the different concentrations of the BrdU marker used. In fact, within the SN, BrdU uptake may not take place exclusively during mitosis; evidence has shown that BrdU can be integrated during nuclear repair (Rakic, 2002). To put this discrepancy to rest, Shan et al. (2006) abandoned the use of BrdU and instead utilized the Nestin reporting (pNes-LacZ) system within non-Parkinson’s disease mice instead. They observed a strong presence of filament proteins, which suggests that there is indeed a basal level of neural progenitor cells in the SN of normal mice, consistent with Zhao’s findings (2003). Furthermore, because Shan et al. (2006) used normal non-PD mice, it is implied that SN neurogenesis is a process not only spurred by DA deficits, but also one that is prevalent in normally functioning substantia nigra as well. The recent prevalence of neurodegenerative diseases that destroy dopaminergic neurons in the adult substantia nigra (SN) has focused attention on the effects of neuronal loss on neurogenesis. Dopamine and dopamine receptors, especially D3, have often been associated with neurogenesis because of their strong presence in germinal neuroepithelial areas of the basal forebrain in rats. Because these dopamine neurotransmitters appear early during the development of these highly proliferating areas, they have been thought to contribute to neurogenesis through precursor cell regulation. There are also elevated D3 levels in the germinal subventricular zone (Diaz et al., 1997). Since the subventricular zone undergoes neurogenesis that populates the olfactory bulb, future research should examine the effect of fluctuating dopamine levels in this area. Zhao et al. (2003) estimated the effect of DA neuron loss on neurogenesis using universal doses of neurotoxin, MPTP (1-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridine). MPTP acts as an artificial inducer of Parkinson’s disease, which is characterized by a gradual decrease in the number of DA neurons in the adult substantia nigra. MPTP injection was followed by the death of nearly half of the SN dopaminergic cells, which subquently led to a doubling of BrdU+ cells approximately three weeks after neurotoxin injection. This rapid growth in response to cell death indicates that neurogenesis is spurred by the systematic loss of DA neurons. These results have been replicated by Shan et al. (2006) using a Nestin reporting system and again by Frielingsdorf, Schwarz, Brundin and Mohapel (2006), 4 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Review who originally challenged the Zhao et al. papers (2004). It is interesting to note, however, that much higher doses of MPTP are necessary to result in the high degree of SN neuronal death seen in primates (Schmidt & Ferger, 2001). It may be useful to study whether this differential in MPTP dosage has any other effect on the rodent brain. Researchers could systematically increase the dosage until the threshold value, which begins impairing rodent SN neurons. Analysis of metabolism and protein function in the rodent brain at this stage may provide useful clues to bolstering primate defenses against the MPTP toxin. The understanding of adult neurogenesis has transformed from a vague theory to a well-established truth in the last two decades. Research with a wide variety of species including rodents, nonhuman primates, and humans, has ellucidated the mechanism, composition, and locations of neural proliferation. Neurogenesis in the adult dentate gyrus, subventricular zone and substantia nigra have deep biological implications, which range from memory expansion, olfactory acuity and neurodegenerative compensation. The regulation of neurogenesis is intrinsically tied to these functions, demonstrating that proliferation acts in accordance with neurological need. Altman, J. (1962). Are new neurons formed in the brains of adult ma mmals? Science, 135, 1127–1128. Bédard, A., Lévesque, M., Bernier, P., & Parent, A. (2002). The rostral migratory stream in adult squirrel monkeys: contribution of new neurons to the olfactory tubercle and involvement of the antiapoptotic protein Bcl-2. European Journal of Neuroscience, 16, 1917-1924. Bédard, A., & Parent, A. (2004). Evidence of newly generated neurons in the human olfactory bulb. Developmental Brain Research. 151, 159-68. Cameron, H., & Mckay, R. (2001). Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. The Journal of Compara tive Neurology, 435, 406 - 417.Bruel-Jungerman, E., Davis, S., Rampon, C., & Laroche, S. (2006). Long-term potentiation enhances neurogenesis in the adult dentate gyrus. The Journal of Neuroscience. 26(22), 58885893. Cecchi, G.A., Petreanu, L.T, Alvarez-Buylla, A., & Magnasco, M.O. (2001). Unsupervised learning and adaptation in a model of adult neu rogenesis. Journal of Comparative Neuroscience, 11, 175–182. Dawley, E.M., Fingerlin, A., Hwang, D., John, S.S., & Stankiewicz, C.A.(2000). Seasonal Cell Proliferation in the Chemosensory Epithe lium and Brain of Red-Backed Salamanders, Plethodon cinereus. Brain, Behaviour and Evolution, 56, 1-13. Diaz, J., Ridray, S., Mignon, V., Griffon, N., Schwartz, J., & Sokoloff, P. (1997). Selective Expression of Dopamine D3 Receptor mRNA in Proliferative Zones during Embry onic Development of the Rat Brain. The Journal of Neuroscience. 17(11), 4282-4292 Dupret D., Revest J.M., Koehl. M., Ichas, F., De Giorgi, F., Costet, P., Abrous, D.N., & Piazza, P.V. (2008) Spatial Relational Memory Re quires Hippocampal Adult Neurogenesis. PLoS ONE, 3, 14. Eldrige, S.R., Tilbury, L.F, Goldsworthy, T.L., & Butterworth, B.S. (1990). Measurement of chemically induced cell proliferation in rodent liver and kidney: a comparison of 5-bromo-2’-deoxyuridine Q and [3H] thymidine administered by injection or osmotic pump. Carcinogenesis, 11(12), 2245-2251. Esposito, M.S., Piatti, V.C., Laplagne, D.A., Morgenstern, N.A., Ferrari, C.C., Pitossi, F.J., & Schinder, A.F. (2005). Neuronal Differentiation in the Adult Hippocampus Recapitulates Embryonic Development. The Journal of Neuroscience. 25(44), 10074 –10086. French, R. M. (2003) Catastrophic Forgetting in Connectionist Networks. Encyclopedia of Cognitive Science. 1, 431 - 435.
Review
NEURO Genesis
Frielingsdorf, H., Schwarz, K., Brundin. B., & Mohapel, P. (2004). No evidence for new dopaminergic neurons in the adult mammalian sub stantia nigra. Proceedings of the National Academy of Sciences. 101(27), 10177-82. Gheusi, G., Cremer, H., McLean, H., Chazal, G., Vincent, J., & Lledo, P. (2000). Importance of newly generated neurons in the adult olfactory bulb for odor discrimination. Proceedings of the National Academy of Sciences.97(4), 1823-1828. Kempermann, G., Gast, D., Kronenberg, G., Yamaguchi, M., & Gage, F.H. (2003). Early determinations and long-term persistence of adultgenerated new neurons in the hippocampus of mice. Development, 130(2), 391-399. Liu, Z., & Martin, L.J. (2003). Olfactory bulb core is a rich source of neural progenitor and stem cells in adult rodent and human, Journal of Com parative Neurology, 459, 368–391. Lois, C., & Alvarez-Buylla, A. (1994). Long-distance neuronal migration in the adult mammalian brain, Science, 264, 1145– 1148. Ngwenya, L.B., Peters, A., & Rosene, D.L. (2006). Maturation sequence of newly generated neurons in the dentate gyrus of the young adult rhesus monkey. The Journal of Comparative Neurology. 498, 204–216. Ormerod, B.K., Lee, T.T., & Galea, L.A. (2003). Estradiol initially enhances but subsequently suppresses (via adrenal steroids) granule cell proliferation in the dentate gyrus of adult female rats. Journal of Neuro biology, 55, 247–260. Pencea V., Bingaman, K.D., Freedman, L.J., & Luskin, M.B.(2001). Neurogenesis in the Subventricular Zone and Rostral Migratory Stream of the Neonatal and Adult Primate Forebrain, Experimental Neurology, 172, 1-16, Rakic, P. (1985). Limits of Neurogenesis in Primates. Science, 227, 10541056 Rakic, P. (2002). Adult neurogenesis in mammals: An identity crisis. Jour nal of Neuroscience, 22, 614–618. Schmidt, N., Ferger, B. (2001). Neurochemical findings in the MPTP model of Parkinson’s disease. Journal of Neural Transmission, 108(11), 1263-1282. Smith, M.T., Pencea, V., Wang, Z., Luskin, M.B., & Insel, T.R. (2001). Increased number of BrdU-labeled neurons in the rostral migratory stream of the estrous prairie vole. Hormonal Behavior, 39, 11-21. Tulving, E., & Markowitsch, H. J. (1998). Episodic and declarative memory: Role of the hippocampus. Hippocampus, 8, 198–204. Wesson, D.W., & Wilson D.A. (2010). Smelling Sounds: Olfactory–Audi tory Sensory Convergence in the Olfactory Tubercle. The Journal of Neuroscience, 30, 3013-3021. Zhao, M., S. Momma, K. Delfani, M. Carlen, R. M. Cassidy, C. B. Johans son, H. Brismar, O. Shupliakov, J. Frisen, & A. M. Janson. (2003) Evi dence for neurogenesis in the adult mammalian substantia nigra. Proceedings of the National Academy of Sciences of the United States of America, 100, 7925-7930.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 5
NEURO Genesis
Abstract
The joining of spatial episodic memories in the rat hippocampus Dennis Kwon1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Dennis Kwon (dennis.kwon@duke.edu) 1
The hippocampus has been implicated in two, critical neural processes. The first involves encoding of episodic memories; studies of hippocampal lesions in humans and rats confirm similar episodic memory deficits. The second is in spatial mapping and navigation; hippocampal place cells in rats fire rapidly when the animal is in a confined space. The relationship between these two functions of the hippocampus has not been fully established. Therefore, the current study addresses this question by examining firing characteristics evoked by two different episodic memories that are spatial in nature and also evoked by an episodic memory that represents a combination of these two separate memories. This was done by having rats run two tracks of different shapes (hat and W tracks) separately for several trials, then joining the tracks so that they run the two as a single track (hat then W) in several subsequent trials. Their position on the tracks and the corresponding firing rates of multiple single hippocampal neurons were recorded using multiple tetrodes surgically placed into the rat hippocampus, and then the data for the two situations were compared. Specifically, the change in the firing of place cells from the separate to the joined condition were analyzed. The peaks in firing rates as the rats run along specific locations on the tracks are called place fields. Most hippocampal neurons recorded were found to have different place fields in the joined condition as opposed to the separate condition. However, some place fields were retained from the separate to the joined condition, and the majority of these were found to fire when the rats ran on the hat track, the first track that the rats ran on (of the two tracks: hat and W) in the joined track. This finding may indicate that when rats learn to run a combined track from two separate tracks they are already familiar with, they encode the first track as an “anchor� for the memory of the new joined track.
6 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
NEURO
Letters
Genesis
Young genes and their regulation as a window to human brain evolution Adrienne Niederriter1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Adrienne Niederriter (adrienne.niederriter@duke.edu) 1
SUMMARY: The brain is one of the most complex and critical parts of human evolution and its own evolution has attracted much attention. Unfortunately, until recently, much of the investigation on the evolution of gene expression was based solely on adult brains rather than on developing brains. Within the past year, reports were published that a large number of evolutionarily “young”, or clade-specific, genes are expressed in the fetal brain of humans. A spatiotemporal expression bias toward early development, mainly in the youngest structure of the brain, the neocortex, suggests that these new genes have great significance in brain evolution. Their upregulation and transcript variation in the early developing human brain also shows that these genes likely play diverse functional roles, which could be a consequence of selection for novel functions in the newly-evolved brain structure. In particular, a significant number of these young genes encode transcription factor-related domains, suggesting that fine tuning of gene regulation by young genes may be a major contributor to “human-specific” traits (Stahl & Wainszelbaum, 2009). In response to this new study, the proposed research will determine when evolution began selecting for brain function in the primate line via recruitment of new genes into the neocortex. In addition, we will investigate if any of these genes are related to transcriptional regulation. Analyses will be conducted using transcriptome data from developing brains in several primates which will determine when, evolutionarily, the recruitment of new genes began. This transcriptional data will also cover several developmental stages to compare differences that may arise in the timing of upregulation of certain genes between species. This study will shed new light not only on the evolution of the human genome, but the evolution of the human brain and what factors may have impacted the selection of the young genes. Furthermore, this comparison may provide important insights into genes responsible for differences in cognitive functions between humans and our primate relatives, as well as help to understand when in the primate lineage such changes came about. As a result, this research may aid in the identification new candidate genes for future studies of neurological and cognitive disorders.
Aims 1. Determine what young genes are upregulated in the neocortex (compared to non-neocortical areas) of multiple primates at two developmental times via TagSeq methods, an ultra-high throughput sequencing of 21 base pair cDNA tags for sensitive and cost-effective gene expression profiling. Such sequencing will be carried out with the purpose of comparing the amount of young genes upregulated, identifying which of those overlap with those found in the human brain, and which are species specific. We expect there will be similar upregulation of genes in the fetal primate brains as with the humans, but that there will be fewer in nonhuman primates. We also expect that there will be non-genic regions upregulated as possible cis-regulatory regions. 2. Determine which of these upregulated young genes are related to transcription factors via InterPro Database, a protein sequence analysis and classification tool, with the purpose of determining a possible correlation between young genes and regulatory function. We expect that a lower percentage of the upregulated genes will
be related to transcription factors, as compared with the human sample shown in (Zhang, Landback, Vibranovski, & Long, 2011). Background Increased brain size and complexity have long been thought to define humans as a species, but it was recently hypothesized that the differences really lie in the development required for such intricate circuitry, which is enabled by precise spatiotemporal regulation of the transcriptome ( Johnson et al., 2009). Transcriptional regulation of key developmental genes is thought to contribute to the formation of distinct neuronal circuits, which likely enabled higher cognitive function, and thus the evolution of the human brain. One mechanism enabling this complexity is alternative splicing, which has also been investigated as a significant source of diversity between humans and chimpanzees (Calarco, Saltzman, Ip, & Blencowe, 2007; Dehay & Kennedy, 2009), adding to the case for regulatory evolution of the primate line. As a result, research has been conducted on human brain gene expression, showing that the adult brain expresses
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 7
NEURO Genesis
more alternatively spliced transcripts than any other tissue (Yeo, Holste, Kreiman, & Burge, 2004). This evidence was then followed by global transcriptome analyses of fetal brains to investigate gene expression more explicitly at the time of neuronal differentiation and establishment of connectivity ( Johnson et al., 2009). In addition to upregulation of genes, certain conserved noncoding sequences are also differentially expressed in the fetal brain. Specifically, neocortical genes that displayed differential expression were often near highly conserved noncoding sequences, many of which appeared to have undergone human-specific accelerated evolution. The authors believed that these sequences were likely to act as cis-regulatory elements, suggesting that the proposed human specific regulatory evolution has greatly contributed to increased spatial specificity for expression of a subset of genes during development ( Johnson et al., 2009). While these data seem promising and innovative, not enough research has been done to make any decisive claims in terms of human evolution, particularly what makes the human brain unique from other primates. In a new study, Zhang (Zhang et al., 2011) has examined similar human transcription data while also taking into account the ages of genes as a means to detect areas of recent evolution in the human brain. Aging of such genes was accomplished by aligning orthologous syntenic regions across the vertebrate phylogeny to identify in which branch a new gene arose (Zhang et al., 2011). However, this was done for only mice and humans, and we aim to go one step further with primate comparisons. Zhang et al found that a much larger proportion of young, primate-specific, genes were expressed in the fetal human brain as compared with rodent-specific genes in mice, and that the higher transcription of the young genes was brain specific, rather than seen throughout all fetal tissues, as mentioned above. Upregulation of such genes during the development of the neocortex, relative to the adult, suggests they have a specific role in fetal brain development. Old genes predating the primaterodent split were found to be roughly equally distributed between early (up to 0.38 years) and late developing brains. It was proposed that this pattern may be distinct to the human lineage, as compared with mice, but because the neocortex is rather small and simple in mice, comparisons are complicated. Because the authors based their analysis on primate and rodent specific genes and the transcriptome data of only human and mouse, it is unclear whether the accelerated origination of development genes is human-specific or rather primate-specific. Our similar comparisons between the more closely related primates will help to narrow down when exactly the recruitment of these genes began in the primate lineage and when 8 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Letters upregulation began to play a role. All of the primates included in this study have highly developed neocortical brain regions, but structural and functional differences are still present, which may not be explained by our data as the primate brain functions, and the related genes are not completely understood. The authors also noted that the young genes were found to be scattered across the entire genome and likely caused by multiple independent events (Zhang et al., 2011). It could be inferred that they have persisted as the result of a systematic force rather than mutational bias, most likely in positive selection, which we believe to be a sign of selection for brain function. Furthermore, out of the young genes identified in humans, around 13% encoded transcription factor related domains- almost twice the amount found in mice (Zhang et al., 2011). Based on this study and those mentioned above, it is possible that fine-tuning gene regulation of human-specific genes (Torgerson et al., 2009; Wagner & Lynch, 2008) may underlie many of the characteristics and behavior believed to be human-specific. Because of the strong association with the neocortex, it is likely that such traits could include sensory perception, motor skills, reasoning, conscious thought, and language. We propose to test this hypothesis by gathering more information on the role of young genes and regulation in the neocortex which will clarify the timing and causation of its evolution and what that means for “human specific� traits. Aims 1. Determine what young genes are upregulated in the neocortex of multiple primates at two developmental times, and identify which are primate-specific vs. species-specific. I. From three separate human, marmoset (New World), and rhesus monkey (Old World) individuals per fetal stage, brain matter will be extracted from the neocortex and non-neocortical regions twice before birth, and once immediately after birth. Because human samples will be obtained from aborted fetuses, times of primate sample collection are adjusted accordingly to gestation period: about 4 and 7 weeks for marmoset, about 4 and 8 weeks for rhesus, and about 7 and 15 weeks for human. Using Tag-Seq methods as established by (Babbitt et al., 2010), changes in global transcript abundance in the neocortex and non-neocortical regions will be collected; tag sequencing functions by capturing RNA by their poly(A) tails, producing short tags from the 3’ ends of mRNA molecules, and then sequencing them on an Illumina platform (Morrissy et al., 2010). From this data, comparisons of the
NEURO
Letters
Genesis
neocortical and non-neocortical regions will be made to determine which genes are upregulated specifically in the neocortex and to identify those genes. Time-wise comparisons of such upregulated neocortical genes will determine the amount of young genes differentially expressed between different developmental stages which may be notably different between primates. Tag-seq also allows for the analysis of noncoding RNA which will be examined in the same way. II. Additionally, chimpanzee brain matter will be obtained by opportunity from the Southwest Foundation for Biomedical Research and/or the New England Primate Center in accordance with their protection laws. If samples are unavailable, data may be used from previous arrays (Babbitt et al., 2010). However these samples will likely not be from the desired time period (about 6 and 12 weeks), a limitation to their inclusion. III. Aging of genes will be accomplished by aligning orthologous syntenic regions across the vertebrate phylogeny to detect in which branch a new gene arose, as established by (Zhang et al., 2011). Those genes found to be upregulated will then be classified as “old”- prior to the primate clade or “young”-since the primate clade. “Young” will then be broken down into primatespecific or species-specific categories. IV. We expect there will be similar upregulation of genes in the fetal brains, but that there will be less in nonhuman primates because, as previous studies have shown, the amount of upregulation within the human brain involves genes that have evolved at various points in the primate line. We also expect that humans will have the highest proportion of both upregulated primatespecific and species-specific genes as we favor the hypothesis that young genes are responsible for much of the development and function of the human neocortex. 2. Determine which of these upregulated young genes are related to transcription factors. I. Using InterPro Database, identify probable protein functions of those genes (Hunter et al., 2009). II. Compare upregulated non-genic areas to ENCODE database.
III. We expect that a lower percentage of the upregulated genes will be related to transcription factors, compared with the human sample, because we favor the hypothesis that differential regulation by human specific genes is a large factor in determining human specific traits, especially in relation to the brain. 3. Determine where in the primate line accelerated origination of certain development genes occurred in the neocortex and if there is any correlation with transcription factors. I. Based on the information gathered from the above aims, a pattern of upregulation of young primate-specific genes in the neocortex will signal recruitment within the primate lineage, but it is also possible that much of this process may have occurred further down the phylogeny, in other vertebrates. Any significant differences between the primate samples will also help to narrow the window of recruitment of specific young genes based on their function, which will highlight the differences between human and primate brain development. We hope that this study will then be used as a platform for future research on human brain function. Conclusion While research of human-specific traits, in particular cognitive ability, is largely popular, we believe that the proposal presented here provides a critical foundation for such topics to advance. With the improvements of sequencing technologies, both in cost and assay type, we are now able to conduct research on the most basic level of what sets humans apart from other species. As proposed, primate comparisons of phylogenetically young genes will provide insight into differences in regulation and transcription within the human brain, a relatively new and unique concept in understanding human evolution. Babbitt, C. C., Fedrigo, O., Pfefferle, A. D., Boyle, A. P., Horvath, J. E., Furey, T. S., & Wray, G. A. (2010). Both noncoding and protein-coding RNAs contribute to gene expression evolution in the primate brain. Genome Biol Evol, 2, 67-79. doi: 10.1093/gbe/evq002 Calarco, J. A., Saltzman, A. L., Ip, J. Y., & Blencowe, B. J. (2007). Technolo gies for the global discovery and analysis of alternative splicing. Adv Exp Med Biol, 623, 64-84. Dehay, C., & Kennedy, H. (2009). Transcriptional regulation and alterna tive splicing make for better brains. Neuron, 62(4), 455-457. doi: 10.1016/j.neuron.2009.05.006 Hunter, S., Apweiler, R., Attwood, T. K., Bairoch, A., Bateman, A., Binns, D., Yeats, C. (2009). InterPro: the integrative protein signature database. Nucleic Acids Res, 37(Database issue), D211-215. doi: 10.1093/nar/gkn785 Johnson, M. B., Kawasawa, Y. I., Mason, C. E., Krsnik, Z., Coppola, G., Bogdanovic, D., . . . Sestan, N. (2009). Functional and evolutionary
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 9
NEURO Genesis
insights into human brain development through global transcriptome analysis. Neuron, 62(4), 494-509. doi: 10.1016/j.neuron.2009.03.027 Morrissy, S., Zhao, Y., Delaney, A., Asano, J., Dhalla, N., Li, I., Marra, M. (2010). Digital gene expression by tag sequencing on the illumina genome analyzer. Curr Protoc Hum Genet, Chapter 11, Unit 11 11 1136. doi: 10.1002/0471142905.hg1111s65 Stahl, P. D., & Wainszelbaum, M. J. (2009). Human-specific genes may of fer a unique window into human cell signaling. Sci Signal, 2(89), pe59. doi: 10.1126/scisignal.289pe59 Torgerson, D. G., Boyko, A. R., Hernandez, R. D., Indap, A., Hu, X., White, T. J., Clark, A. G. (2009). Evolutionary processes acting on candidate cis-regulatory regions in humans inferred from patterns of polymorphism and divergence. PLoS Genet, 5(8), e1000592. doi: 10.1371/journal.pgen.1000592 Wagner, G. P., & Lynch, V. J. (2008). The gene regulatory logic of transcrip tion factor evolution. Trends Ecol Evol, 23(7), 377-385. doi: 10.1016/j. tree.2008.03.006 Yeo, G., Holste, D., Kreiman, G., & Burge, C. B. (2004). Variation in alternative splicing across human tissues. Genome Biol, 5(10), R74. doi: 10.1186/gb-2004-5-10-r74 Zhang, Y. E., Landback, P., Vibranovski, M. D., & Long, M. (2011). Ac celerated recruitment of new brain development genes into the human genome. PLoS Biol, 9(10), e1001179. doi: 10.1371/journal pbio.1001179
10 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Letters
NEURO
Opinion
Genesis
Creativity and the default network: Is the brain at work while the mind is wandering? Elizabeth Beam1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Elizabeth Beam (elizabeth.beam@duke.edu) 1
SUMMARY: Most of us have had the experience of trying to suppress off-task thoughts while writing a paper or working out the solution to a difficult math problem. Contrary to our intuition, research suggests that mind wandering may be an important part of the creative problem solving process. The default mode network (DMN), comprised of regions in the brain that are active when an individual is not engaged in a task, has been implicated in a number of cognitive processes that underlie creativity. During wakeful rest, the DMN functions to generate spontaneous thoughts; this is why it may be such a challenge to keep from daydreaming during class or work, and it is how the DMN facilitates the most basic “creating” element of creativity. Nonetheless, idea generation does not necessarily lead to creative innovation. Ideas must also be novel and useful. In these aspects, too, the DMN may contribute to creativity by making it possible to switch between divergent thoughts and to test out potential solutions in imagined scenarios. Bringing together studies of problem solving and the DMN, there is evidence to suggest that productivity may be enhanced by giving the mind room to wander. struct fictional narratives (Andreasen et al., 1995). These The default network are the same sorts of cognitive processes that are actively By “default” during wakeful rest, the brain initiates initiated during the brainstorming phase of creative a range of diverse processes that underlie daydreams: problem solving. To stimulate the flow of potential soluretrieving memories, planning for the future, assuming tions, individuals may think as if they were at rest while the perspectives of others, and exploring hypothetithey are at work on a problem. Furthermore, because cal scenarios (Buckner et al., 2008). The default mode the DMN is capable of activating thoughts that are not network (DMN) is identified on functional brain scans as deliberately called upon, it has the potential to draw one’s the regions that are more active when an individual is not attention to novel ideas that we may otherwise never performing a task, engaging instead in internally guided have reached by logical reasoning. “default mode” cognition. The DMN components are Unfortunately, the flow of idea generation is not distributed throughout the brain in the frontal, temporal, always easy to maintain, and it is often curtailed during and parietal lobes (Figure 1). problem solving when one “gets stuck” at an impasse. Likewise, the experience of “block” is common among Idea generation creative writers. Problem solvers have thus found it useful What is remarkable about the DMN is its automato take a break (Sio & Ormerod, 2009). When they reticity (Andrews-Hanna et al., 2010). Without exerting turn to the problem, they experience an increase in both any effort, individuals in the resting state may relive past problem-solving speed and in the likelihood of arriving experiences, imagine the experiences of others, and conat an insightful solution (Sio & Ormerod, 2009). The
Figure 1: Brain slices showing DMN components distributed in the frontal, temporal, and parietal lobes.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 11
NEURO Genesis
incubation period is especially productive when individuals are engaged in a task with low processing demands, stimulating the DMN and allowing for the generation of new ideas under relaxed constraints. You have probably observed that thoughts flow more easily to some of your peers than others, while on your own, you may have experienced fluctuations between creative surges and blocks. To understand how the brain mediates this variation, neuroscientists have studied differences in DMN generativity across the population and within individuals. They have found that those with higher DMN activation during rest also report a higher frequency of spontaneous thoughts (Razumnikova, 2007). Patients who suffer dysfunction of the frontotemporal network—which includes medial prefrontal and medial temporal regions that overlap with the DMN—may exhibit compulsive idea generation that varies with the symptoms of their disorder (Flaherty, 2005). DMN abnormalities have been observed in bipolar disorder patients (Gyulai, 1997; Ongur, 2010), who show fluctuations in mood that correspond with changes in fluency of thinking. Although idea generation is stunted during an individual’s depressive phases, their manic phases are characterized by what the DSM-IV calls “flight of ideas” and “pressured speech” (American Psychiatric Association, 2000). In some individuals resembling Wernicke’s aphasics, compulsive idea generation may be expressed as “hypergraphia,” or the drive to write (Waxman & Geschwind, 1974; Flaherty, 2005). This condition is often co-morbid with the manic phase of bipolar disorder, implying that tonic increases in mood valence are correlated with linguistic creativity. The coincidence of hypergraphia and mania has long been suspected in renowned writers, such as George Gordon, Lord Byron ( Jamison, 1989; Jamison, 1993). Future research on individuals who exhibit both exceptional creativity and enhanced creative drive is necessary to learn more about the DMN abnormalities that could confer success in long-term creative work. Flexible thinking Brain imaging studies of resting state cognition have revealed a multitude of underlying processes mediated by regions throughout the brain. Prefrontal and posterior cingulate regions direct one’s wandering mind toward thoughts of oneself or others. The medial prefrontal cortex (MPFC) has been implicated in self-reflection, including judgments that integrate preferences with emotional and social information (Gusnard et al., 2001; Kelley et al., 2002; Mitchell et al., 2006). The posterior cingulate cortex (PCC), which is often active in tasknegative thought, has also been implicated in theory-ofmind tasks that require an individual to take on another person’s point of view (Saxe & Powell, 2006; Rilling et 12 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Opinion al., 2004). Intuitively, the brain should be involved in self-oriented thought while one is not engaged in the external world. Less easily explained is why individuals retrieve memories and simulate imagined events in rich detail when they lack extrinsic or intrinsic motivation to do so. The medial temporal lobe (MTL), a critical component of learning and memory, is the origin of this tendency (Andreasen et al., 1995; Buckner et al., 2005). Anyone who has had the experience of daydreaming can attest to the associational way in which memories are called to mind, which may be why the precuneus—often referred to as the “association cortex”—is active in resting state cognition (Fransson & Marrelec, 2008). It is intriguing, and especially relevant to a study of creative problem-solving, that regions so diverse in anatomical localization and cognitive functioning are activated together when an individual is at rest. The type of thought that characterizes default mode cognition is divergent and flexible, facilitating connections between disparate concepts. For creative problem-solving, this means that engaging the DMN may help individuals to restructure the problem, integrate problem constraints with prior knowledge, and reconsider the problem from alternate perspectives. Electroencephalography (EEG) data from individuals using different strategies to solve anagrams lends support to this theory: those who come upon insightful solutions tend to engage parietal association cortices more so than individuals who use logical reasoning methods (Kounios et al., 2008). What individuals may find most helpful is to take a break from the problem at hand and occupy themselves with an unrelated activity. A fascinating behavioral result that supports this theory comes from a study of decisionmaking among consumer products (Dijksterhuis et al., 2006). Experimenters presented individuals with an array of simple options (between shampoo brands, for example) and complex options (between cars). Only some individuals were permitted to consciously deliberate their choices; others were distracted for the same period of time. When choosing between shampoos, individuals who deliberated over their choice later reported that they were more satisfied. However, when choosing between cars that varied across 12 dimensions, individuals made normatively and subjectively better choices when they were not permitted to think consciously about their choice. While analytical thinking is adequate for making simple calculations, DMN processing during unconscious deliberation enables one to integrate large amounts of complex information. This finding has been replicated in behavioral research on creative problem solving. An analysis across the aggregate of studies on incubation reveals that breaks are most beneficial when they require subjects to perform tasks with low processing demands (Sio & Ormerod, 2009). While
NEURO
Opinion unoccupied rest permits individuals to ruminate on the problem under their initial assumptions, daydreaming while working on a new and dissimilar task enables the mind to process the problem under looser constraints than the analytical processes utilized during conscious problem solving. In the future, neuroimaging during the incubation period may reveal how DMN regions interact to yield creative solutions. Conclusion Creative solutions must be both novel and useful, and in these two regards, the DMN accomplishes creative work while individuals are at rest. Allowing the mind to generate new ideas during unstructured brainstorming may turn up solutions that others would not have considered under the constraints of analytical thinking, while promoting cognitive flexibility throughout the problem solving process prevents fixation on initial assumptions and timeworn approaches to the problem type. Paradoxically, taking a break to engage the DMN may be the most productive way to tackle problems that require creative thinking.
Genesis
writers and artists. Psychiatry, 52, 125-134. Jamison, K. R. (1993). Touched with fire: Manic-depressive illness and the artistic temperament. New York, NY: Free Press Paperbacks. Kelley, W. M., Macrae, C. N., Wyland, C. L., Caglar, S., Inati, S., & Heath erton, T. F. (2002). Finding the self ? An event-related fMRI study. Journal of Cognitive Neuroscience, 14, 785-794. Kounios, J., Fleck, J. I., Green, D. L., Payne, L., Stevenson, J. L., Bowden, E. M., & Jung-Beeman, M. (2008) The origins of insight in resting-state brain activity. Neuropsychologia, 46, 281-291. Mitchell, J. P., Macrae, C. N., Banaji, M. R. (2006). Dissociable medial pre frontal contributions to judgments of similar and dissimilar others. Neuron, 50, 655-663. Razumnikova, O. M. (2007). Creativity related cortex activity in the remote associates task. Brain Research Bulletin,73, 96-102. Rilling, J. K., Sanfey, A. G., Aronson, J. A., Nystrom, L. E., & Cohen, J. D. (2004). The neural correlates of theory of mind within interpersonal interactions. Neuroimage, 22, 1694-1703. Saxe, R., & Powell, L. J. (200^). It’s the thought that counts: specific brain regions for one component of theory of mind. Psychological Science, 17, 692-699. Sio, U. N.,& Ormerod, T. C. (2009). Does incubation enhance problem solving? A meta-analytic review. Psychology Bulletin,135, 94-120. Waxman, S. G., & Geschwind, N. (1974). Hypergraphia in temporal lobe epilepsy. Neurology, 24, 629-636.
Acknowledgements Many of the ideas in this paper were inspired by conversations with Scott Huettel, Duke University, Department of Psychology & Neuroscience. I also thank David V. Smith, Duke University, for providing the figure of the DMN. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders,(IV). Washington, DC. Andreasen, N. C., O’Leary, D. S., Cizadlo, T., Arndt, S., Rezai, K. et al. (1995). Remembering the past: two facets of episodic memory explored with positron emission tomography. American Journal of Psychiatry, 152, 1576-1585. Andrews-Hanna, J. R., Reidler, J. S., Huang, C., & Buckner, R. L. (2010). Evidence for the default network’s role in spontaneous cognition, Jour nal of Neurophysiology, 104, 322-335. Buckner, R. L., Raichle, M. E., Miezin, F. M., & Petersen, S. E. (1996). Functional anatomic studies of memory retrieval for auditory words and pictures. Journal of Neuroscience, 16, 6219-6235. Buckner, R. L., Andrews-Hanna, J. R., Schacter, & D. L. (2008) The brain’s default network: anatomy, function, and relevance to disease. Annals of the New York Academy of Sciences, 1124, 1-38. Dijksterhuis, A., Bos, M. W., Nordgren, L. F.,& van Baaren, R. B. (2006). On making the right choice: the deliberation-without-attention effect. Science,311, 1005-1007. Flaherty, A. W. (2005). Frontotemporal and dopaminergic control of idea generation and creative drive. The Journal of Comparative Neurology, 493, 147-153. Fransson, P., & Marrelec, G. (2008). The precuneus/posterior cingulate cor tex plays a pivotal role in the default mode network: evidence from a partial correlation network analysis. Neuroimage, 42, 1178-1184. Gusnard, D. A., Akbudak, E., Shulman, G. L., & Raichle, M. E. (2001). Medial prefrontal cortex and self-referential mental activity: relation to a default mode of brain function. PNAS, 98, 4259-64. Hermann, E., Call, J., Hernandez-Lloreda, M., Hare, B., & Tomasello, M. (2007). Humans have evolved specialized skills of social cognition: the cultural intelligence hypothesis. Science, 317(5843), 1360-1366. Jamison, K. R. (1989). Mood disorders and patterns of creativity in British
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 13
NEURO
Article
Genesis
Learning and non-learning effects of Ginkgo biloba extract EGb 761 in Aplysia californica Mary Petrosko1 and Robert Calin-Jageman2 1 Department of Psychology, Dominican University Correspondance should be addressed to Mary Petrosko (petr0308@umn.edu) 2 Department of Neuroscience, Dominican University
SUMMARY: Existing research about Ginkgo’s ability to improve memory is mixed, and the neural mechanism by which it may affect behavior has yet to be determined. To explore claims of efficacy, the behavioral and physiological effects of Ginkgo extract EGb 761 (Schwabe Pharmaceuticals) were observed here in Aplysia californica. The effects of EGb 761 were measured on spontaneous activity, reflex sensitivity, rates of learning and forgetting in a long-term habituation paradigm and subsequently in spontaneous and siphon-evoked central nervous system (CNS) activity. Effects were measured after acute and long-term exposure to EGb 761. No significant effects of EGb 761 were observed on any of the dependent measures, suggesting that the EBg 761 compound is not bioactive in the Aplysia CNS. These findings suggest the following: 1) EGb 761’s mechanism of action is outside of the CNS (e.g. cerebral blood flow); 2) its mechanism of action is not conserved across the animal kingdom; or 3) that EGb 761 is not bioactive.
Supplements to improve learning and memory abound in supermarkets and health food stores. These supplements are not regulated by the Food and Drug Administration and do not undergo testing to verify their efficacy. Rigorous and controlled testing must to be conducted to determine the efficacy and safety of the supplements. A popular learning and memory supplement is Ginkgo biloba, which has been used historically for medicinal purposes. The patented extract EGb 761, marketed by Schwabe Pharmaceuticals (Karlsruhe, Germany), contains a formula of 24% flavones and 6% terpenoides derived from the ginkgo plant (DeFeudis & Drieu, 2000). This formula has been marketed as the most ‘effective’ extract of ginkgo (Muller & Chatterjee, 2003). The purported memory effects of ginkgo have been supported by the argument that it increases circulation in the brain and has antioxidant properties (DeFeudis & Drieu, 2000; Gold, Cahill & Wenk, 2003; Petkov, Belcheva & Petkov, 2003). The formula of flavones and terpenoides in EGb 761 are thought to enhance the antioxidant and circulatory properties of the extract. For the purposes of this paper, “EGb 761” will identify the ginkgo extract patented by Schwabe Pharma. Any other extract of ginkgo will be referred to as “ginkgo extract.” Evidence of efficacy The learning and memory effects of ginkgo have been investigated in humans and other mammals, with most studies focusing on the effect of EGb 761. Despite several studies investigating the drug, the literature presents mixed conclusions regarding the efficacy of gingko supplements. Evidence does exist for EGb 761 effectiveness in alleviating learning and memory deficits in human and animal subjects (Kanowshi, Hermann, Stephan, Wierich & Horr, 1996; Lin, Cheng, Hsu & Chang, 2003; 14 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Tchantchau, Xu, Wu, Christen & Luo, 2007; Wang, Wang, Wu & Cai, 2006); on the other hand, some studies report opposing results (Birks & Grimley, 2009; Dekosky et al., 2008). However, there is consensus regarding EGb 761 efficacy in subjects with normally functioning brains; in these populations, EGb 761 provides no improvement in memory (Burns, Bryan, & Nettelbeck, 2006; DeKosky et al., 2008; Solomon et al., 2002). The learning and memory effects of ginkgo have been investigated in humans and other mammals, with most studies focusing on the effect of EGb 761. Despite several studies investigating the drug, the literature is mixed regarding its efficacy. Studies in human and animal subjects with brain abnormalities has supported EGb 761’s effectiveness in alleviating deficits in learning and memory (Kanowshi, Hermann, Stephan, Wierich & Horr, 1996; Lin, Cheng, Hsu & Chang, 2003; Tchantchau, Xu, Wu, Christen & Luo, 2007; Wang, Wang, Wu & Cai, 2006); however, other studies have failed to support this effect (Birks & Grimley, 2009; Dekosky et al., 2008). In normally functioning brains, however, most research indicates that EGb 761 provides patients with no clear improvement in memory (Burns, Bryan, & Nettelbeck, 2006; DeKosky et al., 2008; Solomon et al., 2002). In rodent animal models, ginkgo has demonstrated selective influence on learning and memory. In studies with older or impaired rats, EGb 761 has been shown to be relatively effective in alleviating learning deficits (Kanowshi et al., 1996; Lin et al., 2003; Tchantchau et al., 2007; Wang et al., 2006). Lin et al. (2003) examined the effect of ginkgo in rats with bilateral common carotid artery ligation, which produced chronic cerebral insufficiency. EGb 761 improved impairments in spatial learning and motor function more effectively than an extract of local ginkgo. However, it is important to note
Article that performance did not return to normal function, and impairments in spatial learning persisted beyond EGb 761 treatment . Wang et al. (2006) found that EGb 761 administered to older rats improved their spatial learning for a maze task. Furthermore, they found that hippocampal long-term potentiation (LTP) was increased in subjects treated with EGb 761. This increase in LTP may account for memory improvements in treated subjects, as the hippocampus is critical in forming new memories and spatial learning. Research on gingko efficacy in humans with cognitive deficits is less conclusive. In a large, double-blind, placebo-controlled study conducted by Le Bars et al. (1997), a slight but consistent memory improvement was observed in participants with dementia on gingko supplementation. Participants treated with EGb 761 performed better on the Alzheimer’s Disease Assessment Scale-Cognitive subscale and on the Geriatric Evaluation by Relative’s Rating Instrument. However, the Ginkgo Evaluation of Memory (GEM) by DeKosky et al. (2008) found that EGb 761 failed to reduce the incidence of dementia in older adults (≥75 years old). GEM is a large and wellcontrolled study supported by the National Institutes of Health to test the efficacy of EGb 761 as a preventative treatment for Alzheimer’s disease. The study tracked participants for an average of five years and found that EGb 761 failed to prevent or delay the incident of dementia in older adults. Based on these conflicting results, EGb 761 may work through a mechanism which improves memory in those who have developed dementia but is not involved with pre-dementia memory mechanisms. While results are contradictory in impaired populations, similar studies with normal populations have shown little overall effectiveness of the ginkgo extract. Solomon, Adams, Silver, Zimmer, and DeVeaux (2002) examined the effect of ginkgo extract in healthy individuals aged 60 years or older during six weeks of supplementation. No measurable learning, memory, or attentional improvements were found for treated individuals. Similarly, no significant learning or memory improvement occurred for a younger population of participants--aged 18 to 43 years--observed across twelve weeks of gingko extract supplementation (Burns, Bryan, & Nettelbeck, 2006). In addition, the studies that do report disparate positive effects of EGb 761 in healthy adults, their conclusions may be statistically unreliable because of small sample sizes and unclear outcome measures (Birks & Grimley, 2009; Gertz & Kiefer, 2004). Outstanding questions Overall, the existing evidence suggests that EGb 761 may have neuroprotective qualities but that it does not actively improve learning or memory in normally functioning brains. One limitation with existing research,
NEURO Genesis
however, is the focus on a single indicator of efficacy: long-term recall measured at a single time point. This measure of memory and learning function may not be sensitive to changes in a normal population. To more fully evaluate the effects of EGb 761 in normal populations, it may be useful to collect a broad range of learning indicators, including rate and depth of learning, shortterm memory performance, and the duration of longterm memory. Another limitation in the study of gingko extract is our inability to identify biological mechanisms, which may be linked to the behavioral effects of EGb 761. One suggested mechanism by which EGb 761 may operate lies outside of the CNS; EBg 761 may improve blood circulation and increase the flow of oxygen to the brain (DeFeudis & Drieu, 2000). Other evidence points to the direct effects of EGb 761 in the CNS; older rats treated with EGb 761 exhibited increased LTP in vivo (Wang et al., 2006). Similarly, mice brains with applied EGb 761 exhibited increased LTP and neuronal excitatibility in vitro (Williams, Watanabe, Schultz, Rimbach, and Krucker, 2004). Additional studies have also found that EGb 761 increases neuronal plasticity, and protects against neuronal apoptosis in cell cultures (Defeudis & Drieu, 2000; Luo et al., 2002; Smith et al., 2002). Despite these findings, researchers have not yet explored the neural mechanisms by which EGb 761 influences behavior. Aplysia as a model system The marine gastropod, Aplysia californica (Figure 1), is a useful model organism in which to evaluate the effects of EGb 761. Aplysia provide sensitive measurements of the effects of EGb 761 effects in a biological system that can more easily be connected to behavioral measures. Aplysia have long proven useful for identifying basic mechanisms of learning and memory. One attractive feature is that Aplysia have a relatively small number of nerve cells, numbering less than 10,000 in its CNS (Kandel, 2001). Moreover, Aplysia have large nerve cells (up to 1 mm in diameter), which are both easy to monitor and manipulate (Moroz et al., 2006). These features make Aplysia uniquely useful for relating behavioral and neural function. As a result of their relatively simple neural circuitry, the sensory and motor neural circuit has been identified and studied extensively, relating its activity to the behavior it elicits (Kandel, 2001). In addition, the simple neural circuitry allows for extracellular and intracellular recording, with which neural activity can be analyzed and related to behavioral responses. Aplysia have been used to understand basic physiological mechanisms of habituation, sensitization, operant conditioning, place conditioning, and other forms of learning and memory (Brembs, 2003; Glanzman, 2006; Rankin, 2002; Stopfer & Carew, 1996). Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 15
NEURO
Article
Genesis
Stimulus 30 stimuli, 1 min ISI 1 day
Pre-test
Training
Post-test
Figure 1: Procedure for Experiment 1, to elicit long-term habituation of S-SWR. Animals were assessed for baseline responding with the pretest, 3 stimuli administered at 10 minute ISI. Animals were then trained in 5 sessions, each session consisting of 30 stimuli at 1 minute ISI, with 24 hours between each session. After training, short- and long-term retention post-testing were administered in the same way as the pretest. Post-testing was administered 1 hour after the last training session, then once a day for the subsequent 6 days.
Current study To address the diverse and discordant literature on EGb 761, this study used a three-part approach, which explored possible learning, arousal, and physiological effects of EGb 761 on acute and long-term exposure in Aplysia. The three experiments included measures on multiple time scales and explored possible effects on different behavioral and physiological outcome measures. The combined results offer a sensitive and inclusive analysis of possible behavioral and physiological effects of EGb 761. Effect of EGb 761 on long-term learning The first experiment examined the effect of EGb 761 on long-term habituation after long-term exposure. Habituation is a commonly used model of learning in Aplysia; it is a type of learning in which a subject’s response to a stimulus decreases as a result of repeated presentation of the stimulus. Although simple, habituation shares many attributes with more complex learning mechanisms (Carew, Pinsker, & Kandel, 1972; Stopfer, Chen, Tai, Huang, & Carew, 1996; Zolman & Peretz, 1987). Like more advanced learning mechanisms, habituation response rates can improve with training, are sensitive to the pattern and frequency of training, and can produce both short- and long-term memories (Abramson, 1994). A benefit of studying habituation is that it provides multiple indicators of learning and memory. This first experiment examined four separate indicators: depth of learning, rate of learning, short-term memory retention, and long-term memory retention. The behavior used for habituation training in this experiment was the siphon-elicited siphon withdrawal reflex (S-SWR). The S-SWR is a simple reflex in which the siphon is withdrawn into the mantle after mild stimulation to the siphon (Carew et al., 1972). The S-SWR 16 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
is easily observable and is easily quantified as the time from when the animal withdraws its siphon to the time it begins to relax the siphon back to the resting position. In the S-SWR, habituation can be elicited by repeatedly stimulating the siphon, producing a gradual decrease in the duration of the SWR. Depending on the pattern and duration of stimulation, it is possible to produce both short- and long-term habituation (Carew et al., 1972; Stopfer et al., 1996; Zolman & Peretz, 1987). This experiment explored the possible effects of long-term EGb 761 exposure on four separate learning and memory indicators of long-term habituation of the S-SWR. Effect of EGb 761 on arousal after acute exposure The second experiment explored the effect of EGb 761 on arousal and gross motor movement after acute exposure to EGb 761. Past research suggests that significant effects can be elicited from manipulations other than long-term gingko exposure; a single dose of EGb 761 before the administration of a narcotic resulted in increased arousal in treated animals (Brochet, Chermat, DeFeudis, & Drieu, 1999). Additional studies in humans have shown a slight improvement in reaction times on short-term memory tasks and sustained attention following acute exposure to EGb 761 (Elsabagh, Hartley, Ali, Williamson, & File, 1995; Kennedy, Scholey, & Wesnes, 2000; Rigney, Kimber, & Hindmarch, 1999). To test possible arousal effects of EGb 761, the second experiment examined behavioral measures of locomotion and S-SWR after acute exposure to EGb 761. Previous research in Aplysia has used both locomotion and SSWR as indicators of arousal after exposure to a treatment (Marinesco, Wichremasinghe, Kolkman, & Carew, 2004). Together, these measures serve as sensitive indicators of acute animal arousal.
NEURO
Article
Genesis
Effect of EGb 761 on CNS activity The third experiment examined the physiological affect of EGb 761 on background and nerve-evoked neural activity in the CNS after acute exposure. An investigation of the possible biological mechanism by which EGb 761 affects the organism followed. Research with rats and mice has shown that EGb 761 can increase hippocampal activity and LTP when administered in vivo and in vitro (Lanahan, Lyford, Stevenson, Worley, & Barnes, 1997; Wang et al., 2006; Williams et al., 2004). In vitro studies have shown an increase in neuronal excitability and synaptic transmission when EGb 761 was applied to mice brain slices (Williams et al., 2004). This evidence indicates a correlation between EGb 761 and the neural mechanisms of memory. As Aplysia offer a simple neural circuit from which the activity of the CNS can be easily monitored and quantified, this study directly tested the claim that EGb 761 affects the CNS. The three experiments performed here allowed for a multi-dimensional analysis of the effects of EGb 761 on multiple behavioral and physiological outcome measures; such a comprehensive evaluation of the possible effects of EGb 761 in a model system may provide a link between physiology and behavior. Experiment 1 A number of Aplysia behaviors have been analyzed at the neural level. This particular study explored the effect of EGb 761 on long-term habituation of the S-SWR. To produce long-term decreases in S-SWR behavior, Aplysia received five separate training sessions consisting of 30 stimuli applied to the siphon, each eliciting a S-SWR. Training sessions were administered daily for five days (Figure 2). This protocol has shown significant long-term habituation learning lasting up to 14 days after training (Stopfer et al., 1996).
Multiple indicators of learning and memory were assessed: depth of learning, rate of learning, short-term retention, and long-term retention. Depth of learning was measured as the change from baseline to minimum responding during training. Rate of learning indicated how fast the animal habituated and was measured as the number of trials during training to reach minimum responding. Short- and long-term retention was assessed by comparing each animal’s baseline responding to posttest responding, indicating the amount of learning retained after training. Together, these measures provide a sensitive composite of learning and memory outcomes. By using four separate learning measures, this study improves upon the standard reliance on one measure of memory at a single point in time and may reliably detect subtle changes in learning performance. Method Subjects We used 24 Aplysia californica. Animals (late juveniles, 75-80g) were purchased from the NIH Aplysia resource facility (Miami, FL) and housed in separate plastic containers in a 15° C tank of aerated artificial seawater (ASW). All animals were fed a diet of dried seaweed, delivered three times a week. Aplysia were maintained at a constant 12:12 light/dark schedule (light 6 am–6 pm, dark 6 pm – 6 am), and underwent treatment and training during light cycles. Drug administration Animals were randomly assigned to one of the three experimental groups (n=8): control, low dose and high dose. Animals were pre-treated with EGb 761 for ten days prior to training. The Gingko biloba supplement used was powdered EGb 761 containing 24% ginkgo flavone glycosides and 6% terpene lactones (Schwabe Pharmaceuticals, Germany). The supplement was dissolved
8 High
7
Low Control
Seconds
6 5 4 3 2 1 0
BL TR 1 TR 2 TR 3 TR 4 TR 5 Figure 2: Long-term habituation learning curve data between experimental groups. S-SWR (s) throughout habituation training. Significant effect of habituation training, F(4, 84) = 8.02, p < .01. No significant effect of ginkgo on the rate or magnitude of learning (F < 1).
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 17
NEURO
Article
Genesis
in an incubation chamber containing 2 liters of ASW. Treated animals were soaked in the incubation chamber for 1 hour each day; control animals were soaked in untreated ASW. Animals were housed and trained in their home tank throughout the experiment and were not handled at any time. For the low-dose group, EGb 761 was dissolved in the incubation tank in quantities that achieve a 0.17% solution. When equilibriated, each tank provided a peak dose of 0.133 mg for each 75g animal, or 1.77 mg/kg. The high-dose group received twice the low-dose exposure: incubation in a 0.35% solution providing a peak dose of 3.54 mg/kg. The solution was mixed by an outside researcher, and the researcher responsible for habituating and timing responses was blind to the experimental groups. S-SWR response The S-SWR behavior was quantified as the duration of the withdrawal reflex. A researcher manually stimulated the Aplysia by placing a thin, plastic stimulator against the inner wall of the siphon. The stimulator is then drawn up along the siphon rapidly to elicit the SWR in the mantle shelf. S-SWR behavior was measured in seconds by the researcher after administering the stimulation. Timing began once the siphon was contacted and ended once the siphon began to relax to its normal position. Habituation training Baseline responsiveness was determined as the mean response to three stimuli administered at a ten minute interstimulus interval (ISI). A ten minute ISI does not produce habituation. Habituation training consisted of five sessions of 30 stimuli administered at a 1 minute ISI with a 24-hour interval between training sessions (Figure 1). Memory tests After habituation training, animals were tested for both short- and long-term habituation. Short-term testing occurred one hour after the final training session 7 6
Seconds Âą SEM
5 4 3 2 1 0
1 hr
1 Day
2 Day
3 Day
4 Day
5 Day
6 Day
Figure 3: Forgetting curve for 6 days post habituation training between experimental groups. Average S-SWR response between groups. Significant effect of forgetting, F(5, 105) = 4.38, p < .01. No significant effect of ginkgo on the rate or magnitude of forgetting (F < 1).
18 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
on the fifth day of training. Long-term testing occurred once a day for six days following the final training day. Each memory test consisted of three siphon stimulations with a 10 minute ISI. Statistical approach Data were analyzed using the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA). The basic strategy was to compare task performance between treatment groups (control, low, and high). Univariate analyses of variance (ANOVA) assessed for group differences. Alpha levels below 0.05 were considered statistically significant. Results Baseline responding No significant effect of EGb 761 on baseline responding was found. After ten days of pre-treatment, baseline responding was characterized by timing three elicited S-SWRs at ten minute intervals in all three treatment groups (control, low, and high dose of EGb 761). S-SWR responsiveness was similar across treatment groups: control (M = 6.41, SD = 3.72), low (M = 7.03, SD = 2.72), and high (M = 7.79, SD = 4.60) treatment level (Figure 3). This result was confirmed with a one-way betweengroups analysis of variance (ANOVA); no significant main effect of treatment condition (F(2, 21) = .27, p > .05) was observed. Repeated exposure to EGb 761 did not affect baseline S-SWR. Any difference in learning or memory measures between experimental groups would not be due to unequal initial conditions between experimental groups, but instead due to experimental condition. Learning effects A mixed-factor ANOVA demonstrated that the means between the drug conditions were not significantly different, F(2, 21) = .97, p > .05, during any of the five training sessions (Figure 3). Therefore, animals in each of the three drug conditions acquired habituation at a similar rate, and EGb 761 did not affect rate of learning. A one-way between-groups ANOVA was conducted to evaluate the main effect of EGb 761 on average S-SWR at the end of last day of training. No significant differences were found between the conditions, F(2,21) = .97, p > .05 (Figure 2). This result is expected because no significant differences in rate of learning were found between the three conditions. All groups were given habituation training of 1 training session per day for 5 days (20 stimuli per session, 1 minute ISI). By the end of the training, S-SWR responsiveness had decreased substantially across all groups: control (M = 2.85, SD = .79), low (M = 2.92, SD = 1.55), and high (M = 2.92, SD = .51) treatment level (Figure 3). To evaluate the main effect of habituation training, average S-SWR across the last 5 trials in each training session were analyzed with a repeated measures ANOVA.
NEURO
Article
Genesis
Stimulus
30 min
ASW
1 day
Gingko
Figure 4: Procedure for experiment 2 to determine acute arousal effects of EGb 761. S-SWR measures were taken at the beginning and end of each session, 30 min ISI. Animals were exposed to EGb 761 during Days 3-4 of the experiment.
The ANOVA indicated a significant training effect, F(4, 84) = 8.02, p < .01. Follow-up polynomial contrasts indicated a significant linear effect, with response means decreasing over time, F(1, 21) = 14.64, p < .01. To more fully understand the shape of the learning curve, a variety of simple equations were fit to the training data. For this analysis, all trials were included and data for each drug condition was collapsed because there was no main effect of drug condition and no interaction between drug condition and training session (F(4, 84) = .77, p > .05). The data best fit to an exponential function, which accounted for 75% of the variance in the training data. The parameters for this function were well constrained by the data: at the end of training, the animals’ S-SWR was 51.81% of their baseline responding, 95% confidence between 49.74% - 53.88%. After the first 19.91 stimulations, animals’ response decreased 40.68% from baseline responding. This means the greatest amount of learning typically occurred during the first training session (which consisted of 20 stimulations). Memory Effects No significant group difference was found on S-SWR one hour after the last training session, F(2, 21) = .96, p > .05 (Figure 3). This result indicates that the drug condition did not affect short-term memory retention: control (M = 3.61, SD = 1.60), low (M = 3.23, SD = .27), and high (M = 3.77, SD = 1.31). Similarly, no significant group differences were found on S-SWR for the six days following the end of training, F(2, 21) = .88, p > .05 (Figure 3). These results failed to support an effect of EGb 761 on short and long-term memory. To evaluate habituation training forgetting, a one-way repeated measures ANOVA was conducted to evaluate all animals’ S-SWR for the six days after training ended. The ANOVA indicated a significant forgetting effect, F(5, 105) = 4.38, p < .01 (Figure 3). Therefore, animals displayed significant forgetting during post-testing. A two-way mixed factor ANOVA was conducted to evaluate a possible interaction between EGb 761 condition and the forgetting curve over the six days following training. The ANOVA failed to support a significant
ASW
interaction between EGb 761 and forgetting, F(5, 105) = .84, p > .05. Therefore, the results indicate that animals forgot training at a similar rate regardless of EGb 761 condition. Experiment 2 In addition to possible learning benefits, EGb 761 has been purported to have acute arousal effects. Research indicates that EGb 761 can have an arousal effect in animals and humans after acute exposure (Brochet et al., 1999; Elsabagh et at., 1995; Rigney et al., 1999). To better document this effect in a physiologically tractable system, this experiment examined the acute arousal effects of EGb 761 in Aplysia. Two measures of arousal were used: S-SWR sensitivity and locomotion. Method Experimental design Experiment 2 used a within subjects design with all animals (n = 8) undergoing the same experimental sequence (Figure 4). Animals were transferred to an experimental tank containing 2 liters ASW while remaining in their home containers and were not handled at any time. While in the experimental tank, animals were videotaped to analyze locomotion. Animals were taped once a day for 30 minutes on 6 consecutive days. Drug administration Days 1-2, animals were only exposed to ASW and were not exposed to EGb 761. Days 3-4, animals were exposed to twice the recommended dose of Ginkgo. EGb 761 was dissolved in the 2 liters of ASW in the experimental tanks to achieve a 0.35% solution providing a peak dose of 3.54 mg/kg, double the recommended daily dose (120 mg recommended daily dose, average adult weights 67.5 kg, providing a dose of 1.77 mg/kg). Days 5-6, animals were only exposed to ASW and were not exposed to EGb 761. S-SWR response S-SWR response was measured in seconds by the researcher after administering the stimulation manually, as in Experiment 1. Each animal’s S-SWR was recorded once at the beginning of videotaping and a second time at the end of taping, 30 minute inter-stimulus interval (ISI). Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 19
NEURO
Article
Genesis
Crawling The animals’ amount of crawling was analyzed using SwisTrack software (Lochmatter et al., 2008). SwisTrack provided the amount of pixels each animal moved during the 30 minute recording period by tracking their positions on an XY coordinate. Animal positions were calculated at an interval of 5 times per minute. The total number of pixels moved during recording was calculated using the Pythagorean Theorem. The distance moved in pixels was added together to generate a total crawl amount. The total pixels each animal moved was converted to the equivalent cm moved (Figure 5). The pixel to cm conversion was calculated using a calibration frame of the video; the amount of pixels generated was compared to the specified distance moved by a sample object. Results S-SWR response S-SWR was compared between EGb 761 exposed and non-exposed conditions. The within-subject control was highly reliable. S-SWR responses in each animal were very stable across days, alpha = 0.86 (Figure 6). A within-subjects ANOVA comparing the animals’ S-SWR behavior during pre-exposure, exposure, and postexposure days showed no significant effect of EGb 761: pre-exposure (M = 7.25, SD = 2.17), exposure (M = 6.44, SD = 2.39), and post-exposure (M = 7.01, SD = 2.16), F(2, 14) = .98, p > .05. Crawling The amount of crawling between the two conditions (non-exposed, EGb 761 exposed) was compared to detect a significant effect of ginkgo. A within subjects ANOVA comparing the animals’ total movement during treatment condition showed no significant effect of EGb 761on cm crawled: pre-exposure (M = 1441.28, SD = 1352.19), exposure (M = 2027.82, SD =1068), and post-exposure (M = 1632.22, SD = 1172.38), F(2, 14) = .57, p > .05 (Figure 7). However, the amount each animal crawled varied substantially between days, and the within-subject control was not reliable, alpha = 0.20 (Figure 7 and 8). Experiment 3 While EGb 761 failed to produce an effect in the Video Recording
1
3
Movement Tracking
2 1
2
3
4
4
Figure 5: Example SwisTrack locomotion tracking of 4 Aplysia during 30 min video recording session.
20 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
behavioral measures tested in Experiments 1 and 2, there is the possibility that it could have a direct effect on the CNS which is not manifested behaviorally. The biological mechanism by which EGb 761 is purported to operate is unclear. Some research indicates that it operates outside of the CNS, working through blood flow, etc. (DeFeudis, 2000). Other research, however, with other animal models and humans indicates that EGb 761 can increase neuronal activity and hippocampal function in subjects both in vivo and in vitro (Lanahan et al., 1997; Wang et al., 2006; Williams et al., 2004). A possible mechanism by which EGb 761 can affect the organism is through direct influence in the CNS. Neuronal activity and synaptic transmission has increased in some studies which look at the effect of EGb 761 when directly applied to the brain (Lanahan et al., 1997; Williams et. al., 2004). Experiment 3 tested the direct effects of EGb 761 on the spontaneous and nerve-evoked activity in the CNS. Nerve activity during three conditions (pre-exposure to EGb 761, exposure, and post-exposure) and was assessed by two measures. First, background nerve activity was measured as the normal spontaneous activity occurring in the CNS during the experiment. Second, nerve-evoked evoked activity was measured after an electrical shock was administered to a sensory nerve. The 1 minute preceding and following each nerve shock were excluded from background nerve analysis. Each minute segment was compared as a percentage of baseline responding (nerve activity during first 5 minutes of recording). Method Animal preparation Aplysia (n = 5) were anesthetized (50 mL MgCl) and dissected according to established protocol (Stopfer et al., 1996). The CNS (right and left P9 sensory nerves, ring ganglion, right and left connective nerves, abdominal ganglion, and siphon motor nerve) was removed from the animal and maintained in a 5 mL recording dish with profused ASW. The CNS was revived with ASW and rested 1 hour before recording. Electrophysiology Nerve activity was recorded extracellularly using suction electrodes with flexible Tygon tips. Nerve recordings were taken for the right P9, right connective, and siphon nerve in each animal (10hz high-pass filter, 3khz low-pass, 10k gain, A-M Systems AC/DC Differential Amplifier model 3000). Nerve recordings were quantified as the integral of the absolute value of the nerve trace (20 K/s sampling rate). Spontaneous background nerve activity Background nerve activity was calculated in one minute segments as the average of the nerve trace integral. The 1 minute proceeding and following each nerve shock were excluded from background nerve analysis.
NEURO
Article
Genesis
10
Seconds SEM
8
6
4
2
0
1
2
3
Day
4
5
6
Figure 6: S-SWR in experiment 2. Responses were stable (alpha = .86). No significant effect of EGb 761 on S-SWR (F < 1).
400
CM SEM
300
200
100
0
1
2
3
4
5
Figure 8: Example of SwisTrack locomotion tracking for one Aplysia across the six experimental days.
6
Day
Figure 7: Average locomotion (cm) by day for experiment 2. Responses were not stable across days (alpha = .20). No significant effect of EGb 761 on locomotion (F < 1).
10 min
ASW
Gingko
ASW
Figure 9: Procedure for Experiment 3. This is used to determine background and nerve-elicited nerve activity in the Aplysia central nervous system (right P9 sensory nerve, right connective nerve, siphon motor nerve).
200µV
Figure 10: Sample of spontaneous nerve activity in and out of EGb 176 bath application.
Nerve-elicited nerve activity During nerve-elicited nerve activity trials, an 80-90 µA pulse (0.25 s duration, biphasic) was administered to the left P9 sensory nerve (A-M Systems Isolated Pulse Stimulator model 2100). The stimulus artifact (0.5 s after administration of pulse) was subtracted out of nerve-elicited nerve activity analysis. Nerve-elicited nerve activity was calculated as the average of the nerve trace integral for the 0.3 s following the stimulus artifact (0.5 s – 0.8 s after pulse administration). This protocol was validated by comparing SWR with nerve activity in a set of reduced siphon+tail preperations; the average of the nerve trace integral for the 0.3 s following the stimulus artifact (0.5 s – 0.8 s after pulse administration) was highly correlated with behavioral SWR response. Average nerve response after each pulse was compared as a percentage of baseline responding (nerve activity during first 2 pulses). Drug administration During the drug exposure condition, EGb 761 was added to the ASW in the recording dish (5µL EGb 761/5 mL recording dish capacity).
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 21
NEURO
Article
Genesis
Experimental design Recordings were taken for the right P9 sensory nerve, right connective nerve, and siphon motor nerve continuously for 90 minutes. Ten pulses (80-90 µA) were administered to the left P9 sensory nerve throughout recording 10 minute ISI (Figure 9). For minutes 0-19 the CNS was exposed to ASW, establishing baseline background neural activity and nerve-elicited activity. During minutes 19-51 EGb 761 was directly applied to the CNS. After the 30 minutes of exposure, the profusion tube was reopened, washing out the EGb 761. For minutes 51-90 post-exposure, the CNS was exposed to ASW during post-test background activity and nerve-elicited activity recording. Results Background nerve activity A one-way within-subjects ANOVA was conducted for each nerve (right P9, right connective, siphon), comparing each nerve’s background activity across treatment conditions (pre-exposure to EGb 761, exposure, postexposure). EGb 761 demonstrated no significant effect on total background nerve activity (Figure 10): right P9: F(2, 8) = .50, p > .05; right connective: F(2, 8) = 1.06, p > .05; siphon: F(2, 8) = 2.22, p > .05 (Figure 11). Therefore, background nerve activity was not significantly different on any of the three nerves between the ASW and EGb 761 conditions. A one-way within-subjects ANOVA comparing each nerve’s nerve-elicited activity (right P9, right connective, siphon) across treatment conditions (pre-exposure to EGb 761, exposure, post-exposure) showed no significant effect of EGb 761 on nerve-elicited activity (Figure 12): right P9: F(2, 8) = .88, p > .05; right connective: F(2, 8) = .83, p > .05; siphon: F(2, 8) = .44, p > .05 (Figure 13). Thus, EGb 761 did not significantly change the nerveelicited activity in any of the three monitored nerves. 1.5
Percent of Baseline
SEM
R Connective Siphon
1
R P9
General discussion This study applied rigorous and controlled testing to determine the efficacy of the popular learning and memory supplement, Ginkgo biloba extract EGb 761. Previous research is mixed as to EGb 761’s ability to improve learning and memory (Birks & Grimley, 2009; DeKosky, et al., 2008; Gertz & Kiefer, 2004; Le Bars et al., 1997; Tchantchau et al., 2007) in human and animal models. The first experiment explored the possible learning and memory effects of long-term exposure to EGb 761 on a long-term habituation protocol in the Aplysia. Despite sensitive learning measures which accounted for several factors of learning and memory (rate, depth, short-term and long-term retention), EGb 761 failed to produce an effect in any dependent measure. EGb 761’s purported effect of arousal after acute exposure was explored in Experiment 2 (Brochet, 1999; Rigney, 1999). Two arousal measures, locomotion and S-SWR sensitivity, were tested after acute exposure to the drug. Again, EGb 761 failed to produce an effect on any of the dependent measures. To further explore the suggested mechanisms by which EGb 761 is purported to work, Experiment 3 tested its effects in Aplysia CNS. Lanahan (1997) and Williams et. al. (2004) have suggested that EGb 761 operates on the CNS by increasing neuronal activity and synaptic transmission; however, no main effect of EGb 761 was found on the spontaneous background and nerve-elicited nerve activity in the present study. The general conclusion from these experiments is that EGb 761 is not bioactive in the Aplysia nervous system. Possible interpretations One possible explanation for the lack of positive results in these experiments is a lack of experimental power to detect the effects of EGb 761. The measure of locomotion after acute exposure to EGb 761 showed high within-subject variability and low power. However, this interpretations appears unlikely. Numerous other tested measures of the efficacy of EGb 761 demonstrated low within-group variability and high levels of internal reliability. In addition, the three experiments provide converging and consistent results. The preponderance of evidence provided by these experiments thus indicates no
0.5
0
10
20
30
40
50
60
70
80
Time Figure 11: Background resting neural activity by nerve (P9, right connective, and siphon). EGb 176 does not affect resting neural activity in Aplysia (F<1 for each nerve).
22 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
200µV
90 1s
Figure 12: Representative siphon nerve recording of average nerveevoked activity as a percent of baseline (first 2 stimulations).
NEURO
Article
Genesis
Percent of Baseline
SEM
2
R Connective Siphon
1.5
1 R P9 0.5
0
0
10
20
30
40
Time
50
60
70
80
90
Figure 13: Nerve-elicited nerve activity as a percent of baseline (activity after first two pulses) for each nerve (P9, right connective, and siphon). No significant difference during EGb 761 exposure (F < 1 for each nerve).
effect of EGb 761. A second possible explanation for the null findings is that EGb 761 operates by a mechanism in humans which is not conserved in Aplysia. Aplysia are a useful animal model to study learning and memory, as their simple nervous system allows for identification of neural correlates of behavior. While the simplicity of Aplysia is an advantage in some respects, this simplicity also limits the applicability of invertebrate research to mammals. However, research with Caenorhabditis elegans, another invertebrate animal model, has indicated beneficial effects of EGb 761 (Luo, 2006; Smith & Luo, 2004; Wu et al., 2006). Moreover, recent research by Moroz et al. (2006) to sequence the Aplysia genome indicates that Aplysia may be more similar to humans than C. elegans genomically. Because Aplysia have NMDA receptors and are capable of synaptic plasticity, the animal model has proven useful for understanding mammalian learning and memory. Given the strong evolutionary conservation of the biochemistry of learning, it seems somewhat unlikely that Aplysia and mammals diverge in their sensitivity to EGb 761. Another possible explanation is that EGb 761 works outside of the CNS and its effects cannot be detected in Aplysia. Some research suggests that EGb 761 works through improving blood circulation and increasing the flow of oxygen to the brain (DeFeudis & Drieu, 2000). However, there is a lack of evidence to support EGb 761â&#x20AC;&#x2122;s effects on blood flow through direct monitoring of these biological effects. Possible future research can explore this mechanism through direct monitoring of blood flow and oxygenation as a result of exposure to EGb 761. A fourth possible conclusion is that EGb 761 is not bioactive. The study by Solomon et al. (2006) found that EGb 761 was not active in normal healthy adults. There have been mixed results as to its effectiveness in impaired populations; however, the recent GEM study by DeKosky et al. (2008) concluded that EGb 761 failed to improve memory or prevent dementia in older adults. A recent meta-analysis also found that EGb 761
was not effective in improving memory in older adults overall (Birks & Grimley, 2009). Despite the majority of null results in large and well-controlled studies, mixed evidence to support EGb 761â&#x20AC;&#x2122;s beneficial effects persists. This could possibly be due to publication bias, in which positive results are favored over negative results in publication. The proposed benefits of EGb 761 in positive studies are not consistently demonstrated, and the inconsistent purported effects do not appear in large, well-controlled studies. This study with Aplysia fits into the recent trend of evidence that EGb 761 does not appear to be bioactive in both Aplysia and humans. Abramson, C. (1994). A primer of invertebrate learning: The behavioral perspective. Washington, DC: American Psychological Association. Bear, M.F., Conners, B.W. & Paradiso, M.A. (2007). Neuroscience: Ex ploring the brain, 3rd edition. Philadelphia, PA: Lippincott, Williams, & Wilkins. Birks, J., & Grimley, E.J. (2009). Ginkgo biloba for cognitive impairment and dementia (review). The Cochrane Library, 1. Brembs, B. (2003). Operant reward learning in Aplysia. Current Directions in Psychological Science, 12(6), 218-221. Brochet, D., Chermat, R., DeFeudis, F.V., Drieu, K. (1999). Effects of single intraperitoneal injections of an extract of Ginkgo biloba (EGb 761) and its terpene trilactone constituents on barbital-induced narcosis in the mouse. General Pharmacology, 33, 249-256. Burns, N., Bryan, J., & Nettelbeck, T. (2006). Ginkgo biloba: No robust ef fect on cognitive abilities or mood in healthy young or older adults. Hu man Psychopharmacology, 21, 27-37. Carew, T.J., Pinsker, H.M., & Kandel, E.R. (1972). Long-term habituation of a defensive withdrawal reflex in Aplysia. Science, 175(4020), 451-454. DeFeudis, F.V., & Drieu, K. (2000). Ginkgo biloba extract (EGb 761) and CNS functions: Basic studies and clinical applications. Current Drug Targets, 1, 25-58. DeKosky, S.T., Williamson, J.D., Fitzpatrick, A.L., Kronmal, R.A., Ives, D.G., Saxton, J.A., Lopez, O.L., Burke, G., Carlson, M.C., Fried, L.P., Kuller, L.H., Robbins, J.A., Tracy, R.P., Woolard, N.F., Dunn, L., Snitz, B.E., Nahin, R.L., & Furberg, C.D. (2008). Ginkgo biloba for preven tion of Dementia: A randomized controlled trial. The Journal of the American Medical Association, 300(2), 2253-2262. Elsabagh, S., Hartley, D.E., Ali, O., Williamson, E.M., & File, S.E. (2005). Differential cognitive effects of Ginkgo biloba after acute and chronic treatment in healthy young volunteers. Psychopharmacology, 179, 437446. Gertz, H.J., & Kiefer, M. (2004). Review about Ginkgo biloba special extract EGb 761 (Ginkgo). Current Pharmaceutical Design, 10, 261264. Glanzman, D.L. (2006). The cellular mechanism of learning in Aplysia: Of blind men and elephants. Biological Bulletin, 210, 271-279. Gold, P.E., Cahill, L., & Wenk, G.L. (2003). The lowdown on Ginkgo biloba. Scientific American, 288(4), 86-92. Kandel, E.R. (2001). The molecular biology of memory storage: A dialogue between genes and synapses. Science, 294, 1030-1038. Kanowski, S., Hermann, W.M., Stephan, K., Wierich, W., & Horr, R. (1996). Proof of efficacy of the Ginkgo biloba special extract EGb 761 in outpatients suffering from mild to moderate primary degenera tive dementia of the Alzheimer type or multi-infarct dementia. Pharma copsychiatry, 29(2), 47-56. Kennedy, D.O., Scholey, A.B., & Wesnes, K.A. (2000). The dose-dependent cognitive effects of acute administration of Ginkgo biloba to healthy young volunteers. Psychopharmacology, 151, 416-423. Lanahan, A., Lyford, G., Stevenson, G.S., Worley, P.F., & Barnes, C.A. (1997). Selective alteration of long-term potentiation-induced tran scriptional response in hippocampus of aged, memory-impaired rats. The Journal of Neuroscience, 17, 2876-2885. Le Bars, P.L., Katz, M.M., Berman, N., Itil, T.M., Freedman, A.M., & Schatzberg, A.F. (1997). A placebo-controlled, double-blind, random
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 23
NEURO Genesis
ized trial of an extract of Ginkgo biloba for dementia: North Ameri can EGb study group. The Journal of the American Medical Associa tion, 278(16), 1327-1332. Lin, C.C.K., Cheng, W.L., Hsu, S.H., & Chang, C.M.J. (2003). The effects of Ginkgo biloba extracts on the memory and motor functions of rats with chronic cerebral insufficiency. Neuropsychobiology, 47, 47-51. Lochmatter, T., Roduit, P., Cianci, C., Correll, N., Jacot, J., & Martinoli, A. (2008). SwisTrack: A flexible open source tracking software for multiagent systems. Proceedings of the IEEE/RSJ 2008 International Con ference on Intelligent Robots and Systems, 4004-4010. Luo, Y. (2006) Alzheimer’s disease, the nematode Caenorhabditis elegans, and ginkgo biloba leaf extract. Life Science, 78(18), 2066-2072. Luo, Y., Smith, J., Paramasivam, V., Burdick, A., Curry, K., Buford, J., Khan, I., Netzer, W., Xu, H., & Butko, P. (2002). Inhibition of amyloid-beta aggregation and caspase-3 activation by the Ginkgo biloba extract EGb 761. Proceedings of the National Academy of Sciences of the United States of America, 99, 12197-12202. Marinesco, S., Wichremasinghe, N., Kolkman, K.E., Carew, T.J. (2004). Serotonergic modulation in Aplysia: Cellular and behavioral conse quences of increased serotonergic tone. Journal of Neurophysiology, 92, 2487-2496. Moroz, L.L., Edwards, J.R., Puthanveettil, S.V., Kohn, A.B., Ha, T., Heyland, A., Knudsen, B., Sahni, A., Yu, F., Liu, L., Jezzini, S., Lovell, P., Iannucculli, W., Chen, M., Nguyen, T., Sheng, H., Shaw, R., Kala chikov, S., Panchin, Y.V., Farmerie, W., Russo, J.J., Ju, J., & Kandel, E.R. (2006). Neuronal transcriptome of Aplysia: Neuronal compartments and circuitry. Cell, 127, 1453–1467. Muller, W.E., & Chatterjee, S.S. (2003). Cognitive and other behavioral effects of EGb 761 in animal models. Pharmacopsychiatry, 36, 24-31. Perlman, A.J. (1979). Central and peripheral control of siphon-withdrawal reflex in Aplysia californica. Journal of Neurophysiology, 42(2), 510-529. Petkov, V.D., Belcheva, S., & Petkov, V.V. (2003). Behavioral effects of Ginkgo biloba L., Panax ginseng C.A. Mey. and Gincosan®. The American Journal of Chinese Medicine, 31(6), 841-855. Rankin, C.H. (2002). A bite to remember. Science, 296(5573), 1624-1625. Rigney, U., Kimber, S., & Hindmarch, I. (1999). The effects of acute doses of standardized Ginkgo biloba extract on memory and psychomotor performances in volunteers. Phytotherapy Research, 13, 408-415. Smith, J., Burdick, A., Golik, P., Khan, I., Wallace, D., & Luo Y. (2002). Anti-apoptotic properties of Ginkgo biloba extract EGb 761 in differ entiated PC12 cells. Cellular and Molecular Biology, 48, 699-707. Solomon, P.R., Adams, F., Silver, A., Zimmer, J., & DeVeaux, R. (2002). Ginkgo for memory enhancement: A randomized controlled trial. The Journal of the American Medical Association, 288(7), 835-840. Stopfer, M., & Carew, T.J. (1996). Heterosynaptic facilitation of tail sensory neuron synaptic transmission during habituation in tail-induced tail and siphon withdrawal reflexes of Aplysia. The Journal of Neuroscience, 16, 4933-4948. Stopfer, M., Chen, X., Tai, Y.T., Huang, G.S., & Carew, T.J. (1996). Site specificity of short-term and long-term habituation in the tail-elicited siphon withdrawal reflex of Aplysia. The Journal of Neuroscience, 16(16), 4923-4932. Tchantchou, F., Xu, Y., Wu, Y., Christen, Y., & Luo, Y. (2007). EGb 761 enhances adult hippocampal neurogenesis and phosphorylation of CREB in transgenic mouse model of Alzheimer’s disease. The Journal of the Federation of American Societies for Experimental Biology, 21, 2400-2408. Wang, Y., Wang, L., Wu, J., & Cai, J. (2006). The in vivo synaptic plasticity mechanism of EGb 761-induced enhancement of spatial learning and memory in aged rats. British Journal of Pharmacology, 148, 147-153. Williams, B., Watanabe, C.M., Schultz, P.G., Rimbach, G., & Krucker, T. (2004). Age-related effects of Ginkgo biloba extract on synaptic plastic ity and excitability. Neurobiology of Aging, 25, 955-962. Wu, Z., Smith, J. V., Paramasivam, V., Butko, P., Khan, I., J.R. Cypser, J.R. & Luo, Y. (2002) Ginkgo biloba Extract EGb761 Increases Stress Resistance and Extends Life Span of Caenorhabditis elegans. Cellular and Molecular Biology, 48, 725-731. Zolman, J.F., & Peretz, B. (1987). Motor neuronal function in old Aplysia is improved by long-term stimulation of the siphon/gill reflex. Behav ioral Neuroscience, 101(4), 524-533.
24 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Article
NEURO
Letter
Genesis
Examining the “dys”- ordered schizophrenic brain Barrington Quarrie1 1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Barrington Quarrie (baq@duke.edu)
Papaleo and colleagues (2010) examined the molecular effects of a fairly new protein in the field, dysbindin, on phenotypical changes relating to schizophrenia-like behavior and looked at its possible correlation with the wellknown dopamine hypothesis for schizophrenia. “It’s almost as if a demon might have passed from one host to another.” - John Nash John Nash, the Nobel Prize winning mathematician, refers to the tragic effects of schizophrenia, a debilitating mental illness that affects about 1% of the population (Strobel, 2005). As Nash hints, some of the prime symptoms in the illness are psychosis (e.g., hallucinations and delusions) and thought disorganization (Laruelle et al., 2003). This brain dysfunction may be explained by the dopamine (DA) hypothesis, which states that abnormal behaviors associated with schizophrenia arise from an increase of DA signaling via D2 receptors in several brain regions, particularly the prefrontal cortex (PFC) ( Ji et al., 2009). This D2 postulation has driven researchers to elucidate the molecular pathways by which individuals with seemingly normal, even great, mental capability like John Nash can become delusional and lose their senses of reality. In the past five years, researchers have discovered the dysbindin protein, promising for its possible associations with schizophrenia. The dysbindin protein is localized near synaptic sites in the PFC and may be involved with presynaptic, DA vesicle trafficking ( Ji et al., 2009) as well as postsynaptic, D2 receptor internalization (Iizuka et al., 2007). Further research is necessary, however, to determine the molecular and cellular effects of dysbindin on the expression of schizophrenia-like behavior. Papaleo and colleagues have explored the effects of dysbindin by performing cognitive, emotional, and physiological experiments in the mouse. They utilized knockout mice with an inactivated dysbindin gene and chronic D2 agonist treatment to examine the role of dysbindin in modulating PFC DA signaling. This model may provide evidence for an individualized drug treatment target in schizophrenic patients (Figure 1). Because working memory deficits correlate with schizophrenia and DA functioning in the PFC (Vijayraghavan et al., 2007), Papaleo et al. first analyzed the
working memory of dysbindin knockout mice. With use of a T-maze paradigm under demanding conditions (i.e., increased number of trials with fewer time intervals), they discovered that knockout mice (dys -/-) performed worse than heterozygous dysbindin gene mice (dys +/-) and worse than the wild-type mice (dys +/+). These results indicate that reduced levels of dysbindin impairs working memory under stressful situations, which correlates with schizophrenia-like behavior. In addition, knockout mice exhibited increased stress reactivity but, surprisingly, increased pre-pulse inhibition–a weak initial stimulus stops a neurobiological reaction to a later stronger stimulus. The latter finding presents an anomaly to previously tested schizophrenia-like models and necessitates further inquiry. Since dysfunctional working memory is associated with the PFC and prior research has localized dysbindin to this anterior, cortical region, Papaleo and colleagues further
Figure 1: Model Dysbindin plays an important role in influencing prefrontal cortex (PFC) functioning through modulation of D2 receptor activation and of CaMK components in the DA signaling pathway, ultimately affecting schizophrenia-like behavior
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 25
NEURO Genesis
examined the pyramidal neurons within layer II/III of the medial PFC (mPFC). The mPFC is critical for the formation of cognitive networks as well as DA modulation and thus an appropriate region of interest to determine whether varying levels of dysbindin differentially affects neuronal excitability. With use of whole-cell current clamp recording, it was found that pyramidal neurons were more excited in null dysbindin mice relative to wild-type mice, suggesting that reduced dysbindin levels may affect the excitability of the pyramidal neurons in the PFC (Papaleo et al., 2010). Ca2+/calmodulin kinase (CaMK) components are involved in working memory and DA regulation of the neuronal excitability in this region (Gonzalez-Islas et al., 2003, Runyan et al., 2005). When researchers examined dysbindin’s connection with CaMK, it was found that null mice had lower levels of CaMK, suggesting that dysbindin may also influence the excitability of pyramidal neurons through CaMK modulation. Although these results show a phenotypical connection to dysbindin, the underlying mechanism is still in question. Based on prior research documenting dysbindin’s connection with dopamine, however, Papaleo and colleagues (2010) examined the possible influence of dysbindin on schizophrenic-like behavior through regulation of DA signaling. Chronic drug treatment studies with quinpirole, a D2 agonist, were conducted to determine whether mimicking increased DA signaling correlates with results observed in null dysbindin mice. Pyramidal neurons in the null dysbindin mice were shown to be more sensitive to acute D2 receptor activation compared to wild-type, which indicates a possible increased amount of D2 receptors on the surface of the neurons. Using the same experimental protocol, the researchers also used the chronic D2 receptor stimulation model to test another group of wildtype mice. Relative to wild type mice, quinpirole–treated mice performed poorly on the more demanding T-maze paradigm and exhibited increased pre-pulsed inhibition and decreased CaMK levels, which correlate with the results seen in the null dysbindin mice. These similar results suggest that the schizophrenia-like behavior seen with reduced dysbindin correlates with increased DA signaling. Moving forward, questions regarding dysbindin’s role in the biological mechanism underlying schizophrenia have yet to be answered. The most obvious question is why there is an increase of pre-pulse inhibition seen in both the null dysbindin and quinpirole-treated mice. Although the researchers argue that it may be caused by a change in sensitization, more research needs to examine this abnormality in the results. In addition, the pathway through which DA signaling modulates the excitability of pyramidal neurons in mPFC layer II/III needs to be explored. Previous research has shown inhibitory effects on the PFC pyramidal neurons via D2 activation (Vijayraghavan et al., 2007). While enhanced baseline excitability in PFC pyramidal neurons result from a decrease in dysbindin (Papaleo et al., 2010), the mechanism by which this occurs has not been 26 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Letter uncovered because of experimental constrains. Overall, the implication behind these results is that dysbindin plays an important role in influencing PFC functioning through the modulation of D2 receptor activation and CaMK components in the DA signaling pathway, which ultimately affect schizophrenia-like behavior. Current research supports that dysbindin critically influences PFC function through modulation of D2 receptor activation and CaMK components in the DA signaling pathway, which is implicated in schizophrenia-like behavior. This present study indicates that genetic variations within DTNBP-1(the dysbindin-encoding gene) may be genetic marker for schizophrenia risk. A deeper investigation of the mechanisms by which dysbindin interacts with various biological molecular structures may unveil how dopaminergic signaling can be regulated to normalize PFC function. Such an understanding may facilitate pharmaceutical or psychiatric treatment research for schizophrenia. Gonzalez-Islas, C., and Hablitz, J.J. (2003). Dopamine enhances EPSCs in layer II-III pyramidal neurons in rat prefrontal cortex. J Neurosci 23, 867-875. Iizuka, Y., Sei, Y., Weinberger, D.R., and Straub, R.E. (2007). Evidence that the BLOC-1 protein dysbindin modulates dopamine D2 receptor internalization and signaling but not D1 internalization. J Neurosci 27, 12390-12395. Ji, Y., Yang, F., Papaleo, F., Wang, H.X., Gao, W.J., Weinberger, D.R., and Lu, B. (2009). Role of dysbindin in dopamine receptor trafficking and cortical GABA function. Proc Natl Acad Sci U S A 106, 19593-19598. Laruelle, M., Kegeles, L.S., and Abi-Dargham, A. (2003). Glutamate, dopamine, and schizophrenia: from pathophysiology to treatment. Ann N Y Acad Sci 1003, 138-158. Papaleo, F., Yang, F., Garcia, S., Chen, J., Lu, B., Crawley, J.N., and Weinberger, D.R. (2010). Dysbindin-1 modulates prefrontal cortical activity and schizophrenia-like behaviors via dopamine/D2 pathways. Mol Psychiatry, 1-14. Runyan, J.D., Moore, A.N., and Dash, P.K. (2005). A role for prefrontal calcium-sensitive protein phosphatase and kinase activities in working memory. Learn Mem 12, 103-110. Strobel, G. (Interviewer) & Murray, Robin. (Interviewee). (2005). Schizophrenia Research Forum [Interview transcript]. Retrieved from Schizophrenia Research Forum Web site: http://www.schizophreniaforum.org/ for/int//Murray/murray.asp Vijayraghavan, S., Wang, M., Birnbaum, S.G., Williams, G.V., and Arnsten, A.F. (2007). Inverted-U dopamine D1 receptor actions on prefrontal neurons engaged in working memory. Nat Neurosci 10, 376-384.
NEURO
Article
Genesis
Improved decoding methods to reduce reaction time in brain - machine interface systems Olga Mutter1 1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Olga Mutter (olga.mutter@duke.edu)
SUMMARY: Brain machine interfaces (BMIs) use recorded neuronal activity to establish direct, real-time communication between the brain and external actuators such as prosthetic limbs (Lebedev & Nicolelis, 2006, 2009). Improved understanding of the relationship between modulations in sensorimotor cortical neurons and the complexities of realistic arm kinematics is a critical step in facilitating recovery in patients with motor deficits. Reaction time in the human motor system is a consequence of delays in motor and pre-motor processing steps as well as corticospinal projections. There exist many applications where this inherent response delay is undesirable such as vehicle control and military operations. Faster motor responses, even on the order of milliseconds, can drastically improve safety and performance in many common time-sensitive scenarios. Prior to initiation of movement, neurons in the motor (M1) and sensory (S1) cortices encode information about the timing and type of movement that will be initiated. Extraction of these parameters in real-time enables a reduction in latency between motor plan selection and initiation of movement. In this project, we characterized M1 and S1 neural activity during a reaction-time reaching task and identified highly modulated sensorimotor neuronal subpopulations to optimize our decoding methods and enact BMI operation at reduced reaction times. We were able to create a model that successfully predicts and executes movements to a specified target faster than a monkey itself can execute those same movements. This model proved effective, reaching the correct target for the majority of trials at a faster rate than the monkey model, thus reducing the response time in the execution of planned movement.
Introduction Sensorimotor defects resulting from neurologic injuries, diseases, or limb loss affect millions of people worldwide. In the United States alone, five million people are currently afflicted with some form of paralysis according to data from Medical News Today (Paddock, 2009). Such paralyzing disorders substantially limit independence, mobility, and communication. Despite severe motor deficits due to damage to the spinal cord, nerves, or muscles, many patients retain fully intact cortical and subcortical motor networks that are capable of motor processing (Mattia et al., 2009). These areas can adapt to new controls due to innate brain plasticity (Hosp & Luft, 2003; Winstein et al., 2003). To bypass the site of neural lesion, activity from healthy motor regions such as M1 or S1 can be connected to a neural prosthetic through an interface, called a brainmachine interface (BMI) (Lebedev & Nicolelis, 2006). Thus, artificial actuators such as an exoskeleton or artificial limbutilizing neurophysiological signals from undamaged components of the central nervous system allow for direct interaction between the brain and the outside world (Andersen, et al., 2004; Jackson et al., 2004; Lebedev et al., 2006). Many research groups are currently pursuing this goal with the hope that BMIs with increasingly sophisticated technologies and decoding strategies will serve to augment partial and full body mobility in paralyzed patients (Andersen, et al., 2004; Birbaumer et al., 2007; Fetz,
2007; Lebedev et al., 2006; Mussa-Ivaldi et al., 2003, Nicolelis et al., 2009, Schwartz et al., 2006). Development of BMI systems In recent years, there has been much development in the quality of recordings extracted from neuronal ensembles with the aim of creating improved BMIs to drive neuroprosthetics. Initially, single-electrode implants in the brain showed promise for providing the source of signals to drive artificial devices in restoration of mobility after paralysis (Schmidt et al., 1980). Advancing on this technique, the development of the novel electrophysiological model of multi-electrode recordings, such as the Utah Intracortical Electrode Array (UIEA) emerged (Maynard, et al., 1996). The UIEA demonstrated the ability of neuronal populations to perform control tasks and showed that the number of neurons present in a recording is significant, as recordings from small populations of neurons rather than single units are more reliable for brain-computer interface application. The introduction of this effective model was almost simultaneous with the development of BMIs (Schmidt, 1980). Chronic implants containing multielectrode arrays in multiple cortical areas of the rhesus monkeys brain are now able to record extracellular electrical activity of hundreds of neurons (Carmena, et al., 2003; Fitzsimmons et al., 2009; Lebedev et al., 2005; Nicolelis et al., 2003; Lebedev & Nicolelis, 2011). With these novel electrophysiological techniques executing simultaneous Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 27
NEURO Genesis
recordings from populations of neurons in distinct brain areas, significant information can be derived from ensemble encoding (Nicolelis et al., 1995, 1997, 2003; Kralik et al., 2001; Nicolelis & Ribeiro, 2002). After nearly two decades, the ability of highly distributed populations of broadly tuned neurons to sustain the continuous production of motor behaviors in real-time has been evaluated and better-understood (Serruya, et al., 2002; Taylor et al., 2002, Wessber et. al, 2000). Many studies have demonstrated that cortical neurons modulate their firing during voluntary movements and therefore encode a variety of movement parameters (Georgopoulos et al., 1986, 1992; Ashe & Georgopoulos, 1994; Fu et al., 1995; Sergio & Kalaska, 1998). However, it is well known that individual neurons largely do not demonstrate a oneto-one relationship with any motor parameter (Ashe & Georgopoulos, 1994; Sergio & Kalaska, 1998; Carmena et al., 2003) as they exhibit considerable variability in their neural activity (Lee et al., 1998; Shadlen & Newsome, 1998; Cohen & Nicolelis, 2004; Wessberg & Nicolelis, 2004). Rather, it is the cumulative activity of large and diverse neuronal ensembles that regulates execution of precise movement (Wessberg et al., 2000; Carmena et al., 2003).Incorporating motor, sensory, and cognitive signals, encodings of such activity show high complexity (Cheney & Fetz, 1980; Alexander & Crutcher, 1990; Lebedev et al., 1994; Donchin et al., 1998; Kakei et al., 1999). Therefore, it has been demonstrated that large neuronal populations maintain the ability to generate motor commands during planning and execution of movements despite being disconnected from the body’s effectors. Thus, it is possible to extract neuronal signals from real-time recordings of such populations, obtain relevant movement parameters, and utilize them in the control of BMI driven neuroprosthetics (Andersen, et al., 2004; Carmena et al., 2003; Chapin et al., 1999; Jackson et al., 2004; Patil et al., 2004; Nicolelis, 2001; Serruya et al., 2002; Taylor et al., 2002; Wessberg et al., 2000; Wessberg & Nicolelis, 2004). Previous studies have also shown that the arm region of the motor cortex specifically encodes kinematics of reach parameters including velocity (Moran and Schwartz, 1999; Lebedev et al., 2005) and information about movement onset and offset (Lebedev et al., 2008). For example, timevarying speed of movement was represented in cortical activity in addition to the well-studied average directional selectivity, or preferred direction of a single cell (Moran and Schwartz, 1999). In Lebedev et al., 2008, researchers identified neuronal ensembles capable of encoding information regarding temporal intervals during self-timed delay tasks. This neuronal ensemble activity was used to generate predictions that discriminated between delay periods and movement periods of the task. This study, and the study presented in this thesis, seeks to enhance the temporal resolution by which we can extract information from the brain. Neural coding operates at a millisecond time scale and improving BMI efficacy will depend on fast and accurate interpretation of neuronal population activ28 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Article ity. Another factor to consider in the creation of BMIs is the inherent plasticity of the brain and autonomic adjustments. As early as 1960, Clynes and Kline described the interaction of artificial and biological components. In their work, they discussed, “If man in space, in addition to flying his vehicle, must continuously be checking on things and making adjustments merely to keep himself alive, he becomes a slave to the machine”. They sought to create a mechanism by which “such robot-like problems are taken care of automatically and unconsciously, leaving man free to explore, to create, to think, and to feel” (Clynes & Kline,1960). In recent research, Mussa-Ivaldi and Miller explore the importance of plastic changes in the brain and rapid feedback in order to enact a more successful BMI. There have been several attempts to induce controlled plastic changes in the brain to simulate how a BMI would react to such a change. Such experiments provide the idea of activating central sensory areas directly using electrical stimulation as a means of reducing feedback delays (Mussa-Ivaldi & Miller, 2003). Serruya et al. and Taylor et al. utilized paradigms in which subjects receive visual feedback of brain-controlled movement and learning-induced changes in neural activity patterns are tracked (Serruya et al., 2002; Taylor et al., 2002). Serruya discussed use of a linear filter method that is constructed during neuronal control to test whether hand trajectory could be reconstructed from neural activity. Reconstruction accurately reflected hand trajectory, and movement to targets was nearly as good with brain control as it was with hand control, with time required to reach the target only slightly greater with neural signaling. In addition, the study demonstrated that visual and other forms of feedback, paired with a subject’s dynamic learning, can compensate for inaccuracies in a BMI model to provide a voluntarily adjusted control signal. Furthermore, their results demonstrated feasibility for human application with an electrode array suitable for human use (Serruya et al., 2002). Advancing on this, Taylor et al. created a model that incorporates learning-induced changes in neuronal activity during brain-controlled movements, bypassing the need for physical limb movement even while training the model. Researchers explored how visual feedback affects movements derived from cortical signals by comparing movements enacted during closed-loop paradigms (monkeys received visual feedback) and open-loop paradigms (monkeys did not receive visual feedback). Movement trajectories from closed-loops were more accurate on average than open-loops, and animals were able to improve their closed-loop target hit rate over the course of the experiment. These results imply that the subjects learned to modulate their brain signals more effectively when visual feedback was given, consistent with the discoveries of Serruya et al. Because hand-controlled cursor movements cannot be carried out in patients with movement deficits, a co-adap
Article tive movement prediction algorithm was developed. Such a model did not require physical limb movement, and learning induced changes in cell tuning properties were successfully tracked. Individual as well as groups of cells showed substantial differences in preferred directions between the two tasks. These results showed how paralyzed patients could make 3D cursor movements by co-adapting a prediction algorithm to their dynamic cell turning properties. This study also demonstrates the need to record neuronal activity from brain control elicited movements as well as movements physically enacted by the body in order to create BMIs that function to effectively represent natural movement (Taylor, et al., 2002). Realizing this, Musallam, et al., sought to extract movement goals from monkeys specifically without the monkeys enacting physical behavior related to the movement. They were able to decode intended goals of three monkeys using neuronal activity from brain-controlled trials as the monkeys positioned cursors on a computer screen independent of physical movement of the body. This demonstrated the ability to use high-level neural signals from the premotor and parietal cortices in driving a neural prosthetic (Musallam et al., 2004). Progress has been made in developing BMI models that can extract increasingly sufficient and valuable neuronal data, but there still remains much more to understand and improve upon in order to create a BMI model capable of producing natural voluntary and autonomous movement in humans. Monkey to human models In order to execute a BMI driven neuroprosthetic that enacts movements indicative of natural human voluntary movement, numerous facets of movement must be explored. As a close relative to humans, results using rhesus monkeys can provide a reasonable first estimate for what may or may not work in humans. Rhesus monkeys have become the animal model of choice in BMI research although previously, rodents, cats, owl monkeys, and humans have been used (Chapin, et al., 1999, Kennedy and Bakay, 1998; Stanley et al., 1999, Wessberg et al., 2000).
The rhesus monkey is the optimal model for several reasons, including their advanced reaching and grasping abilities and sophisticated hand manipulation skills. They are also able to quickly learn how to voluntarily control the firing rates of individual and multiple neurons in the primary motor cortex if rewarded for generating successful patterns (Schmidt et al., 1978). Additionally, rhesus monkeys have deeply cleft and furrowed brains, making them better models for human neurophysiology than owl monkeys (Carmena et al., 2003; Lebedev et al., 2005). Rhesus monkeys serve as more effective test subjects than human models for widespread research as they are similar in neural complexity and capability, but are less vulnerable to complications (Kennedy and Bakay, 1998). BMIs developed with these animals will provide feasibility for the main safety concerns of long-term implants and estimates of the neuroprosthetics performance. Using rhesus mon-
NEURO Genesis
keys, numerous studies have demonstrated that signaling from neuronal ensembles during movement related tasks provides substantial information to successfully control a computer cursor or robotic arm (Carmena et al., 2003; Lebedev et al., 2006; Musallam et al., 2004; Serruya et al., 2002; Taylor et al., 2002). Thus, with discoveries from BMI experimentation with monkey models proving promising, BMI technology may be applied to treatment of motor deficits in humans through the use of neural prosthetics in coming years. However, rigorous clinical trials validating the use of BMI control must be established. Several studies utilizing signaling from neuronal ensembles to control a computer cursor have also been enacted in human models, demonstrating applicability of BMIs for human patients with motor deficits (Kennedy et al., 2000; Leuthardt, et al., 2004; Hochberg, et al., 2005; Patil et al., 2004; Wolpaw and McFarland, 2004). In Kennedy et al., 2000, a neurotrophic electrode that uses trophic factors to encourage growth of neural tissue was implanted into the outer layers of the human cortex for two human subjects in order to synthetically produce speech and typing in patients who cannot effectively communicate. Results indicated that recorded neural signals can drive a cursor across a screen onto targets in order to accurately select icons or letter squares and such tasks demonstrated positive learning curves. The rate at which letters could be accurately selected in order to spell, 3 letters/min, is similar to alternative techniques (Kennedy et al., 2000). Such success yielded hope for BMI models in humans able to enact movements necessary for everyday tasks. A major advancement came from the research of Hochberg et al. in 2006 when a tetraplegic human participant in an ongoing pilot clinical trial was able to operate a neuromotor prostheses (NMP) controlled by spike activity from M1 neuronal ensembles. Even years after a spinal cord injury, it was found that neural spiking remained in the M1 arm area and intention of movement could be extracted. A cursor was controlled through extracted neural signals from M1 and the patient was able to open simulated e-mail, operate a television, open and close a prosthetic hand, and perform actions with a multi-jointed robotic arm, all while speaking. NMPs have the potential to be scaled so that parallel commands could be simultaneously extracted from multiple sensors, each in discrete cortical regions. Independent outputs could then emanate bilaterally to both arms and legs, for example (Hochberg et al., 2006). In order to most effectively and realistically execute motor actions using neuroprosthetics during everyday tasks effectively, feedback from the external environment in which these movements are made is vital. Examples include feedback regarding facets of mechanosensation such as touch and temperature. Such advancements are crucial in situations where visual feedback is insufficient such as picking up a textbook. In order to gauge how much force to exert in lifting the book, mechanosensory feedback is necessary throughout the duration of movement. Missing,
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 29
NEURO Genesis
injured, or paralyzed limbs lack sensory input to the CNS, in addition to motor functionality. Recent work has sought to further enhance BMIsby providing a sensory modality to the artificial limb. To implement sensory feedback into a BMI, a recent study developed a brain-machine-brain interface (BMBI), or simply a BMI system featuring intracortical microstimulation of S1 (O’Doherty et al, 2011). This sensorized BMI, unlike previous motor BMIs, acts to provide information regarding tactile sensory information in the environment explored by the actuated prosthetic limb (O’Doherty et al, 2011). In this novel experimental paradigm, rhesus monkeys enacted upper limb control using a brain-machine-brain interface that controlled the exploratory reaching movements of an external actuator while ICMS of primary somatosensory cortex provided a means for feedback. Distinct temporal patterns of ICMS encoded texture representationsof three visually identical textures. Monkeys operated this BMI using activity of motor neurons to move a realistic virtual avatar arm into position to identify and discriminate the unique artificial textures associated with each of three targets. Thus, monkeys were able to identify and experience tactile sensation solely using brain activity and feedback to the cortex. In his words, Dr. Miguel Nicolelis explains “Such an interaction between the brain and a virtual avatar was totally independent of the animal’s real body because the animals did not move their real arms and hands, nor did they use their real skin to touch the objects and identify their texture. It may be possible to create an exoskeleton that severely paralyzed patients could wear in order to explore and receive feedback from the outside word. Such an exoskeleton would be directly controlled by the patient’s voluntary brain activity in order to allow the patient to move autonomously. Simultaneously, sensors distributed across the exoskeleton would generate the type of tactile feedback needed for the patient’s brain to identify the texture, shape and temperature of objects, as well as many features of the surface upon which they walk.” Such work yields promise for the construction of BMIs that allow patients with motor deficits to regain function that closely parallels that of natural movement and interaction with the external environment. Extraction of movement parameters prior to movement onset Research into various facets of movement has provided hope for the restoration of autonomous and voluntary movement in patients with motor deficits; yet, there is an incredible amount that must still be learned. The ability to discover and improve the function of BMIs is essentially limitless, and must therefore be explored one step at a time. This research focuses on one specific and crucial facet of planned motor movements; the period following application of a stimulus but prior to the onset of movement, known as the reaction time. Further investigation into the reduction of reaction time would be widely applicable to 30 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Article the successful operation of BMI driven neuroprosthetics in the execution of everyday time-sensitive tasks. In monkeys, the activity of neuronal ensembles in the motor cortex (M1) during movement has been extensively analyzed, but there is much to be discovered from neuronal activity during the period prior to movement onset. Once cortical neurons demonstrated the ability to modulate their activity prior to movement, researchers have experimented with using these signals to control various prosthetic devices (Craggs et al., 1975; Wolpaw et al., 2004). This window where the movement intention is present but the action is yet to occur is the ideal temporal epoch to derive information for BMIs. Advances in chronic recording electrodes and signal-processing technology reveal the possibility to use these cortical signals in real-time. Real-time BMI systems operate most effectively when salient information is extracted prior to the movement event occurring. Much research has focused on improving this neural decoding algorithms to exploit the biological lead time of motor intentions, including work in rats (Chapin, et al., 1999) and nonhuman primates (Carmena et al., 2003; Moritz et al., 2008; Taylor et al., 2002; Velliste et al., 2008; Wessberg et al., 2000). Neural recordings from the M1 region of rhesus monkeys have demonstrated increased activity 200-500ms prior to the onset of voluntary movement (Kubota and Hamada, 1979). More specifically, a temporal segregation was discovered between parameters of direction, target position, and movement distance, with direction related discharge (115ms before movement onset) followed by target position (57ms after movement onset) and movement distance (248ms after movement onset), although some overlap was evident (Fu et al., 1995). Furthermore, it was also found that several of these motor parameters can be represented by a single cell(Fu et al., 1995). From the culmination of extensive research, the direction of movement is encoded by cortical neurons 150-100ms before the onset of the voluntary movement (Georgopoulos, 1995, 1995, 1988, 1989). Neurons in the primary motor cortex are often characterized by the preferred direction of the cell, i.e. the direction of movement, which corresponds to the largest increases in neural firing rate (Georgopoulos et al., 1989). The directional modulation of neurons is one of the primary motor parameters encoded during the reaction time period and is a parameter with important implications for motor BMIs. Combining the activity of large numbers of directionally tuned cells in M1, an estimate of the intended movement direction can be made, termed the population vector (Georgopoulos et al., 1989). The direction of the population vector (calculated at 20 ms intervals) during the reaction time was able to predict the direction of upcoming movement. Additionally, it has been found that during the reaction time period, target size influenced the slope of the rise in firing rate, and during movement, cortical representations were dependent on movement velocity (Ifft et al., 2011).
NEURO
Article
Genesis
Efforts to improve the speed of control for a computer cursor, simulating a shortened reaction time in execution of planned movements, have been the subject of a recent study with monkeys (Santhanam et al., 2006). In a key selection task, monkeys were able to perform center-out reaches of a prosthetic cursorusing a BMI driven by unprecedentedly brief neural recordings, on the order of hundreds of milliseconds. A higher performance BMI than previously reported, capable of operating up to 6.5 bits per second or ~15 words per minutes, was demonstrated. The results of this performance implicate a system design, which, in conjunction with other human studies, could substantially increase the clinical viability for BMIs in humans (Santhanam et al., 2006). Other such studies with humans include the development of directional specificity during the reaction time period, with specification evolving over a 200ms time period about 100ms after target presentation(Ghez et al., 1997). This study aims to advance on previous research conducted on neuronal activity extracted during the reaction time period in order to reduce the time course within which voluntary reaching movements can be executed. The goal of this study is to directly translate neuronal activity into a movement selection, such as the control over
1A
1B
1D
1C
movements of a cursor to a decoded target location. This approach is termed end-point control and speeds up actions by bypassing continuous BMI control, favoring discrete outcome selection. The ability of a BMI to correctly predict desired target direction from activity of neuronal ensembles during a specified epoch of the reaction time period was demonstrated. Longer windows yielded more accurate predictions but decreased the number of predictions that can be made per second. The approach used in the present study is to extract motor commands and identify the intended direction of movement early in the reaction time period.Linear discriminant analysis was employed to make categorical predictions of target location using binned neural activity to train the model. This strategy bypasses the need to predict continuous parameters such as position and velocity. As such, the time required to initiate a movement can be reduced. Using monkeys implanted with multielectrode arrays to record neuronal ensembles, we analyzed the activity of neuronal populations during the reaction time period in order to accomplish the following specific aims 1. Determine the relationship between pre-movement neural activity and the accuracy of classifier predictions of reach direction and 2. Demonstrate that pre-movement activity can be decoded to instruct the intended reach location faster than the monkeyâ&#x20AC;&#x2122;s own movements with high fidelity and accuracy.
Methods Electrode implants and recording mechanism Rhesus monkeys M and N (male, female respectively) were chronically implanted with multi-electrode arrays in right and left hemisphere M1 and S1 according to previously discussed surgical methods (Nicolelis et al., 2003). Two 96 channel microelectrode arrays were inserted into each hemisphere in cortical areas corresponding to representations of arm and leg (Figure 1A). Each array consisted of grids of independently moving electrode triplets, each of which was comprised of electrodes Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 31
NEURO Genesis
of different lengths. This allowed for sampling of neuronal activity from different depths of the cortex. For the purposes of this study, neural activity was only recorded in the arm representation area of right hemisphere M1 and S1 in both monkeys. A multichannel recording system was used to amplify, digitize, and filter recorded signals. (Plexon Inc, Dallas, TX, USA). Spike-sorting software utilizing on-line waveform template matching and threshold features was used to sort neuronal spikes. All studies were conducted with a protocol approved by the Duke University Institutional Animal Care and Use Committee and were in accordance with the NIH guidelines for the Care and Use of Laboratory Animals. Behavioral task Monkeys M and N were trained to control the position of a computer cursor in a two-dimensional reaching task using a hand-held joystick. X (left-right) and Y (forward-backwards) positions of the joystick were translated to the X (left-right) and Y (up-down) positions of the cursor on the display screen, placed 45 cm from the monkeys’ eyes (Figure 1B). They performed center-out movements to peripheral targets, similar to a previous human study studying pointing movements (Smyrnis et al., 2000). The joystick was placed on the side of the working hand (left), at waist level. The left hand was chosen to hold the joystick because the quality of neuronal recordings was better in the right hemisphere of each monkey. The monkey positioned its hand on the joystick to initiate a trial. If hand contact was broken with the joystick, the trial was discarded. Once the monkeys touched the joystick, a computer cursor (diameter of 0.5 cm) and a circle (diameter of 3cm) appeared at the center of the screen. The monkey then moved the cursor inside the circle, holding the cursor steady for a random interval between 800 and 1500ms (Figure 1D). After this interval concluded, the central circle disappeared and a peripheral target appeared at angles 45, 135, 225, or 315 degrees on a thin circle aligned on the center of the screen. The monkey was required to move the cursor to this peripheral target, a thickened arc of 8, 15, or 22 degrees (Figure 1C). When the cursor crossed the target from the inside of the circle out, a juice reward was given. If the cursor was moved outside of the boundary of the circle without crossing the target, the trial was terminated and a 500ms timeout was given. There was a 5s limit to each trial. Three sessions in monkey M and four sessions in monkey N were used for analysis. Data analysis PETH Analysis Neural activity was analyzed using peri-event time histograms (PETHs) aligned on target onset (Awiszus, 1997). Action potential (or spike) timestamp data from the experiments was recorded for offline for each cell. Bins of PETH represent number of spikes occurring in discrete increments of time over the task interval and 32 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Article were 25 ms in width. For each neuron, PETHs were calculated for each rewarded trial and averaged across trials for each possible movement direction (4 possibilities). For each neuron, this average modulation profile was normalized by subtracting the mean bin count and dividing by the standard deviation of the cells’ bin count. After normalizing, PETH depicts the statistical z-score, or the modulations as a fraction of the overall modulations. Neural decoding Linear discrimination analysis (LDA) is a linear classifier that uses a history of training data to make categorical predictions (Fisher, 1936). In the present study, LDA was used to decode the movement direction of a single trial from neuronal activity sampled across the task interval. A sliding window approach was used, where input for the LDA classifier was a vector of neural data, indicating the number of spikes occurring during a 100ms window for each cell. The window was slid at 25ms time steps from 500ms before and 1000ms after target appearance. The data was divided up such that 80% of trials for a given session serve as sample data, and 20% as training data for the decoder. LDA predicted variables of movement direction 1-4 (corresponding to 45°, 135°, 225°, and 315°, respectively) from ensembles of neurons. The accuracy of the LDA prediction analysis (fraction correct) was evaluated by dividing correct predictions by the number of possible correct predictions. The fraction correct was found at each time step of the sliding window throughout the length of the trial using M1 cells for monkey N and M1 and S1 cells for monkey M (Figure 3). Response profile An analysis was conducted to determine neuronal firing rates that exceeded an empirically determined threshold. First, the recorded spikes across a session for each cell were binned into 50ms bins. The mean spikes per bin were averaged across all neurons to generate population activity. To obtain the neuronal firing threshold value, we computed the mean population activity over a full session for each monkey. We defined threshold to be the initial time when population activity exceeded mean firing rate+2 standard deviations (Figure 4). This data was then analyzed with movement onset to determine the alignment between nearest threshold crossings (T0) and movement onset (M0). Movement onset was subtracted from nearest threshold crossing and the difference was obtained for each trial for monkeys M and N (Figure 5). The metric T0 – M0 was computed to better understand the temporal relationship between these two events. A t-test was performed on the distribution of T0 – M0 to determine if it was significantly shifted from 0 mean, as we would expect. Temporal effects on LDA prediction quality We sought to determine the effect of bin size on the quality of LDA predictions. Neuronal data was organized into 100ms bins, and four separate analyses were conducted using 1-4 bins, corresponding to a window length 100400ms. The first bin always began 400ms prior to movement onset. The time was chosen because the mean
Article reaction time of these trials was approximately 500ms. The average across bins was taken and input into the decoder representing the mean firing rate for each neuron from the population of cells for the 100-400ms window of time. LDA once again predicted the movement direction from ensembles of neurons with 20% of trials serving as training data and 80% of trials serving as sample data for the decoder. Fraction of correct predictions was individually calculated for each separate analysis of bin size. Analysis was conducted on four sessions for monkey N and two sessions for monkey M using both M1 and S1 cells (Figure 6).
Executing accurate movements faster than the monkey Our previous analyses yielded promise for a model that can decode a monkey’s pre-movement brain activity in order to instruct the intended reach location of each trial. We next sought to evaluate whether this information could be decoded and used to execute accurate movements towards a target faster than the monkey was able to execute such movements. LDA was used to generate predictions, with 30% of trials serving as training data and 70% as sample data. For each rewarded trial, LDA was used to predict movement direction from consecutive bins of neuronal data. Thus, LDA was applied at 100ms intervals, or bins, and the corresponding neuronal activity was decoded to determine the intended reach location at the earliest possible time interval while still maintaining accuracy for each trial. An algorithm for the computer model was constructed. Every 100ms (10 Hz) from the time of target appearance (TA) until the time of entry into the target,a prediction of movement direction was made. If three consecutive predictions of a discrete movement direction were found
NEURO Genesis
between the nearest target appearance (TA) and TA + 800ms, it was determined whether 1. The prediction was correct and 2. The prediction was made prior to the time that the monkey reached the target. If both criteria were satisfied, the computer model was given a ‘point’. If the prediction of movement direction was incorrect or the monkey reached the target before the prediction was made, the monkey model was given a ‘point’. If three consecutive bins were not found between TA and TA + 800ms, we sought to find 2/3 bins with the same movement direction from the time period of TA + 800ms to entry into the target. If 2/3 bins with the correct movement prediction were found and the above two criteria were satisfied, the computer model was given a ‘point’. If not, a ‘point’ was given to the monkey model for that trial. Thus, for every trial, if the LDA model was able to make an accurate prediction of intended reach location prior to the time that the monkey reached the target, the computer model received a point and ‘won’ that trial. However, if the computer model made either an incorrect prediction of movement or the monkey reached the target faster than an accurate prediction could be made, the monkey received a point and ‘won’ for that trial. For every trial,a single point or ‘win’ was given to either the computer or monkey model- whichever was able to execute the correct movement faster. Analysis was conducted on four sessions in monkey N and two sessions in monkey Musing both M1 and S1 cells (Figure 7). Results Data were collected from four daily recording sessions in monkey N (2126 trials) and from three sessions in monkey M (1305 trials). Neural activity was recorded from 64-69 M1 neurons (range depends on different
Figure 2: Neuronal activity by cell during course of trials for monkey N. Greatest neural activity occurred after target onset for most cells.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 33
NEURO
Article
Genesis
3A Monkey N, M1 cells
3C Monkey M, S1 cells
3B Monkey M, M1 cells
3D
Movement onset. Monkey N, M1 cells
3E
Movement onset. Monkey M, M1 cells
3F
Movement onset. Monkey M, S1 cells
3G
Entry into target. Monkey N, M1 cells
3H
Entry into target. Monkey M, M1 cells
3I
Entry into target. Monkey M, S1 cells
Figure 3: Fraction correct predictions of movement direction during the course of a trial. 3A., 3B., 3C. Values became significant upon onset of target appearance (confidence interval with standard deviation (SD) of 2, dashed line represents chance prediction value). 3D., 3E., 3F. Mean onset of movement for each session for either M1 or S1 cells (confidence interval with SD of 2, dashed line represents mean value). 3G., 3H., 3I.Mean entry into target for each session for either M1 or S1 cells (confidence interval with SD of 2, dashed line represents mean value).
34 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
NEURO
Article
Genesis
recording sessions) in monkey N, and from 92-111 M1 neurons and 83-91 S1 neurons in monkey M. PETH From the PETH analysis, it was evident that neuronal activity was greatest after target appearance for all four movement directions for the majority of cells. The response profiles differed slightly depending on movement direction reflecting the directional tuning of most M1 and S1 neurons (Figure 2). LDA
LDA analysis was used as a classifier to predict movement direction from neuronal ensembles. Before target appearance, LDA was able to make correct predictions with ~25% accuracy, equivalent to chance predictions. Within several milliseconds of target appearance, fraction of correct direction predictions increased. Mean +/- 2 standard deviations ofmovement onset(Figure 3D-F) and target entry (Figure 3G-I) times for each session for monkey N(M1 cells) and monkey M (M1 cells and S1 cells) are depicted. LDA predictions were deemed significant if they exceeded the 95% confidence interval (grey band in Figure 3A-C). The confidence interval was generated using the 1-proportion z-test at alpha = 0.05, following previous methods (Ifft et al., 2011).
4A
5A
Response profile In conducting a typical response profile, neural activity of the population was examined in the absence of temporal information about the behavior (see Methods). Neural activity throughout the course of the session from both monkeys was analyzed to depict a population response profile (Figure 4). Activity was averaged over the full population to yield mean response and response greater than 2 standard deviations above the mean (signifying threshold crossing). Population activity demonstrated sufficient amounts of threshold crossings throughout the course of the session for both monkeys, with data from monkey N shown (Figure 4). Further analysis was subsequently conducted on the temporal correlation between movement onset and nearest threshold crossing. We observed a rapid increase in firing rate that often occurred shortly prior to movement onset. In other words, a significant, detectable burst across the recorded subpopulation of neurons was seen at 318.3+711.6 (monkey N, p<0.05) and -179.9+1854.4 (monkey M, p <0.05) prior to movement onset (Figure 5). It was determined that threshold crossings occurred before movement onset for both monkeys N and M. A ttest was conducted for both monkeys M and N yielding highly significant results, affirming that neural ensembles
4B
5B
Figure 4: Population response profile. 4A. The recorded spikes across a session for each cell were binned into 50ms bins; population activity reflects the mean spikes per bin when averaged across all neurons for a single session. Sufficient threshold crossings were observed to conduct further analysis on the temporal relationship between movement onset and nearest threshold crossing. Both M1 and S1 cells from monkey N are depicted for one session. 4B. Select portion of the session (Movement onsets represented on the x-axis). Figure 5: Alignment of firing onset and movement onset. Movement onset was subtracted from nearest threshold crossing. Both monkeys M and N showed rapid firing onset (first instance of exceeding the + two standard deviation threshold) on average prior to movement onset (p<0.05). 5A. Monkey N, all cells. 5B. Monkey M, all cells.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 35
NEURO
Article
Genesis
6A
6B
6C
6D
peaked in activity prior to movement onset. Temporal effects on LDA prediction quality
LDA analysis was used as a classifier to predict behavioral direction from neuronal ensembles using window lengths of 100-400ms. For all four sessions for monkey N, fraction of correct predictions increased with increasing window length. Similar results were observed using two trials from monkey M, with an increase in the quality of LDA predictions as the length of the window was increased (Figure 6). Executing accurate movements faster than the monkey Four sessions from monkey N were used for analysis. Two sessions from monkey M were used for analysis. For the majority of trials, the computer model was able to decode intended reach location from neuronal activity to execute movements to a target; successfully completing a trial faster than the monkey could complete the same trial. Our results demonstrate that pre-movement activity can be effectively decoded to determine intended reach location, reducing the amount of time needed for a movement to be executed. Monkey M exhibited a higher percentage of tested trials. This is likely due to the superior quality of neuronal data extracted from analysis from monkey M because of the greater number and quality of cells used for recording (Figure 7). Discussion We were able to create a model that can successfully predict and execute movements to a specified target faster than a monkey itself can execute those same movements. 36 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Figure 6: Effect of bin size on LDA prediction quality. As window length increases, an increase in fraction of correct predictions is apparent. 6A. Results of four individual sessions in monkey N. 6B. Summated results over four sessions in monkey N. 6C. Results of two individual sessions in monkey M. 6D. Summated results over two sessions in monkey M.
The computer model proved extremely effective, successfully reaching the target and completing the majority of trials at a faster rate than the monkey model. This demonstrates the ability to accurately execute movements in a reduced amount of time, effectively eliminating a portion of the response time in planned movement. In addition, this method eliminates the time necessary to complete the movement itself, as the target is reached almost immediately after an accurate prediction is made. Thus, we accomplished our second aim by demonstrating that pre-movement activity can be decoded to instruct the intended reach location faster than the monkeyâ&#x20AC;&#x2122;s own movements with both high fidelity and accuracy. In order to determine feasibility for such a model, we conducted several analyses on the neuronal activity prior to and during movement. Each analysis yielded useful information that was utilized in order to develop our computer model. From the PETHs analysis it is evident that there are several components of reaction time. The first is a brief period of neuronal response after target appearance, during which no elevated activity is observed. A second interval after appearance of the target commonly consists of increased cell activity. Maximum activity is reached approximately 300-400ms after target appearance, although is varied between neurons. Results from the LDA analysis show a performance equal to chance before target appearance and predictions reach statistically significant values less than 100ms after target appearance while the brain is processing the stimulus and generating a motor command, consistent with previous
NEURO
Article 7A
Genesis
7B
studies (Ifft et al., 2011). The results from the threshold crossing data relative to movement direction demonstrate that if there is a synchronous burst of activity there will be likely be a movement to follow within the next second. It is from this pre-movement burst of population activity that rich information about upcoming movements can be extracted. Real-time BMIs will need to operate without any knowledge of task-dictated timing (such as target appearance). Laboratory experiments often enforce a rigid stimulus-cued movement paradigm which is marginally realistic and difficult to generalize to other behaviors. This result, therefore, suggests that BMIs could use neural activity itself as the movement trigger, obviating external timing signals from the decoding model. From our analysis, it is evident that neuronal ensembles are most active within 1 second prior to movement onset. We can then make accurate predictions of movement direction at this point in neuronal activity during the action of responding to a target via movement. Thus, we effectively accomplished our first aim of determining the relationship between pre-movement neural activity and the accuracy of classifier predictions of reach direction. With this mechanism of executing predictions, the time interval between this point and the onset of movement can be reduced, improving reaction time between target onset and initiation of movement. We then sought to investigate the effect of bin size on the quality of LDA predictions. From this analysis, it was evident that increased bin sizes of neuronal data yield more accurate predictions. However, while using larger bin sizes provides increased quantities of neuronal information and therefore more accurate predictions, it sacrifices the frequency at which predictions can be made. Using smaller bin sizes yields less neuronal information and less accurate predictions, but allows for greater frequency of predictions. Thus, a balance must be found between executing movements faster using less neuronal data, and executing movements more accurately with larger quantities of data. Our model uses bin sizes of 100ms, demonstrating a high temporal resolution of predictions while still maintaining sufficient accuracy.
Figure 7: Percentage of correct movements executed with reduced response times for monkey M and N. For 1/4 sessions in Monkey N and 2/2 sessions in Monkey M, the computer model was able to accurately execute movements to reach the target and successfully complete trials faster than the monkey model 7A. Monkey N, 4 sessions, 1526 trials. 7B. Monkey M, 2 sessions, 659 trials.
Executing movements in a reduced amount of time has widespread applicability to a range of fields such as machine control, video gaming, and competitive athletics. Most importantly, it is applicable to real-life scenarios where a fast response is necessary, such as operating vehicles and responding to dangerous situations. We are currently working on an advanced model using solely neural activity that is able to make predictions without any time related information. Such a model would have prediction windows aligned on threshold crossings so that it can be implemented in real-time. This model is likely to be effective as threshold crossings are strong predictors of movement onset.Thus, the next step in this research is the development of a BMI that generates continuous predictions in real-time as a monkey model executes movements. With success from such experiments, this model could then be applied towards human clinical trials. With the discoveries presented in this research, we can ultimately improve reaction time to allow for more natural function with neuroprosthetic devices, and enact faster movements in time-critical situations. Alexander, G.E., Crutcher, M.D. (1990). Neural representations of the target (goal) of visually guided arm movements in three motor areas of the monkey. J Neurophysiol64: 164-178. Andersen R.A., Musallam S., Pesaran B. (2004). Selecting the signals for a brain-machine interface. Curr Opin Neurobiol14:720–26. Ashe, J., Georgopoulos, A.P. (1994). Movement parameters and neural activity in motor cortex and area 5. Cereb Cortex4: 590-600. Awiszus, F. (1997). Spike train analysis. J Neurosci Methods74: 155-166. Birbaumer N., and Cohen L.G. (2007). Brain-computer interfaces: communication and restoration of movement in paralysis. J Physiol579:621–36. Carmena, J.M., Lebedev, M.A., Crist, R.E., O’Doherty, J.E., Santucci, D,M., Dimitrov, D.F., Patil, P.G., Henriquez, C.S. et al (2003). Learning to control a brain-machine interface for reaching and grasping by primates.PLoS Biology1 (2): E42. Chapin, J. K., Moxon, K. A., Markowitz, R. S. & Nicolelis, M. A. L. (1999). Real-time control of a robot arm using simultaneously recorded neurons in the motor cortex. Nature Neurosci 2: 664–670. Cheney, P.D., Fetz, E.E. (1980). Functional classes of primate corticomotoneuronal cells and their relation to active force. J Neurophysiol 44: 773-791. Clynes, M.E. and Kline, N.S. (1960) Cyborgs and space, in Astronautics pp. 26–27, 74–75, American Rocket Society. Cohen, D., and Nicolelis, M.A.L. (2004). Reduction of single-neuron firing uncertainty by cortical ensembles during motor skill learning. J
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 37
NEURO Genesis
Neurosci 24: 3574-3582. Craggs, M.D. (1975). Cortical control of motor prostheses: using the cordtransected baboon as the primate model for human paraplegia. Adv Neurol10:91–101. Donchin, O., Gribova, A., Steinberg, O., Bergman, H., Vaadia, E. (1998). Primary motor cortex is involved in bimanual coordination. Nature 395: 274-278. Fetz, E.E. (2007). Volitional control of neural activity: implications for brain-computer interfaces. J Physiol579:571–9. Fitzsimmons, N.A., Lebedev, M.A., Peikon, I.D., Nicolelis, M.A. (2009). Extracting kinematic parameters for monkey bipedal walking from cortical neuronal ensemble activity. Front Integr Neurosci 3:3. Fu, Q.G., Flament, D., Coltz, J.D., Ebner, T.J. (1995). Temporal encoding of movement kinematics in the discharge of primate primary motor and premotor neurons. J Neurophysiol 73: 836-854. Georgopoulos, A.P.(1995). Current issues in directional motor control. Trends Neurosci 18:506–510. Georgopoulos, A.P., Ashe, J., Smyrnis, N., Taira, M. (1992). The motor cortex and the coding of force. Science 256: 1692-1695. Georgopoulos, A.P., Kettner, R.E., Schwartz, A.B.(1988). Primate motor cortex and free-arm movements to visual targets in three-dimensional space. II. Coding of the direction of movements by a neuronal population. J Neurosci 8:2928–2937. Georgopoulos, A.P., Lurito, J.T., Petrides, M., Schwartz, A.B., Massey, J.T.(1989). Mental rotation of the neuronal population vector. Science243:234–236. Georgopoulos, A.P., Schwartz, A.B., Kettner, R.E. (1986). Neuronal population coding of movement direction. Science 233: 1416-1419. Georgopoulos, A.P., Pellizzer, G.(1995). The mental and the neural: psychological and neural studies of mental rotation and memory scanning. Neuropsychologia 33:1531–1547. Ghez, C., Favilla, M., Ghilardi, M.F., Gordon, J., Bermejo, R., Pullmann, S.(1997). Discrete and continuous planning of hand movements and isometric force trajectories. Exp Brain Res 115:217–233. Hochberg, L. R., Mukand, J., Polykoff, G., Friehs, G. & Donoghue, J. Braingate.Neuromotor prosthesis: Nature and use of neural control signals. Program No.520.17, 2005 Abstract Viewer/Itinerary Planner (Society for Neuroscience, Washington DC, 2005). Ifft, P., Lebedev, M. and Nicolelis, M.A. (2011). Cortical correlates of Fitts’ Law. Front Integr Neurosci 5:85. doi:10.3389/fnint.2011.00085. Jackson, A. B., Dijkers, M., Devivo, M. J. & Poczatek, R. B. (2004). A demographic profile of new traumatic spinal cord injuries: change and stability over 30 years. Arch. Phys Med Rehabil 85: 1740–1748. Hosp, J.A. and Luft, A.R. (2011). Cortical plasticity during motor learning and recovery after ischemic stroke. Neural Plasticity, vol. 2011: 871296. Kakei, S., Hoffman, D.S., Strick, P.L. (1999). Muscle and movement representations in the primary motor cortex. Science 285: 2136-2139. Kennedy, P.R. and Bakay, R.A. (1998). Restoration of neural output from a paralyzed patient by a direct brain connection. Neuroreport9 (8): 1707–11. Kennedy, P. R., Bakay, R. A. E., Moore, M. M., Adams, K. & Goldwaithe, J. (2000).Direct control of a computer from the human central nervous system. IEEE Trans Rehabil8: 198-202. Kralik, J.D., Dimitrov, D.F., Krupa, D.J., Katz, D.B., Cohen, D., Nicolelis, M.A.L. (2001). Techniques for long-term multisite neuronal ensemble recordings in behaving animals. Methods25:121-150. Kubota K. and Hamada, I.(1979) Preparatory activity of monkey pyramidal tract neurons related to quick movement onset during visual task performance. Brain Res 168:435–439. Lebedev, M.A., Carmena, J.M., O’doherty, J.E., Zacksenhouse, M., Henriquez, C.S., Principe, J.C., and Nicolelis, M.A. (2005). Cortical ensemble adaptation to represent velocity of an artificial actuator controlled by a brain-machine interface. JNeurosci 25:4681-4693. Lebedev, M.A., Denton, J.M., Nelson, R.J. (1994). Vibration-entrained and pre-movement activity in monkey primary somatosensory cortex. J Neurophysiol 72: 1654-1673. Lebedev, M. A. & Nicolelis, M. A. (2006). Brain-machine interfaces: past, present and future. Trends Neurosci 29: 536–546. Lebedev, M.A., O’Doherty, J.E., Nicolelis, M.A.L.(2008). Decoding of
38 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Article temporal intervals from cortical ensemble activity. J Neurophysiol 99:166–186. Lee, D., Port, N.L., Kruse, W., Georgopoulos, A.P. (1998). Variability and correlated noise in the discharge of neurons in motor and parietal areas of the primate cortex. J Neurosci 18: 1161-1170. Leuthardt, E. C., Schalk, G., Wolpaw, J. R., Ojemann, J. G. & Moran, D. W. (2004). Abrain-computer interface using electrocorticographic signals in humans. J Neural Eng1: 63–-71. Mattia, D., Cincotti, F., Astolfi, L., de Vico Fallani, F., Scivoletto, G., Marciani, M.G., et al. (2009). Motor cortical responsiveness to attempted movements in tetraplegia: evidence from neuroelectrical imaging. Clin Neurophysiol 120:181–9. Maynard, E.M., Nordhausen C.T., Normann, R.A. (1996). The Utah Intracortical Electrode Array: A recording structure for potential braincomputer interfaces. Electroencephalography and Clinical Neurophysiology.102: 228-239. Moran, D.W., and Schwartz, A.B. (1999). Motor cortical representation of speed and direction during reaching. J Neurophysiol 82:2676-2692. Moritz C.T., Perlmutter S.I., Fetz E.E. (2008). Direct control of paralysed muscles by cortical neurons. Nature456:639–42. Musallam, S., Corneil, B. D., Greger, B., Scherberger, H. & Andersen, R. A. (2004). Cognitive control signals for neural prosthetics. Science305: 258–-262. Mussa-Ivaldi F.A., Miller L.E. (2003). Brain-machine interfaces: computational demands and clinical needs meet basic neuroscience. Trend Neurosci 26: 329–34. Nicolelis, M.A.L. (2001). Actions from thoughts. Nature409: 403-407. Nicolelis, M.A., Baccala, L.A., Lin, R.C., Chapin, J.K. (1995). Sensorimotor encoding by synchronous neural ensemble activity at multiple levels of the somatosensory system. Science268: 1353-1358. Nicolelis, M.A., Dimitrov, D., Carmena, J.M., Crist, R., Lehew, G., Kralik, J.D., and Wise, S.P. 845 (2003). Chronic, multisite, multielectrode recordings in macaque monkeys. Proc Natl Acad Sci USA100: 1104111046. Nicolelis, M.A.L., Ghazanfar, A.A., Faggin, B.M., Votaw, S., Oliveira, L.M. (1997). Reconstructing the engram: simultaneous, multisite, many single neuron recordings. Neuron18: 529-537. Nicolelis, M. A. & Lebedev, M. A. (2009). Principles of neural ensemble physiology underlying the operation of brain-machine interfaces. Nature Rev Neurosci 10: 530–540. Nicolelis, M.A.L., Ribeiro, S. (2002). Multielectrode recordings: the next steps. Curr Opin Neurobiol12: 602-606. Paddock C. Paralysis affects more Americans than previously thought. Available at:http://www.medicalnewstoday.com/articles/146819.php. Patil, P. G., Carmena, J. M., Nicolelis, M. A. L. & Turner, D. A. (2004). Ensemblerecordings of human subcortical neurons as a source of motor control signals for a brain-machine interface. Neurosurgery55: 27–-38. Santhanam, G., Ryu, S.I., Yu B.M., Afshar A., Shenoy K.V. (2006). A high-performance brain-computer interface. Nature442: 195-197. Schmidt, E. M. (1980). Single neuron recording from motor cortex as a possible source of signals for control of external devices. Ann. Biomed. Eng. 8: 339–349. Schwartz A.B., Cui X.T., Weber D.J., Moran D.W. (2006). Braincontrolled interfaces: movement restoration with neural prosthetics. Neuron52:205–20. Sergio, L.E., Kalaska, J.F. (1998). Changes in the temporal pattern of primary motor cortex activity in a directional isometric force versus limb movement task. J Neurophysiol 80: 1577-1583. Serruya, M. D., Hatsopoulos, N. G., Paninski, L., Fellows, M. R. & Donoghue, J. (2002). Instant neural control of a movement signal. Nature416: 141–-142. Shadlen, M.N., Newsome, W.T. (1998). The variable discharge of cortical neurons: implications for connectivity, computation, and information coding. J Neurosci 18:3870-3896. Schmidt, EM; McIntosh, JS; Durelli, L; Bak, MJ (1978). “Fine control of operantly conditioned firing patterns of cortical neurons.”. Experimental neurology61 (2): 349–69. Smyrnis, N., Evdokimidis, I., Constantinidis, T.S., and Kastrinakis, G. (2000). Speed-accuracy trade-off in the performance of pointing
Article
NEURO Genesis
movements in different directions in two dimensional space. Exp Brain Res134: 21-31. Stanley, GB; Li, FF; Dan, Y (1999). Reconstruction of natural scenes from ensemble responses in the lateral geniculate nucleus. J Neurosci19 (18): 8036–42. Taylor, D. M., Tillery, S. I. H. & Schwartz, A. B. (2002). Direct cortical control of 3Dneuroprosthetic devices. Science296: 1829–-1832. Velliste M., Perel S., Spalding M.C., Whitford A.S., Schwartz A.B. (2008). Cortical control of a prosthetic arm for self-feeding. Nature 453:1098–101. Wessberg, J., Stambaugh, C.R., Kralik, J.D., Beck, P.D., Laubach, M., Chapin, J.K., Kim, J., Biggs, S.J., Srinivasan, M.A., Nicolelis, M.A. (2000). Real-time prediction of hand trajectory by ensembles of cortical neurons in primates. Nature 408: 361-365. Wessberg, J.and Nicolelis, M.A.L. (2004). Optimizing a linear algorithm for real-time robotic control using chronic cortical ensemble recordings in monkeys. J Cogn Neurosci 16: 1022-1035. Winstein, C.J, Wing, AM, Whitall, J. (2003). Motor control and learning principles for rehabilitation of upper limb movements after brain injury. In: Grafman J, Robertson I, eds. Plasticity and Rehabilitation, Amsterdam: Elsevier Science BV:77-137. (Handbook of Neuropsychology; 2nd Edition, vol. 9.). Wolpaw, J. & McFarland, D. (2004). Control of a two-dimensional movement signal bya noninvasive brain-computer interface in humans. Proc Natl Acad Sci USA 101: 17849-17854.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 39
NEURO
Letter
Genesis
Autism as a case study for the examination of Rushworth’s reinforcement-guided decision-making model and the functional neuroanatomy of the anterior cingulate cortex Colin Martz1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Colin Martz (colin.martz@duke.edu) 1
Background The medial frontal cortex (in particular, the orbital frontal cortex and the anterior cingulate cortex) has been implicated in a number of cognitive functions. Prominent among these is reinforcement-guided decision-making, or choice among alternative actions that is dictated by a comparison of values associated with those actions and based upon past experiences (Rushworth & Behrens, 2008). These action values are thought be calculated in the ACC and to integrate not only potential benefits but also potential costs (in the form of energy expended or time wasted) associated with a given course of action (Walton et al., 2003; Rudebeck et al., 2006a). Other functions ascribed to the ACC include generation of exploratory actions (actions that allow individuals to gather more information about novel environments), monitoring of action outcomes, and attribution of value to social information (Yoshida & Ishii, 2006; Walton et al., 2003; Rudebeck et al., 2006b). To what extent these various functions are related and supported by either overlapping or discrete regions of the ACC is unclear (Rushworth, Behrens & Rudebeck, 2007). The role of the ACC in valuation of socially relevant information is evidenced by the lack of interest in conspecifics shown by male macaques with ACC gyrus lesions (Rudebeck et al., 2006). In accord with these lesion studies, neuroimaging in humans has demonstrated increased activation in the rostral ACC while subjects perform interactive-decision making games such as the prisoner’s dilemma task (Rilling, 2002). It is interesting to consider the features of Autism Spectrum Disorders in relation to these findings, as these afflictions are strongly associated with deficits in social cognition (Volkmar, 2011). Indeed, previous studies have suggested that differences in both gross and microscopic anatomy of the ACC are present in autistic individuals. One MRI study found reduced right ACC volumes, and another described cytoarchitectonic differences in the form of decreased cell size and density in several ACC layers (Haznedar et al. 1997; Simms et 40 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
al., 2009). However, whether functional differences in the ACC—in addition to these anatomical differences—are present in autism is unclear. Research questions and direction In autistic individuals, the role of the ACC in the valuation of social information appears to be disrupted, so a preliminary goal of this investigation is to demonstrate this functional deficit using neuroimaging techniques. An fMRI analysis of activations in this region during a task requiring attention to and processing of social information would indicate whether there are indeed functional differences in the brains of autistics when compared to those of healthy controls. Such an experiment may reveal ACC activations in autistics distinct from those of a normal cohort. With this functional deficit established, it will be possible to determine whether other functions attributed to the ACC—such as the generation of exploratory actions (as in navigating a maze) and maintenance of a reward-outcome history—are supported by areas of the ACC distinct from those that are crucial for valuation of social information. The theory behind this experimental program is as follows: if the regions of the ACC involved in attribution of value to socially relevant information are separate from those necessary for its other functions, then autistic individuals engaged in ACC-taxing tasks should demonstrate fMRI activations and behavior similar to those of control subjects. If, on the other hand, the regions of the ACC that are dysfunctional in autism are the same as those required for proper performance of its other cognitive roles, then autistics should show behavioral deficits as well as altered activations that mirror the activation patterns observed during their performance of a social information task. In short, such functional imaging experiments comparing autistics and controls will help to establish a higher-resolution understanding of the functional neuroanatomy of the ACC.
Letter Methods Ideally, a large group of similarly-aged, right-handed adults with Autism Spectrum Disorder will be compared to a group of age- and sex-matched healthy control subjects. The autistic subjects will have no comorbid neurological or psychiatric disorders, and both groups will have no history of drug or alcohol abuse, head injury, or seizures. For the experiment considering differences in attribution of value to social information, subjects will complete social information tasks while in the scanner. These tasks include recognition of emotions and higher order mental states (e.g. flirtatiousness) in faces and viewing of video vignettes involving social interactions and then answering questions with button presses regarding the exchanges that occurred. Previous research indicates that autistics have difficultly identifying emotions in faces (Adolphs et al. 2001). Both fMRI data and behavioral data (i.e. response accuracy and reaction times) will be collected in order to not only replicate these behavioral findings but also associate them with specific neuronal activity patterns. In a second experiment, the same subject groups will navigate a virtual maze similar to the one used by Yoshida and Ishii (2006) with no directions while in the fMRI scanner. This task will necessitate the generation of exploratory action and should tax the ACC. Again, both functional imaging data and behavioral data will be collected. The measure of behavioral performance will be time taken to complete the maze, in addition to any more subjective indicators of difficulty with the maze or reluctance to explore it (e.g. asking for help). Possible results Given the observation that autism is characterized by deficits in attending to, acquiring and manipulating socially relevant information, it is likely that in the social information tasks there will be significant differences in both ACC activation and behavioral performance between the two subject groups (Dawson et al., 2004). In particular, it is predicted here that autistic individuals will exhibit smaller activations, lower accuracy and longer reaction times when identifying emotions and higher-order mental states in faces and when answering questions regarding social vignettes. Of course, the alternative is that ACC activations in autistics and controls are not significantly different. Such a result does not seem likely, however, given the known anatomical ACC abnormality in autistic brains (Ecker et al., 2012). The maze task would likely produce one of two different results. In the first scenario, ACC activations in autistics and controls would not be significantly different. The same would be true of their behavioral scores (that is, maze completion times would not differ substantially). This result would indicate that the area of the ACC that is affected in autism (and that gives rise to deficits in valuation of social information) is distinct from the area or areas of the ACC this are charged with generation of exploratory actions. Moreover, this region would appear to
NEURO Genesis
be intact in autism. The second possible result would be that the ACC activation and behavioral scores for autistics and controls are in fact significantly different. It seems probable that if this were the case, then the relative activations would parallel the result obtained previously in the social information task (reduced ACC activation), though in principle hyperactivity could be observed as well. Behaviorally, the autistic group might struggle more with exploring the maze freely than the normal cohort, which would be reflected in longer completion times and perhaps greater aversion to the task as measured by complaints or requests for assistance. Which pattern of results is the study more likely to observe? It is prudent to consider the behavioral symptomatology of Autism Spectrum Disorder when answering this question. The disorder is frequently associated with a number of behaviors that are suggestive of a general resistance to (or perhaps deficit in) novel action. These include an affinity for routines (often described as a strong need for “sameness”), perseveration, or getting stuck on a single topic or task, and the use of repetitive body movements (Gomot & Wicker, 2012). Each of these tendencies could be explained by a fundamental abnormality in the ACC circuitry necessary for the generation of exploratory actions. Thus, superficial signs suggest that the areas of the ACC that are dysfunctional in autism are involved in both social information valuation and exploration. The results of the virtual maze imaging experiment would provide evidence in support of or counter to this intuition. Potential problems One issue that could prove limiting to this program of research is the effect of autism on general cognition. For instance, it could be that the reason for autistics’ poor behavioral performances on social information tasks and exploratory action tasks is not a deficit in these areas of cognition per se but rather a general lack of intelligence that prevents them from comprehending the task well enough to perform at the level of healthy controls. This objection could be countered by being careful to select only high-functioning autistics for participation in this study. An IQ test could be administered to potential subjects prior to their full recruitment to the study, and a threshold score could be established to screen away those who might struggle with the task demands for reasons other than a specific abnormality in ACC circuitry. Another factor that could complicate this study is that as the name implies, Autism Spectrum Disorder is not a single neurological disorder but rather a range of disorders with different clinical manifestations (Happé, Ronald & Plomin, 2006). Whether the heterogeneity of autism cases stems from diverse etiologies with convergent phenotypes or from common etiologies with divergent phenotypes is unknown. In any case, the diversity of the disorder’s manifestations may mean that a range of functional ACC abnormalities is present in individuals with Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 41
NEURO Genesis
cases of varying severity. It may be that these abnormalities are only very significant on the end of the spectrum that does not include high-functioning autistics. If this were the case, then a study that screened out subjects from these populations for reasons pertaining to cognitive impairment described above would fail to characterize ACC functional properties in cases of autism with severe cognitive impairment. Despite this shortcoming, however, a study of only high-functioning autistics would still hold potential to reveal important functional neuroanatomical features of autism spectrum disorder. Conclusions and areas for further development The ACC has been implicated in the attribution of value to social information, the generation of novel exploratory actions, and decision-making based on the maintenance of records of rewards and costs associated with various courses of actions (Rudebeck et al., 2006b; Yoshida & Ishii, 2006; Walton et al., 2003). To what extent these functions occupy overlapping regions within the ACC is uncertain. The program of research described here suggests that experiments comparing brain activations and behaviors of autistics—individuals with a clear deficit in one of these ACC functions—to those of healthy controls might resolve some of this uncertainty. In particular, these experiments aim to establish clear differences in activations during tasks involving social information between these two groups. They also promise to elucidate the relationship between regions of the ACC that support social information valuation and exploration. A general distaste for novelty and change evidenced by a number of common autistic behaviors suggests that these two cognitive functions may depend on the same brain regions (and that these regions are dysfunctional in autism). If the approach of using comparisons between ACC activations and behavioral traces of both autistics and controls is successful in delineating the shared or distinct areas supporting social information valuation and exploration, then future studies could carry out a similar analysis of the ACC’s function in decision making. A well-designed task could reveal whether reinforcement-guided decisionmaking too requires ACC circuitry that is pathological in autism. Information gained from studies like these in conjunction with further characterization of anatomical and mechanistic features of autism will help to elucidate the roles of the various regions of the ACC in executive function.
42 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Articles Adolphs, R. et al. (2001) Abnormal processing of social information from faces in autism. J. Cogn. Neurosci. 13, 232-240. Dawson, G. et al. (2004) Early social attention impairments in autism: social orienting, joint attention and attention to distress. Dev. Psychol. 40, 271-283. Ecker, C. et al (2012) Brain anatomy and its relationship to behavior in adults with autism spectrum disorder. Arch. Gen. Psychiatry. 69, 195209. Gomot, M. and Wicker, B. (2012) A challenging, unpredictable world for people with Autism Spectrum Disorder. Int. J. Psychophysiol. 83, 240247. Happé, F., Ronald, A., and Plomin, R. (2006) Time to give up on a single explanation for autism. Nat. Neurosci. 9, 1218-1220. Haznedar, M.M. et al. (1997) Anterior cingulate gyrus volume and glucose metabolism in autistic disorder. Am. J. Psychiatry 154, 1047-1050. Rilling, J.K. (2002) A neural basis for social cooperation. Neuron 35, 395405. Rudebeck, P.H. et al. (2006a) Separate neuronal pathways process different decision costs. Nat. Neurosci. 9, 1161-1168. Rudebeck, P.H. et al. (2006b) A role for the macaque anterior cingulate gyrus in social valuation. Science 313, 1310-1312. Rushworth, M.F., Behrens, T.E., Rudebeck, P.H., and Walton, M.E. (2007) Contrasting roles for cingulate and orbitofrontal cortex in decisions and social behaviour. Trends Cogn. Sci. 11, 168-176. Rushworth, M.F. (2008) Intention, choice, and the medial frontal cortex. Ann. N Y Acad. Sci. 1124, 181-207. Simms, M.L. et al. (2009) The anterior cingulate cortex in autism: heterogeneity of qualitative and quantitative cytoarchitectonic features suggests possible subgroups. Acta Neuropathol. 118, 673-684. Volkmar, F.R. (2011) Understanding the social brain in autism. Dev. Pyschobiol. 53, 428-434. Walton, M.E. et al. (2003) Functional specialization within medial frontal cortex of the anterior cingulate for evaluating effort-related decisions. J. Neurosci. 23, 6475-6479. Yoshida, W. and Ishii, S. (2006) Resolution of uncertainty in prefrontal cortex. Neuron 50, 781-789.
NEURO
Opinion
Genesis
Society’s stake in biomedical cognitive enhancement Akash Shah1 1 Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Akash Shah (akash.shah@duke.edu)
SUMMARY: As biomedical technology grows more and more powerful, the ethical questions that arise become increasingly important to answer. As a society, we fear new technologies that threaten to change our biology – biomedical cognitive enhancement (BCE) is one such technology. I argue that this fear is unwarranted. Arguments leveled against biomedical cognitive enhancement–that it is unnatural, inauthentic, and may widen gaps between sectors of society–can easily be rebutted. Furthermore, arguments in favor of BCE show that it has much to offer society. It can increase the cognitive capacity of the overall society, reducing the risk of many social and economic problems, including bad health, accidents, and mortality. An appropriate government policy regarding BCE is also outlined, the objectives of which are to benefit society while maintaining impartiality and equality. If BCE is indeed found to have large undeniable benefits, it would be in the best interest of the government to subsidize BCE. A system of independent professional panels and thorough background checks would be used to most efficiently assess the risks associated with BCE. Given the positive effects this technology may have on society, it is morally necessary to pursue BCE. Introduction Biomedical cognitive enhancement (BCE) involves the use of drugs or therapies to change the biology of individuals, leading to cognitive enhancement. Neurobiologist Anders Sandberg and philosopher Nick Bostrom define cognitive enhancement as an amplification of the processes used to organize and use information to guide behavior (Bostrom & Sandberg, 2009). Cognitive enhancement, as with any other type of enhancement, is often met with vehement opposition at first mention. Upon closer consideration of what cognitive enhancement truly is, however, many would agree that it has much to offer society. After all, examples of cognitive enhancements such as literacy and numeracy underline the powerful effect that widespread cognitive enhancement can have. We may consider literacy, numeracy, and even computers as part of our normal capacity, but this is, of course, not the case. These were all vast cognitive enhancements at one point in human history, without which modern society as we know it would not exist. Taking these examples into account, it is conceivable that many opponents of BCE would concede that cognitive enhancement has the potential to dramatically better society. Biomedical cognitive enhancement, however, would likely not receive such a concession. The public is vastly misinformed on the biomedical enhancements. As the prospect of BCE grows more and more real, the necessity for an intelligent debate on the matter has become more pressing. Contrary to the claims of BCE opponents, BCE is not morally wrong. In fact, not pursuing BCE, as a society. would be morally inexcusable. Once we accept this, an appropriate government policy of subsidization must be implemented.
Arguments against biomedical cognitive enhancement Since a discussion regarding BCE policy would only be reasonable given that BCE is deemed biologically safe, let us assume, that no safety concerns exist with BCE. Having established that, we can now move forward in discussing the arguments made against BCE. A common argument made against BCE is that it is an affront to the “wisdom of nature” or that it goes against “eons of gradual and exacting evolution” (Sandel, 2004). While such phrases appeal to emotion, it is important to realize that nature is neither wise nor foolish. Evolution simply leads to organisms that are well-adapted to their respective environments. Even then, a population is positively adapted through the eradication of its less fit individuals–a process that often lasts millions of years. Evolution does not have a goal, nor is it efficient. Allen Buchanan argues, evolution simply cobbles together unstable products, most of which will fail anyway. In fact, this supposed wisdom of nature is what modern medicine has been trying to counter for the last century. Since natural selection does not act on post-reproductive traits, diseases which strike during the post-reproductive age are not actively selected against. Surely, opponents of BCE do not oppose modern medicine acting against the natural. Another charge leveled at BCE is that it is unnatural. In this line of reasoning, critics of BCE are referring to the natural as that which is found in nature; the unnatural, therefore, is artificial. Accepting that the unnatural is a departure from the natural to the artificial, it is interesting that the unnatural holds a negative connotation. After all, human civilization itself is an affront to nature. As John Stuart Mill said: “If the artificial is not better than the natural, to what end are all the arts of life? To dig, to
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 43
NEURO Genesis
plough, to build, to wear clothes, are direct infringements of the injunction to follow nature” (Mill, 1874). So why is it that the unnatural holds such a negative connotation? The answer to this question is often answered by religion. Arguments from religion often state that nature is God’s way and should thus not be fiddled with. They contend that if it is natural, it must be correct, or else it would not be there. Unfortunately, they are sorely mistaken. Natural disasters that wipe out animal and human populations do not provide any benefit. Conversion of certain molecules into toxic isomers provides no advantage. The imperfect nature of DNA replication leads millions to suffer from debilitating cancers. The examples can go on and on. Without diving into the theological implications of these examples, one can at least accept that nature is defective. Nature does not run smoothly; it is capricious and impersonal, often leading to pain and suffering. Thus, it soon becomes clear that simply because something is natural does not make it good. Likewise, simply because something is unnatural does not make it bad. Another argument used against BCE is that it is inauthentic. In this case, the conception of authenticity is that of being true to oneself. Critics of BCE charge that changing ones given intelligence through the use of BCE is inauthentic since it would change the given self; and so, the user would no longer be acting true to oneself. However, BCE need not be authentic if the individual’s use of it is in accordance with his/her conception of himself/herself. That is, if an individual truly believes he/she needs to be cognitively enhanced, the use of BCE is justified since it is in agreement with the individual’s notion of himself/ herself. Another aspect of BCE that bothers many of its opponents is the fact that that BCE may change human nature. However, what is wrong with changing human nature? Human nature includes both good and bad characteristics. There is no reason to believe that these characteristics are so interconnected that it is impossible to correct the bad without damaging the good. Many critics charge proponents of BCE for being motivated by arrogance and hubris. I have shown that changing oneself through BCE is not morally problematic, so the problem that must be addressed is cognitively enhancing one’s children. Critics say that parents are motivated by hubris when cognitively enhancing their children. It is foolish to think that this enhancement is about the parents. There is no reason to believe that the parents are acting out of hubris or some desire to master the mystery of nature, a stance that political philosopher Michael Sandel, among many others, strongly supports (Sandel, 2004). Unfortunately, Sandel is missing the entire point of why parents would pursue enhancement. One must realize that the parents are enhancing their children not out of hubris but out of love. It is not that the parents are refusing to accept unenhanced children; they simply want their children to be better, in certain respects, than they would be if they were to remain unenhanced. 44 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Opinion Opponents of BCE are also worried about inequality due to enhancement. They feel that cognitive enhancement would widen the already large intelligence gap between different sectors of society. However, there is no reason to believe that cognitive enhancement will create any more gaps than are already present in society. For example, gaps in intelligence exist between those in a university and those in an isolated village. This is due, of course, to differences in academic facilities and resources. However, such inequalities are abundant, and there is no convincing reason to believe that BCE will exacerbate these existing inequalities. Interestingly, such questions of inequality are not raised when new technological advancements are made – even when these technologies may have the very effect that critics of BCE are afraid of. As a society, we must be careful not to fall into the trap of biomedical enhancement exceptionalism. That is, we must not hold biomedical enhancement to a different standard than that of other kinds of enhancements. BCE critics go further to say that BCE would give an unfair advantage to those who receive the enhancement. Such unfair advantages already exist in the form of family wealth, resources, private tutors, etc. As in the previous case, it is unreasonable to assume that BCE will exacerbate these inequalities. However, one should not dismiss this worry as it is a serious one. Whether an innovation narrows or widens inequalities may be within our control. Appropriate policy measures could be put in place to ensure that inequalities are not widened. Arguments for biomedical cognitive enhancement History has shown that cognitive enhancements can have great effects; one needs only to look at the examples of numeracy, language, and the Internet to see this. But is there reason to believe that BCE will have a positive effect on society? When contemplating the kinds of enhancements BCE could provide, it soon becomes clear that BCE can potentially have a dramatic positive effect on society. BCE would promote more educated choices from individuals, since the ability to retain reliable information and to use it accordingly would be enhanced. Since cognitive capacity is positively correlated with desirable outcomes, BCE would reduce the risk of many social and economic problems, such as bad health, accidents, and mortality (Bostrom, & Roache, 2009). Furthermore, improving the cognition of many individuals could lead to huge benefits for society; BCE is not zero-sum. A concern some have is that cognitive enhancement only provides positional goods and would therefore lead to a kind of competitive arms race that would lead to no overall societal advantage (Bostrom & Roache, 2009). If BCE was to become widespread, the cognitive capacity of the overall society would obviously increase. A society that makes better use of its brainpower would allow it to perform more efficiently in relation to a previous un-enhanced society. Theoretically, individual capacity would increase, as
Opinion would the group’s capacity. Society would coordinate better as a result. Therefore, there is a large societal gain in pursuing BCE. We must not allow the bias of the status quo to let society miss out on valuable goods. If it is definitively found that society stands much to gain from BCE – as it may well turn out – it may become morally necessary to pursue BCE for all. Not doing so would be morally unacceptable. Forcing society to operate at a lower level than it could be is morally inexcusable. Appropriate government policy Since the prospect of implementing BCE in the public sphere is growing more realistic, I feel that it is important to outline an appropriate government policy. The objectives of this policy are to ensure that society benefits while impartiality, equality, and justice are still maintained. This policy outline does not address germ-line BCE because given our current understanding – or lack thereof – of the effect of human germ-line changes, I feel that it would be unwise to include it in the policy at this stage. Controlled trials of germ-line modification are needed to observe its effects on human biology. If BCE is found to yield such large and undeniable benefits, it would be in the interest of the government to subsidize it. After all, to combat the ills of present-day society, we have to give ourselves such enhancement. If we do not, it is inevitable that another nation will enhance itself, putting our nation at a distinct disadvantage. Furthermore, as argued above, there is a public interest. An example of a subsidization of an enhancement is public education. Basic education was subsidized by the government because the government realized that it had a stake in the betterment of society due to education. An important question that must be addressed is: what should be enhanced? I do not think that the specifics of what cognitive abilities to enhance are particularly important. Any enhancement that falls under the domain of cognition (memory, computation, etc.) should be allowable. What is important, however, is to accept that there is risk involved. It would be foolish to think that BCE is completely risk-free – very few advances are. And so, a minimum allowable risk should be determined. That is, an acceptable risk-to-benefit ratio must be established. Since risks do exist, there should be a process of deciding who should have access – especially in the case of particularly powerful enhancements. Though equal access is a tenet of this policy, Ingmar Persson’s and Julian Savulescu’s point regarding irresponsible use of BCE cannot be ignored (2008). BCE in unstable hands, or in the hands of those with evil intentions, could have the potential to do much harm. So, the question stands: how do we assess who should or should not receive BCE? I propose multiple independent panels of psychologists and psychiatrists to judge whether an individual is mentally stable enough for
NEURO Genesis
BCE. To minimum government intervention, these panels would be determined by a non-governmental organization such as the American Medical Association. It would be a waste of time and resources, however, to have every individual who wants BCE face the panels. Having the panels conduct background checks to pinpoint individuals that would pose a substantial risk is a much more efficient way of assessing risk. A committee responsible for ensuring that the implemented BCE is safe would also be established. To ensure as little government intervention as possible, the committee would be selected by an independent scientific organization such as the Howard Hughes Medical Institute. Having established how to decide who gets access, the next question that arises is regarding the eligibility of individuals for enhancement. By this, I mean: what biological conditions must one have in order to receive the enhancement? In this particular policy, I would say that no medical problem is required in order to receive BCE. After all, if the intervention was a response to a medical problem, it would no longer be enhancement but therapy. As long as an individual passes the initial background check, he/she is eligible to receive the enhancement. By requiring that only those with disadvantages are eligible for BCE, one of the key purposes of government subsidization of BCE would be defeated. The benefit of BCE to society would be maximized by having as many people as possible undergoing enhancement. Before complete subsidization by the government, a voluntary enhancement program should be put into place. To be a part of this enhancement program, individuals must first pass a background check as well as the psychology/psychiatry panel. Also, interested individuals must pay a sizeable fee to receive the enhancement. It is conceivable that there would be many wealthy individuals willing to pay in order to receive an enhancement that would increase their cognitive powers. These fees will help fund the future subsidization of BCE. The individuals would be put under medical surveillance to monitor the effects of the BCE. This pilot program will lead to a better product for the rest of society and potentially cut costs. This is often the case with new technology. The rich buy a sub-par product at a higher price and after the initial flaws are ironed out, the rest of society gets a better product at drastically reduced costs. It is imperative that BCE remains the individual’s choice. The individual must be told of all the risks and benefits involved, but then left to let him/her make the final choice. If the individual chooses to pursue BCE, he/ she must pass a written test before he/she can proceed to apply for the background check. If he/she chooses not to, then the individual’s choice must be respected. Though it is of utmost importance that the individual makes the final choice, it would be allowable to incentivize enhancement if it is found that society would further benefit if the entire population was enhanced. Using economic incentives – Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 45
NEURO Genesis
or even economic punitive measures – would be an effective method of getting individuals to enhance themselves. Furthermore, we must make sure that social coercion and pressure do not exist. Furthermore, the government must ensure that individuals without BCE are not put at a disadvantage, be it socially, economically, or politically. It should create a new body responsible for overseeing the effects of BCE on society that can help avoid inequalities due to the enhancements. This body would be composed of social scientists and medical professionals from the private sector. This policy is well and good, resources permitting. But resources, of course, are not necessarily permitting. If they are limited, how can we determine who should get BCE? The panels and subsidization should still exist, simply for a more select candidate pool. The question is: who is this group comprised of? There are two options: should we enhance the poorest and least educated or should we enhance the most educated? I would propose that the latter be enhanced. There is more for society to gain should the academics and researchers be enhanced, because they are the ones whose cognition significantly contributes to the groundbreaking discoveries/inventions that the rest of society benefits from. A way of thinking about this is to imagine society to be a bank of 1,000 computers of varying speeds. Now imagine that there is an enhancement that would allow the computers to become 50% faster. However, there is only enough software for 100 computers. Which choice would increase efficiency: enhancing the 100 slowest computers or enhancing the 100 fastest computers? If the former was chosen, the slowest computers would be brought up to par to the computers of average speed, but the overall capacity of the computer bank does not increase. If the latter was chosen, however, the capacity of the computer bank increases due to higher speeds than ever before. Critics may say that such a stance is heartless, cold, utilitarian, etc. I admit, this is not the ideal situation, just one with limited resources. That being said, parallels drawn between this stance and heartless utilitarianism are wholly unwarranted. No one is being hurt; it is not as if a select group of people are being cured of a deadly disease while the rest are left to suffer. No one is at an advantage or disadvantage, since BCE in this case is not a positional good. However, since there are many who will remain unenhanced, considerations of fairness are relevant. I would argue that the indirect benefits to the unenhanced along with the benefit of overall efficiency outweigh the costs of leaving some individuals unenhanced. Conclusion Most of the opposition to biomedical cognitive enhancement is simply due to the fact that it is new technology. The medical procedure of in vitro fertilization suffered through the same opposition; people are always apprehensive about new technologies even if they do not have good reason to be. This is certainly the case with 46 | neurogenesisjournal.com | Fall 2012 | Vol2 Issue 1
Opinion BCE. And so, it is important to realize that, contrary to the claims of its vocal opponents, BCE is not immoral. In fact, a closer look at BCE shows that it has much to offer society – namely, risk reduction of social/economic misfortunes and improving social productivity. Given these positive effects BCE may have on society, perhaps not pursuing BCE is an immoral action. Since society has much to gain from BCE, a policy where BCE is subsidized by the government is necessary for the betterment of society. Bostrom, N, & Sandberg, A. (2009). Cognitive enhancement: methods, ethics, regulatory challenges. Science and Engineering Ethics, 311-341. Sandel, M. (2004, April). The case against perfection. The Atlantic Monthly, 50-62. Buchanan, A. “Conservatism and the enhancement project.” (ms). 14. Mill, JS. (1874). Nature, the utility of religion, and theism. Watts and Co. Sandel, M. (2004, April). The case against perfection. The Atlantic Monthly, 50-62. Bostrom, N, & Roache, R. (2009). “Smart policy: cognitive enhancement and the public interest.” Enhancing Human Capabilities. Ed. Savulescu, J., ter Muelen, R. & Kahane, G. Oxford: Wiley-Blackwell, 2009. Persson, I, & Savulescu, J. (2008). The perils of cognitive enhancement and the urgent imperative to enhance the moral character of humanity. Journal of Applied Philosophy, 25(3), 162-177.
NEURO
Articles
Genesis
The α4β2 neuronal nicotinic acetylcholine receptor: Developing new therapeutic agents for smoking cessation based on cytisine and bupropion Sheetal R. Hegde1, Akash Pandhare, and Michael P. Blanton. Duke University, Durham, North Carolina 27708 Correspondence should be addressed to Sheetal Hegde (srh34@duke.edu) 1
SUMMARY: Nicotinic acetylcholine receptor (nAChR) agonists have potential therapeutic benefits both for various neurological diseases and the treatment of nicotine addiction. This study investigated the ligand binding site of α4β2, the most common subtype of nAChR in the brain. Lowry protein assays were conducted in order to determine the protein concentrations of α4β2 receptor-containing HEK (Human Embryonic Kidney) cell membranes. Saturation radioligand binding was conducted to determine the number of specific binding sites on receptors for the nAChR agonist epibatidine. Bupropion, an antidepressant also involved in smoking cessation, was examined. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed to analyze photoaffinity labeling of receptors using the bupropion analog [125I] SADU3-72 and to demonstrate differences in labeling between [125I] SADU3-91 and [125I] SADU3-72. These processes allowed for the observation of photoreactivity and identification of binding sites in both Torpedo (a type of fish) and α4β2 nAChRs. It was observed that the efficiency of [125I] SADU3-91 labeling as compared to [125I] SADU3-72 was less than 1%. [125I] SADU3-72 was photoreactive due to the presence of the photoreactive azido group (N3). The pattern of labeling as well as inhibition by bupropion are consistent with a specific binding site in the Torpedo and α4β2 nAChRs that was further mapped to the channel of the Torpedo nAChR. Introduction Nicotinic acetylcholine receptors (nAChRs) are members of the Cys-loop superfamily of ligand gated ion channels expressed in the central nervous system, the peripheral nervous system, and the neuromuscular junction (Exley et al., 2006). The α4β2 receptor is a heteromeric as well as pentameric membrane protein formed by two α4 and three β2 subunits and is the most common nAChR subtype found within the brain. Important to note are the agonist binding sites (ABS) of the nAChRs located in the N-terminal extracellular domain. The α4β2 nAChR has two ABS. Nicotine serves as an agonist, stimulating the receptor. This receptor has gained attention due to its association with diseases such as Alzheimer’s disease, Parkinson’s disease, epilepsy, and schizophrenia, as well as its importance to learning and memory. However, the α4β2 receptor is also largely related to nicotine addiction and dependence. It has a high-affinity binding site for nicotine and is localized to a region of the brain that regulates release of the neurotransmitter dopamine (mesolimbic dopamine neurons). This purports a relationship between the above-mentioned diseases and nicotine addiction while emphasizing the importance of specifically targeting the α4β2 receptor.
In order to reduce side effects of current drugs, subtypeselective ligands must be developed. Furthermore, direct targeting of nAChRs has the potential to decelerate the development of neurodegenerative diseases. Nicotinic acetylcholine receptors are importnat presynaptically and postsynaptically. Presynaptically, these receptors help regulate the release of neurotransmitters .(University of California, San Francisco, 2011). Postsynaptically, the receptors facilitate fast synaptic transmission. Nicotine addiction has shown to affect about a third of the global population (World Health Organization, 2011). Smokers depend on smoking in order to sustain the brain’s nicotine levels and evade the harmful effects of withdrawal. Daily smokers are also shown to demonstrate higher stress levels than nonsmokers (Rosenblum, 1938). Varenicline, a drug for smoking cessation, is a nAChR partial agonist that binds to the receptor. Although varenicline is a partial agonist for α4β2, it is a full agonist at α7 and other neuronal nAChRs. This induces harmful side effects such as neuropsychiatric problems (including suicide). Cytisine, a partial agonist currently used in Europe for smoking cessation, is photoreactive and has a high affinity for the α4β2 receptor. Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 47
NEURO Genesis
Articles Experiment Materials [3H]Epibatidine (45 Ci/mmol) was obtained from PerkinElmer Life Sciences and stored in 95% ethanol at 4°C. Carbamylocholine chloride, diisopropylfluorophosphate, and carbamylcholine chloride were obtained from Sigma. Epibatidine and D-tubocurarine (dTC) were obtained from Tocris (Ellisville, MO), and Staphylococcus aureus endoproteinase Glu-C (V8 protease) was from Worthington. Endoproteinase Lys-C (EndoLys-C) and
Figure 1. α4β2 receptor schematic (Rollema et al., 2007).
Nicotine binding also involves a cation-π interaction with the receptor. The high affinity for nicotine stems from the strong, noncovalent cation-π interaction with TrpB, an aromatic amino acid (Xiu et al., 2009). This binding is much tighter than that of the muscle type nicotinic receptor. Nicotine has a stronger interaction with the neuronal receptor as opposed to the muscle-type receptor primarily due to an amino acid difference near TrpB. Moreover, at a pH greater than 5, nicotine becomes deprotonated which makes it highly membrane-permeant. Thus, nicotine has the ability to rapidly cross the blood brain barrier. Desensitization and upregulation of α4β2 are also responsible for the addictive nature of nicotine (Zhang et al., 2003). Acetylcholine also has a strong cation-π interaction with TrpB. Bupropion is an antidepressant which also aids in smoking cessation by reducing the effects of withdrawal (Drugs Information Online, 2011). Bupropion is an antagonist which functions by blocking the dopamine transporter which, when unblocked, facilitates the uptake of dopamine into neurons (Chen et al., 2000). However, bupropion binds to this transporter and increases the level of dopamine in the synapse which provides bupropion with its antidepressant quality. Bupropion also serves to inhibit the nAChR but has biorhythmic side effects such as nausea, headaches, and weight loss (Mental Health Channel, 2011). Lowry protein assays were conducted to determine the total protein concentration of α4β2 receptor-containing HEK cell membranes. Saturation radioligand binding was then conducted to determine the amount of specific binding sites for epibatidine. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was conducted to analyze photoaffinity labeling of the receptor with bupropion analogs and provided a comparison between [125I] SADU3-91 and [125I] SADU3-72 labeling. This test allowed for the observation of photoreactivity and identification of bupropion binding sites in the α4β2 receptor. 48 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Figure 2. Structure of cytisine (C11H14N2O).
protease inhibitor mixture set III were obtained from Calbiochem. Sodium cholate and CHAPS were obtained from U.S. Biochemical Corp. Affi-Gel 10 was obtained from Bio-Rad. Synthetic lipids (dioleoyl phosphatidic acid and dioleoyl phosphatidylcholine), cholesterol, asolectin, and total lipid extract from porcine brain were obtained from Avanti Polar Lipids (Alabaster, AL). Cell culture and membrane preparation HEK-293 cells transfected with human α4β2 nAChRs (HEK-hα4β2) were obtained from Dr. Joseph H. Steinbach (Department of Anesthesiology, Washington University School of Medicine, St. Louis, MO). The cells were grown at 37⁰C in a humidified incubator at 5% CO2, in 140 mm tissue culture dishes. The dishes were maintained in Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F-12 (Mediatech, Inc.) which was supplemented with 10% fetal bovine serum, 100 units/mL penicillin G, 100 µg/mL streptomycin, and 450 µg/mL Geneticin (G418) as a selection agent. Gentle scraping was employed in 5 mL growth medium with protease inhibitor cocktail III (Calbiochem, 0.2 µL/mL) present. The cells were then pelleted by centrifuging (210 g for 4 min), resuspended in vesicle dialysis buffer (VDB; 100 mM NaCl, 0.1 mM EDTA, 0.02% NaN3, 10 mM MOPS, pH 7.5), and pelleted again by centrifuging. The final cell pellet was stored
Figure 3. Structure of nicotine (C10H14N2).
NEURO
Articles
Genesis
at -80⁰C. For membrane preparation, HEK-hα4β2 cells were thawed and homogenized in VDB with protease inhibitor cocktail III present (Calbiochem, 1 µL/mL) using a glass homogenizer. Membrane fractions were pelleted by centrifuging (39000g for 1 h). Then, the membrane pellets were resuspended in VDB (approximately 0.5 mL/140 mm dish). Protein assay 250 µL of 0.85% NaCl followed by 250 µL of SIGMA protein standard (Bovine albumin 2 mg/mL) were added to the Standard vial. The contents in the vial were then vortexed. 0.85% NaCl was added to 9 triplicates in the following concentrations: 100 µL, 98 µL, 96 µL, 94 µL, 92 µL, 90 µL, 86 µL, 80 µL, and 70 µL. 90 µL were added to 3 separate vials reserved for the protein samples. Next, the Standard solution was added to each vial of the 9 triplicates in the following concentrations: 0 µL, 2 µL, 4 µL, 6 µL, 8 µL, 10 µL, 14 µL, 20 µL, and 30 µL. Each vial’s concentration of the 9 triplicates should be 100 µL. 10 µL of protein sample were added to each of the vials reserved for the samples. Solution D (40 mL 2% Na2CO3 in 0.1N NaOH + 400 µL 2% Na/K tartrate in H2O + 400 µL 1% CuSO4 5H2O in H2O) was then prepared and stirred. 1 mL of D was added to each vial while stirring. Each tube was then vortexed and incubated for 30 min. 1:1 diluted Folin and Ciocalteu’s Phenol Reagent was then prepared by adding 2 mL reagent to 2 mL water. The contents were then stirred, and 100 µL of 1:1 diluted Folin and Ciocalteu’s Phenol Reagent was then added to each vial while vortexing. The vials were then allowed to incubate for 30 minutes and were then read the A750 using a spectrophotometer (Beckman Instruments). Preparation of Torpedo nAChR- Torpedo californica nAChR-rich membranes for radioligand binding studies were isolated from frozen electric organs (Aquatic Research Consultants, San Pedro, CA). Torpedo nAChRrich membranes at 1 mg/mL protein were solubilized in 1% sodium cholate in vesicle dialysis buffer (VDB; 100 mM NaCl, 0.1 mM EDTA, 0.02% NaN3, 10 mM MOPS, pH 7.5) and treated with 0.1 mM diisopropylfluorophosphate after insoluble material was pelleted by centrifuging (91,000 x g for 1 h).
Figure 4. Structure of bupropion (C13H18ClNO).
The nAChR was affinity-purified on a bromoacetylcholine bromide-derivatized Affi-Gel 10 column and then reconstituted into lipid vesicles made of dioleoyl phosphatidic acid/dioleoyl phosphatidylcholine/cholesterol (molar ratio 3:1:1). The lipid to nAChR ratio was adjusted to 400:1. The nAChR-rich membranes and purified nAChRs were stored at -80⁰C. Saturation radioligand binding Vesicle dialysis buffer (VDB) was added to 16 Eppendorf tubes as follows: 99.0 µL, 98.0 µL, 97.0 µL, 96.0 µL, 95.0 µL, 94.0 µL, 92.0 µL, 90.0 µL, 85.0 µL, 80.0 µL, 75.0 µL, 70.0 µL, 65.0 µL, 60.0 µL, 55.0 µL, and 50.0 µL. Then, 35.91 nM [3H] stock solution was added to the respective tubes as follows: 1.0 µL, 2.0 µL, 3.0 µL, 4.0 µL, 5.0 µL, 6.0 µL, 8.0 µL, 10.0 µL, 15.0 µL, 20.0 µL, 25.0 µL, 30.0 µL, 35.0 µL, 40.0 µL, 45.0 µL, and 50.0 µL. 50 µL α4β2 membranes were added to each tube. Each tube was vortexed well and allowed to incubate overnight at room temperature. The Eppendorf tubes were centrifuged in a JA-20 rotor (18,000 for 1 h at 4⁰C) using a rubber adaptor. 50 µL were transferred from each supernatant of the tubes reserved for total binding to a scintillation vial. 3 mL Bio-Safe II scintillation fluid were added and the vials were allowed to incubate for 1 h to obtain free [3H] Epibatidine radioactivity. They vials were then counted for 5 min. The supernatants were then removed, and the tubes were inverted to allow drainage for 30 min. Residual liquid was carefully removed. The pellets were then resuspended in 200 µL of 10% SDS and incubated overnight. They were then transferred to 5 mL scintillation vials. The tubes were then spun down in a Eppendorf microfuge briefly for about 6-8 seconds, and the remaining volume was transferred to the vial. 3 mL Bio-Safe II scintillation fluid was added, and the vials were allowed sit for 1 h. They were then counted for 5 min. The same procedure as listed was used except in the presence of 150 µM Epibatidine in order to obtain nonspecific [3H] Epibatidine binding. Photolabeling of native Torpedo AChR membranes One tube of native Torpedo AChR membranes (6/09 Box- 1 mg per tube) was thawed. Lights were dimmed for photoreactivity. Membranes were diluted to 2500 µL and 25 µL [125I] SADU3-91 (6.75 uCi; 1.37 nM) were added and mixed thoroughly. These membranes were divided into 2 x 1250 µL aliquots. For the [-] set, 25 µL of 250 uM αBgTx (5 uM αBgTx) were added. For the [+] set, 25 µL 20 mM carbamylcholine (400 uM) were added. Each set was mixed thoroughly. These membranes were divided into 4 x 625 µL aliquots. The -/- set was designated as the resting state control (with αBgTx). 8 µL of 10 mM tetracaine (130 uM) were added to the -/+ set (with αBgTx). The +/- set was designated the desensitized state control (with carbamylocholine). 8 Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 49
NEURO Genesis
µL of 10 mM tenocyclidine (TCP) (130 uM) were added to the +/+ set. The aliquots were mixed thoroughly and incubated for 1 hour at room temperature. The aliquots were photolyzed at 254 nm for 10 minutes at < 1cm. The aliquots for each condition were mixed, and then 312 µL were transferred to an Eppendorf tube. The 8 tubes were then spun at 18,000 rpm for 1 hour in JA-20. Each pellet was then resuspended in 70 µL of 1X electrophoresis sample buffer (ESB). The solutions were then run on a 1.0 mm 8% polyacrylamide gel (10-well comb). The gel was stained for 30 minutes and then destained for 3-6 hours. The gel was dried down and put on film for 4 days using an intensifying screen. It can be observed that the average analytical absorbance consistently increases with increasing concentrations. Moreover, the high correlation coefficients confirm the accuracy of these results. Saturation radioligand binding experiments were then conducted and the results are as follows. The saturation radioligand binding served to isolate the receptor in order to find the concentration by adding a ligand (Epibatidine)
Figure 5. Structure of TCP (C15H23NS).
Figure 6. Human α4β2 HEK-293 Membranes (+ nicotine).
50 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Articles which only binds to the available nAChR. The Bmax is the maximum specific binding (total – non-specific), and is observed by the middle curve of Figure 6A. Here, it is displayed that the Bmax is 1.8 pmol bound epibatidine/mg protein. However, the receptor has two binding sites allowing two molecules of epibatidine bind to one receptor. Thus, the concentration of receptor is 0.9 pmol/mg. Kd, the concentration of ligand that binds fifty percent of the receptors, was found to be 1.1 nM. Photoaffinity labeling and SDS-PAGE Photoaffinity labeling allows macromolecules to be labeled at or near the active binding site. A label was attached to the active site of the large molecule and allowed for the identification of the binding site. SDS-PAGE, a gel electrophoresis method, utilizes electricity to induce mobility of proteins through the gel and allow for separation by size. From the graphs, it is evident that the efficiency of labeling of [125I] SADU3-91 as compared to [125I] SADU372 was less than 1%. [125I] SADU3-72 was clearly photoreactive and was expected to have more labeling than [125I] SADU3-91 due to the presence of the photoreactive azido group (N3). This photoreactive behavior allows for future proteolysis (digestion of proteins) which then facilitates the identification of amino acids. For the sets containing tetracaine (-/+), there was a decrease in labeling which was expected due to tetracaine’s ability to block channels and thus displace the radioactive ligand. This represented the resting state. However, nAChRs are prone to desensitization, and TCP has more affinity for the desensitized state. TCP also serves to displace the radioactive drug. By adding tetracaine and TCP, specificity is shown. It is also shown that the drug blocks the channel. Future research This research aims to determine the ligand binding site which may allow for development of drugs that specifically bind to the α4β2 receptor. Drugs that specifically target the α4β2 nAChR will allow for refinement of current drugs utilized for smoking cessation and a reduction in the side effects associated with these drugs. Moreover, the photoaffinity and SDS-PAGE results demonstrate that the photoreactive behavior of the azido group is significant to this study because it covalently tags the binding site and allows for proteolysis steps which will allow for the identification of labeled amino acids within the binding site. This identification will provide detailed information about the binding site to help develop new drugs based on bupropion and cytisine. The next step in this process would include a repetition of the photolabeling except with use of the purified α4β2 receptor.
NEURO
Articles
Genesis
Figure 8. Photolabeling of [125I] SADU3-91 (top) and [125I] SADU3-72 (bottom).
Figure 7. Autoradiographs of [125I] SADU3-91 (top) and [125I] SADU3-72 (bottom). After one week, the [125I] SADU3-91 demonstrates only slight trace labeling while after only 6.5 hours, the [125I] SADU3-72 displays significant photolabeling.
Figure 9. Structure of Azido Group.
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 51
NEURO Genesis
Articles
Acknowledgments I would like to thank Dr. Michael P. Blanton and Akash Pandhare for providing me with the opportunity to work with them. This experience has been incredible, and I have learned so much from their guidance. I would also like to thank Dr. Joseph H. Steinbach (Washington University School of Medicine) for providing the stable α4β2 nAChR HEK-293 cell line. I would also like to thank Dr. Jose-Luis Redondo (Department of Pharmacology and Neuroscience, Texas Tech University Health Sciences Center) for assistance in cell culture. I would also like acknowledge Dr. Michael San Francisco and Ms. Lynda Durham for their encouragement throughout this process. Their patience and support always encouraged me and made this experience even more wonderful. I am also extremely grateful for my family for supporting me in all my endeavors. Finally, I would like to express my deep gratitude towards the Anson L. Clark Foundation and Dr. John M. Burns, the sponsor of this work and Founding Director of this program, respectively. The Anson L. Clark Foundation provided me with the opportunity to conduct this research, and it was an amazing experience. Chen, N. and Reith, ME. (2000), Structure and function of the dopamine transporter. Eur. J. Pharmacol. 405, 329-339. Drugs Information Online. (2011). Bupropion. Exley, R., Moroni, M., Sasdelli, F., Houlihan, L. M., Lukas, R. J., Sher, E., Zwart, R. and Bermudez, I. (2006), Chaperone protein 14-3-3 and protein kinase A increase the relative abundance of low agonist sensitivity human α4β2 nicotinic acetylcholine receptors in Xenopus oocytes. Journal of Neurochemistry 98, 876–885. Rollema et al 2007, TRENDS in Pharmacol Sci. Rosenblum, H. (1938), Cigarette Smoking, American Journal of Preventive Medicine 49, 357. The Mental Health Channel. (2011). Bupropion. University of California, San Francisco. (2011) Neuroscience and Physiology at University of California, San Francisco. World Health Organization. (2011). World Health Organization. Xiu, X., Puskar, NL, Shanata JA, Lester HA, and Dougherty DA. (2009), Nicotine binding to brain receptors requires strong cation-pi interaction. Nature 458, 534-537. Zhang, J. and Steinbach, J. H. (2003), Cytisine binds with similar affinity to nicotinic a4b2 receptors on the cell surface and in homogenates. Brain Res. 959, 98-102.
52 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
Mean Concentration: 8.6145 mg/mL
NEURO
Articles
Genesis
Mean Concentration: 8.6145 mg/mL
Mean Concentration: 8.6145 mg/mL
Mean Concentration: 8.0757 mg/mL
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 53
NEURO Genesis
Edna Andrews, PhD eda@duke.edu
This lab is involved in two major projects: (1) fMRI studies on multilingualism, and (2) speech perception studies, specifically studies of the perception of spoken and sung phonemes across languages. The fMRI multilingualism project is on-going and includes both discrete and longitudinal studies. Students who work in this lab will be trained in computer technologies appropriate to fMRI research. Students who work on this project are not required to know the language of the study.
Roberto Cabeza, PhD cabeza@duke.edu
This lab investigates the neural mechanisms of memory in young and older adults using functional neuroimaging techniques, such as fMRI. Research topics include emotional memory, false memory, and memory-attention interactions.
Christine Drea, PhD cdrea@duke.edu
The Drea lab studies mammalian social and reproductive behavior in primates. Through a combined laboratory and field approach, they investigate reproductive and socio-endocrinology, genital and developmental morphology, and social behavior. They study how males and females differentiate to meet their respective sex roles, how they negotiate social interactions with group members, and how they solve everyday problems in the context of group living.
Cagla Eroglu, PhD c.eroglu@cellbio.duke.edu
The Eroglu lab is interested in understanding how central nervous system (CNS) synapses are formed. They believe astrocytes are an important player in the formation of such synapses. Their lab seeks to answer questions such as what are the secreted signals coming from astrocytes that regulate synapse formation, how do astrocyte-secreted factors lead to synapse formation, and what is the role of astrocyte-induced synapse formation in the development and maintenance of CNS.
Research Opportunities 54 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
NEURO Genesis
Lasana Harris, PhD bsclab@gmail.com
The Boundaries of Social Cognition lab solves philosophical questions about human behavior by combining social psychology, affective/cognitive neuroscience, and philosophy of the mind. They address questions related to morality, economic decisions involving social and monetary rewards and punishment, and the effect of competition and cooperation on decision-making. Using fMRI, EMG, and GSR, the BSC lab focuses on human perception and decision-making.
Scott Huettel, PhD scott.huettel@duke.edu
Huettel Laboratory uses a combination of behavioral, genetic, physiological, and neuroscience techniques to discover the mechanisms that underlie economic and social decision making. This broad research programâ&#x20AC;&#x201C; which includes collaborations with neuroscientists, psychologists, behavioral economists, and business and medical faculty â&#x20AC;&#x201C; falls within the emerging interdiscipline of neuroeconomics.
Richard Keefe, PhD michael.kraus@duke.edu
Keefe lab studies cognition and perception in schizophrenia and related clinical disorders. They are conducting studies in Singapore of the perceptual and cognitive deficits that may predict psychosis. Here in Durham, they are completing clinical trials on behavioral and pharmacologic treatments for cognitive impairment in schizophrenia, and are working to understand the auditory perceptual abnormalities that may underlie emotion identification difficulties.
Kevin LaBar, PhD klabar@duke.edu
Research in this laboratory focuses on understanding how emotional events modulate cognitive processes. They aim to identify brain regions that encode the emotional properties of sensory stimuli, and to show how these regions interact with neural systems supporting social cognition, executive control, and memory functions. Techniques used are psychophysiological monitoring, fMRI, and behavioral studies in healthy adults as well as psychiatric patients. This approach capitalizes on advances in the field and may lead to insights into cognitiveemotional interactions in the brain.
Hiro Matsunami, PhD matsu004@mc.duke.edu
Matsunami lab is interested in the chemical senses, the sense of smell and taste. Using molecular genetics, genomic, and cell biology approaches, they study how chemicals such as odors, tastants and pheromones interact with the receptors expressed on the surface of the sensory cells and elicit unique perception and/ or behavior. Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 55
NEURO Genesis
Jim McNamara, MD jmc@neuro.duke.edu
Dr. McNamaraâ&#x20AC;&#x2122;s laboratory seeks to elucidate the mechanisms of epileptogenesis, the process by which a normal brain becomes epileptic. Understanding the mechanisms of epileptogenesis in molecular terms may provide novel targets for pharmacologic interventions aimed at prevention of epilepsy or limiting its progression. They seek to elucidate the signaling pathways activated by TrkB, or the neurotrophin receptor, that promote limbic epileptogenesis. They are also using genetically modified mice to examine the effects of TrkB in additional models of limbic epileptogenesis.
Steve Mitroff, PhD elise.darling@duke.edu
The Mitroff lab explores malleability of visual cognition - how various experiences (e.g. videogame playing) and traits (e.g. tendencies towards ADHD and Autism) affect visual and attentional abilities. Through an array of experimental tasks and survey responses, this lab is geared towards identifying and enhancing superior searching abilities.
Michael Platt, PhD platt@neuro.duke.edu
The Platt Laboratory probes the ways in which information about the current state of the world gathered by the senses is combined with estimates of costs and benefits, uncertainty, social context, and individual-specific variables like internal state, social status, and risk tolerance to guide behavior in monkeys, adult and developing humans, mice, and other animals.
Richard Premont, PhD richard.premont@duke.edu
Premont lab studies how multiple signaling pathways initiated by neurotransmitters and neuromodulators interact and coordinate within target cells to produce integrated physiological responses. They use cell culture systems to address mechanistic questions, and knockout mice to assess global behavioral functions. As a model system, they focus on the GIT/PIX complex. They are examining mice lacking GIT1 as a model for mental retardation, and mice lacking GIT2 as a model for post-traumatic stress disorder.
Nestor Schmajuk, PhD nestor@duke.edu
Dr. Schmajuk has developed neural network models of classical conditioning, operant conditioning, animal communication, spatial learning, cognitive mapping, and prepulse inhibition. Using these neural networks he has described the effects of hippocampal, cortical, and cerebellar lesions, as well as the results of the administration of dopaminergic and cholinergic drugs, in different sensory, learning and cognitive paradigms. 56 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
NEURO Genesis
Laboratories Debra Silver, PhD debra.silver@duke.edu
Silver lab studies regulation of neural progenitors in the developing brain. They employ genetics, genomics and cell biology techniques to identify genes that influence neural stem cell division, neuron production, and brain size. Defects in these processes are associated with broad classes of neurodevelopmental diseases.
Jim Voyvodic, PhD jim.voyvodic@duke.edu
Research in this lab focuses on using fMRI to study information processing mechanisms in the nervous system. Particularly, Dr. Voyvodic is interested in how the visual system represents 3-D depth when processing visual scenes. The lab’s second main focus is applying fMRI as a clinical tool to aid in the treatment of brain tumors and epilepsy.
Anne West, MD PhD west@neuro.duke.edu
This laboratory seeks to understand at a cellular and molecular level how neuronal activity regulates the formation and maturation of synapses both during brain development and in response to plasticity-inducing stimuli, and ultimately to use genetic model systems to investigate how defects in this process lead to cognitive and behavioral dysfunction.
Leonard White, PhD len.white@duke.edu
White lab’s primary research interest is to understand how sensorimotor experience in early life influences—for better or worse—the formation and maturation of functional neural circuits. Work in his lab uses neuroanatomical techniques to relate structure to function in the developing cerebral cortex and in the brainstem.
Marty Woldorff, PhD woldorff@duke.edu or ken.roberts@duke.edu
The fundamental cognitive function of attention enables us to select and extract from moment to moment the most important information from our complex multisensory world and ongoing neural processes. This lab uses a combination of electrophysiological (EEG, ERP), functional neuroimaging (fMRI), and behavioral measures to study the cognitive neuroscience of attention. Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 57
NEURO Genesis
Kelly R. Murphy
Rory Lubner
Editor-in-Chief
Editor-in-Chief
Christine Lee
Biqi Zhang
Publishing Editor
Publishing Editor
Ha Tran
Managing Editor
Bean Sharif-Askary Managing Editor
Tiffany Chien Design Editor
Not Shown: Kyle Rand, Online Editor 58 | neurogenesisjournal.com | Fall 2012 | Vol 2 Issue 1
NEURO Genesis
Neurogenesis
Editors Undergraduate Publication Board Funding provided by The Bassett Fund Photography by Ashley Tsai
Vol 2 Issue 1 | Fall 2012 | neurogenesisjournal.com | 59
Neurogenesis is a member of Duke Undergraduate Publication Board Š Copyright 2012