PennScience Spring 2021 Issue: Health & Medicine by PennScience Journal

PENNSCIENCE JOURNAL OF UNDERGRADUATE RESEARCH Volume 19 Issue 2 Spring 2021

ARTIFICIAL INTELLIGENCE IN MEDICINE PAGE 29

COVID-19 LONG HAULERS PAGE 13

The Rheological Behavior of Firn: Experimental Observations of Dislocation Creep via Grain Boundary Sliding Page 34

HOSPITAL EMERGENCY

PENNSCIENCE

PENNSCIENCE HEALTH & MEDICINE

PENNSCIENCE PennScience is a peer-reviewed journal of undergraduate research and related content published by the Science and Technology Wing at The University of Pennsylvania and advised by a board of faculty members. PennScience presents relevant science features, interviews, and research articles from many disciplines, including the biological sciences, chemistry, physics, mathematics, geological science, and computer science. PennScience is funded by the Student Activities Council. For additional information about the journal including submission guidelines, visit www.pennscience.org or email pennscience@gmail.com.

EDITORS IN CHIEF Helen Jiang

Magnolia WANG

Faculty Advisors Dr. M. Krimo Bokreta

Dr. Jorge Santiago-Aviles

Writing Managers: Editing Managers: Design ManagerS: DAHYEON CHOI SAGAR GUPTA

Brian Song Mehek Dedhia

Amara Okafor BIANCA VAMA

Writing:

Editing:

Design:

Ashwin Sannecy Benjamin Beyer Brian Lee Emily Berkow Jonathan Tran Harish Bayana Kevin Guo Michelle Paolicelli Rebecca Nadler Sarah Pham

Abigail Gringeri Anushka Dasgupta AVNIASH SINGH Brian Song Daniel Rodriguez HOPE HAWTHORNE IBRAHIM El-Morsy Jessica Lvov linda lin nova meng sukhmani kaur

PENNSCIENCE JOURNAL | SPRING 2021

Caitlyn Prabowo Jessica Hao Lynne Kim Phuong NGO

BUSINESS MANAGERS: Glen Kahan

Business:

Ara Ponugupati Eashwar Kantemneni Elena Cruz-Adames Neha Shetty Sneha Sebastian

Technology Manager: Grace Lee

TABLE OF CONTENTS 6

COVID & INFANT LANGUAGE ACQUISITION........................Emily Berkow

MATHEMATICAL MODELS OF COVID-19......................................Brian Lee

COVID-19 LONG HAULERS....................................................Rebecca Nadler

CHRONIC PAIN MANAGEMENT & COVID-19......................Sarah Pham

HOW THE COVID-19 PANDEMIC HAS IMPACTED RESEARCH AT PENN...................................Ashwin Sannecy

AN UNCERTAIN FUTURE FOR ANIMAL TESTING.............Jonathan Tran

EXTRACELLULAR MATRIX APPLICATION IN REGENERATIVE MEDICINE................................................Michelle Paolicelli

STEM CELL THERAPEUTICS: THE SOLUTION OF TOMORROW............................................Harish Bayana

THE ROLE OF AI IN MEDICINE AND THE COVID PANDEMIC...................................................................Kevin Guo

APPLICATIONS OF QUANTUM MECHANICS IN MAGNETOENCEPHALOGRAPHY.......................................Benhamin Beyer

THE RHEOLOGICAL BEHAVIOR OF FIRN: EXPERIMENTAL OBSERVATIONS OF DISLOCATION CREEP VIA GRAIN BOUNDARY SLIDING................................................D.R. FURMAN

SPRING 2021 | PENNSCIENCE JOURNAL

Dear Readers, On behalf of the entire PennScience team, we are thrilled to present the Spring 2021 issue of the nineteenth volume of the PennScience Journal of Undergraduate Research. For many years, PennScience has explored a variety of topics across the fields of science, technology, and engineering, and in light of the global pandemic, we decided to feature the theme of Health and Medicine. In this issue, Emily Berkow examines the impact of COVID-19 on infant language acquisition while Brian Lee focuses on understanding mathematical models during COVID-19 and future pandemics. Rebecca Nadler discusses the rise of post-viral chronic disorders in COVID-19 survivors. Along a similar vein, Sarah Pham reviews the impact of COVID-19 on chronic pain treatment options. Ashwin Sannecy interviews faculty members and students to investigate how COVID-19 has impacted research at Penn. Meanwhile, Jonathan Tran addresses the controversies of animal testing in research. Michelle Paolicelli explores the applications of extracellular matrix in the field of regenerative medicine. Harish Bayana describes stem cell therapeutics as a potential solution for the future of medicine. Kevin Guo discusses the role of artificial intelligence during the COVID-19 pandemic. Finally, Benjamin Beyer analyzes the applications of quantum mechanics in magnetoencephalography a neuroimaging technique that can be used to identify brain disorders. PennScience continues to remain dedicated to showcasing the breadth and depth of undergraduate scientific research for the Penn community. In this issue, we are pleased to feature an examination of the rheological behavior of the firn, a thesis paper written by Daniel Furman under the guidance of Dr. David L. Goldsby in the Department of Earth Science. We are incredibly grateful and proud of the members of PennScience who worked diligently throughout the semester to assemble this journal - a sincere thank you to our Writing, Editing, Design, Business, and Technology Committees for their dedication and enthusiasm. We would also like to extend our gratitude to the students who submitted their novel research findings for publication, as well as the members of the Penn community who attended our events and were deeply engaged in scientific discourse, even in the remote setting. The Journal would not have been possible without the efforts of our scientifically inquisitive and passionate undergraduate members. We owe our generous funding to the Science and Technology Wing of the King’s Court College House, as well as the Student Activities Council that have made this issue possible. We are extremely thankful for our faculty mentors, Dr. Krimo Bokreta and Dr. Jorge Santiago-Aviles, for their never-ending support and guidance for the Journal. Last but not least, we would like to thank all of our readers! Your support towards PennScience is highly appreciated, and we hope you enjoy reading this issue. Sincerely, Helen Jiang (C’22) and Magnolia Wang (C’23)

PENNSCIENCE JOURNAL | SPRING 2021

Notice

LOOKING FOR A CHANCE TO

PUBLISH YOUR RESEARCH? Penn Science is accepting Research in any scientific submissions for our Fall field will be considered, 2021 issue! including but not limited to: Submit your independent study projects, snior design Biochemistry projects, reviews, and other Biological Sciences original research articles to Biotechnology share your work wiht fellow Chemistry undergraduates at Penn Computer Science and beyond. Email Engineering submissions and any Geology questions to Mathematics pennscience@gmail.com Physics Physiology

SPRING 2021 | PENNSCIENCE JOURNAL

covid-19

& infant

language

aquistion written by emily berkow

PENNSCIENCE JOURNAL | SPRING 2021

When most of a speaker’s face is covered by a

mask, we lose many nuances of communication. What are they feeling? Are they frowning? Smiling? Many adults relate to this phenomenon during COVID-era communication, but what are the effects on babies born during COVID lockdowns and young children whose communication skills have not fully developed? Some cognition experts believe that masks might be detrimental to babies’ speech and language development. In an opinion piece for Scientific American, Dr. David Lewkowicz points to several prepandemic studies that reveal the importance of early childhood development of lip-reading abilities1. This research suggests that increased attention to a speaker’s mouth is helpful for infant language acquisition and retention, and that this ability is particularly important for bilingual infants. Though comprehensive studies on this current generation of ‘“COVID babies’” have not yet been conducted, an understanding of regular language acquisition patterns suggests that masks likely hinder aspects of this process. Name 1

Name 1

Language develops primarily during face-toface interactions between infants and their caretakers2. Visible social and contextual cues provide key information for communication that speech alone is unable to fully transmit, and there is substantial evidence that a baby’s attention to information in the social scene, such as the speaker’s facial expressions, is positively correlated with language development2. Current research investigates the importance of gaze following — the ability to follow a speaker’s attentional focus towards the object of discussion —and face scanning — the ability to observe several areas of a speaker’s face —

in infant language development2. Features Features Face scanning provides infants with important information, especially around the mouth and reflects an infant’s developing ability to gather linguistically relevant information from sources other than just speech2. One study using eye-tracking and behavioral data found that greater attention to the mother’s mouth during interaction predicted higher levels of expressive language and greater rates of language development growth at the 24-month outcome of the study3. In a similar study of infant face scanning, researchers found that 3-to-4-month-old infants looked equally at the mouth and eyes of an adult speaker whereas 9-month olds spent more time looking at the eyes than the mouth. These findings suggest that younger infants rely on the mouth as a source of information, as it can reveal cues about the emotional state of the speaker as well as how to analyze the speech stream4. With fewer opportunities for infants to visualize the speaker’s mouth due to widespread mask-wearing, this method of developing expressive language may be stifled. Name 1

Name 1

The mouth is important for an infant’s language development because it aids in understanding the association between mouth shape and speech sounds2. For example, an infant might take note of a horizontally or vertically opened mouth and the difference in sound, such as [i] vs. [a] sounds, it creates2. Vocal imitation, the act of recreating other’s verbal utterances, is considered to play a large role in infant language acquisition5. Evidence reveals that infants are capable of matching sounds with appropriate orofacial movements,

SPRING 2021 | PENNSCIENCE JOURNAL

and are therefore capable of reproducing sounds more accurately5. Researchers presented infants with video recordings of a speaker producing /a/̇ like and /u/̇like sounds and counted the instances of vocal imitation, or when the infant’s response matched the speaker’s sound. They found that when speech cues were presented from an inverted face, infants imitated speakers’ sounds less because it was difficult for infants to integrate the presented visual and auditory information. Their findings demonstrate the importance of observing a speaker’s mouth for infants to display vocal imitation and ultimately develop language skills. Access to visualizing the speaker’s mouth may be even more important for bilingual infants. Infants growing up in bilingual environments are able to acquire two languages at similar rates to monolingual infants, although their task is significantly more difficult because they must distinguish between two language systems6. The mechanisms that enable this language development are well-studied. Pons et. al hypothesized that bilingual infants exploit audiovisual speech cues, such as the shapes made by a speaker’s lips corresponding with different speech sounds, more than monolingual infants who do not need to process and separate different languages6. They found that at 4 months, bilingual infants looked more at the mouth than monolingual infants. Thus,

PENNSCIENCE JOURNAL | SPRING 2021

these findings suggest that bilingual infants rely on audiovisual speech cues located in a speaker’s mouth more than monolingual infants as they develop language skills during infancy6. Other research found that the more a 24-month-old bilingual toddler looked at the mouth of a speaker, the greater chance they retained a new term7. Together, these findings suggest that a lack of visual access to a speaker’s lips may deprive bilingual infants of important audiovisual information that helps them distinguish between different languages. The exploration of infant language acquisition presented here is not inclusive of all known research in the field. There are many other factors known to affect language acquisition that do not involve the speaker’s face, such as physically gesturing to the object being referenced in speech8. Additionally, babies spend much of their time with caregivers who do not need to wear a mask around them. These primary caregivers are typically most responsible for an infant’s language growth2. Child psychologists and pediatricians advise concerned caretakers and parents to engage in developmentally appropriate play with their growing infants to promote their language and social skills — advice that is especially important during this pandemic isolation where infants have fewer socialization opportunities9. These factors of language development, among many others, are not

greatly affected by the pandemic. Thus, research does not suggest that mask-wearing will hinder all aspects of infant language development that occurs rapidly during the first months of life. Still, while the long-term linguistic developmental differences caused by the pandemic are not yet

known, our current understanding of early childhood language development demonstrates that the effects of mask-wearing are certainly worth considering.

Features

references 1) Lewkowicz, D. J. (2021, February 11). Masks can be detrimental to babies’ speech and language development. Retrieved April 04, 2021, from https://www.scientificamerican. com/article/masks-can-be-detrimental-to-babies-speech-andlanguage-development1/ 2) Tenenbaum, E. J., Sobel, D. M., Sheinkopf, S. J., Malle, B. F., & Morgan, J. L. (2015). Attention to the mouth and gaze following in infancy predict language development. Journal of Child Language, 42(6), 1173-1190. 3) Young, G. S., Merin, N., Rogers, S. J., & Ozonoff, S. (2009). Gaze behavior and affect at 6 months: predicting clinical outcomes and language development in typically developing infants and infants at risk for autism. Developmental science, 12(5), 798-814. 4) Wilcox, T., Stubbs, J. A., Wheeler, L., & Alexander, G. M. (2013). Infants’ scanning of dynamic faces during the first year. Infant Behavior and Development, 36(4), 513-516.

5) Imafuku, M., Kanakogi, Y., Butler, D., & Myowa, M. (2019). Demystifying infant vocal imitation: The roles of mouth looking and speaker’s gaze. Developmental science, 22(6), e12825. 6) Pons, F., Bosch, L., & Lewkowicz, D. J. (2015). Bilingualism modulates infants’ selective attention to the mouth of a talking face. Psychological science, 26(4), 490-498. 7) Weatherhead, D., Arredondo, M. M., Nácar Garcia, L., & Werker, J. F. (2021). The Role of Audiovisual Speech in FastMapping and Novel Word Retention in Monolingual and Bilingual 24-Month-Olds. Brain Sciences, 11(1), 114. 8) Özçalı̇kan, ̇., & Dimitrova, N. (2013). How gesture input provides a helping hand to language development. Semin Speech Lang, 34(4). 9) Yogman, M., Garner, A., Hutchinson, J., Hirsh-Pasek, K., Golinkoff, R. M., & Committee on Psychosocial Aspects of Child and Family Health. (2018). The power of play: A pediatric role in enhancing development in young children. Pediatrics, 142(3).

SPRING 2021 | PENNSCIENCE JOURNAL

MATHEMATICAL MODELS OF COVID-19 Written by: Brian Lee Designed by: Phuong Ngo Week by week throughout the past year, numerous individuals, physicians, and policy-makers have relied upon both expert opinions and a sea of statistical predictions, ranging from individual COVID death forecasts to case counts, to make decisions that affected and continue to affect our schools, workplaces, and hospitals. With many of these decisions impacting the everyday lives of millions of people, it is important to gain a basic understanding of how, and on what grounds, some of these models were built and ponder on the lessons we may carry forward from the past several months in case of another global health emergency.

Introducing the SIR Model

How do scientists and clinicians make estimates

about the size and severity of a pandemic? In epidemiology, there are several potential models for the many diseases of the world. One of the most prominent and simplest general models includes the SIR model1. The name comes from the way it partitions the general population into one of three groups: those who are Susceptible to a disease, those who are Infected with the disease, and those who have Recovered. As time passes, the number of individuals in each of these three groups changes: the number of susceptible individuals is highest at the beginning of a pandemic but lowers over time as individuals contract the disease, while the number of infected and recovered individuals is lowest when a pandemic starts but gradually (or in the case of COVID-19, frighteningly quickly) increases over time. How the sizes of these three groups evolve throughout the pandemic can be modeled by a series of differential equations1.

Given this conceptual paradigm and some real-life data regarding the epidemic in question from the local community and healthcare system, it is possible to calculate the specific equations that predict the numbers of susceptible, infected, and recovered individuals after a certain period of time (in days). As such, epidemiologists can estimate statistics such as how quickly COVID-19 10 PENNSCIENCE JOURNAL | SPRING 2021

spreads, the total number of infected individuals, and the duration of the pandemic. Multiple groups have attempted to do so using computational approaches to learn more about the COVID-19 pandemic. For example, some teams calculated the specific coefficients that satisfy the SIR model’s proposed equations using data from earlier on in the pandemic2. What’s the Best Model for COVID?

Due to a few generally faulty assumptions, however, the SIR model may not yield perfect predictions for COVID-19. First, this model assumes all individuals in the population are equally likely to come in contact with any other person, which does not accurately represent the nature of interpersonal and familial relationships. Second, the SIR model assumes there are no changes to the population size (i.e. births or deaths) due to causes other than COVID-193. Given these two oversimplifying assumptions, many scientists and statisticians argue the SIR model is not the best one to model the spread of COVID-19 across the US. There are two important levels of complexity in particular that make modelling COVID-19 in the United States especially difficult. The first layer of complexity involves the presence of — and necessity to evaluate and synthesize —several alternative models (some

Features Figure 1: Sample fitting of SIR model to early COVID-19 data. This shows one group’s results predicting the outcomes of the pandemic in the US using early COVID-19 CDC data2.

inspired by the SIR model and some not) that attempt to account for more complex diseases. As different scientists have different hypotheses about how COVID-19 is transmitted within and between communities, there are currently several competing paradigms to predict the future of COVID-19. Many of the more prominent models are derived from the SIR model but include extra steps to account for knowledge gained from epidemiological research, including the SEIR and SIRD models1, 4. In an effort to incorporate sociological knowledge into transmission predictions, some researchers have turned to network transmission models that account for ‘connectedness’ of individuals within a community5. This model’s most notable assumptions are based on the intuition that individuals who share a strong relationship (such as a group of Penn students in the same dorm ‘pod’ or a group of close coworkers) are more likely to either be sick simultaneously and/or transmit COVID-19 to each

other. Lastly, in addition to these approaches making powerful biological assumptions, some researchers have resorted to more traditional computational (machine-learning based) models that require large amounts of raw patient and outcome data. A prime example of this includes the Institute for Health Metrics and Evaluation’s COVID-19 pandemic model, which fits curves to regularly-collected case count and death data6. The second layer of complexity involves factoring for the heterogeneous nature of the US. Unfortunately, models created using data from one portion of the country (i.e. a small town in Idaho) may fail to predict COVID-19 cases and deaths in another portion of the country (including our native West Philadelphia), necessitating researchers from across the country to both share their COVID-19 data and potentially create their own models. More worryingly, researchers may not fully grasp the SPRING 2021 | PENNSCIENCE JOURNAL

References 1. Abou-Ismail A. (2020). Compartmental Models of the COVID-19 Pandemic for Physicians and Physician-Scientists. SN comprehensive clinical medicine, 1–7. Advance online publication. https://doi. org/10.1007/s42399-020-00330-z 2. Cooper, I., Mondal, A., & Antonopoulos, C. G. (2020). A SIR model assumption for the spread of COVID-19 in different communities. Chaos, solitons, and fractals, 139, 110057. https://doi.org/10.1016/j. chaos.2020.110057 3. Tolles, J., & Luong, T. (2020, June 23). Modeling Epidemics With Compartmental Models. JAMA Guide to Statistics and Methods, 323(24), 2515-2516. 10.1001/jama.2020.8420 4. Calafiore, G. C., Novara, C., & Possieri, C. (2020). A time-varying SIRD model for the COVID-19 contagion in Italy. Annual reviews in control, 50, 361–372. https://doi.org/10.1016/j.arcontrol.2020.10.005 5. Zlojutro, A., Rey, D., & Gardner, L. (2019). A decision-support framework to optimize border control for global outbreak mitigation. Scientific reports, 9(1), 2216. https://doi.org/10.1038/s41598-019-38665-w 6. IHME COVID-19 health service utilization forecasting team & Murray, C. J. (2020, March). Forecasting COVID-19 impact on hospital bed-days, ICU-days, ventilator-days and deaths by US state in the next 4 months. medRxiv. 10.1101/2020.03.27.20043752 7. The Centers for Disease Control. (2021, February 12). Health Equity Considerations and Racial and Ethnic Minority Groups. COVID-19. https://www.cdc. gov/coronavirus/2019-ncov/community/health-equity/ race-ethnicity.html 8. The Centers for Disease Control. (2020, August 7). COVID-19 Mathematical Modeling. COVID-19. https://www.cdc.gov/coronavirus/2019-ncov/covid-data/ mathematical-modeling.html 9. Ray, E. L., Wattanachit, N., Niemi, J., Kanji, A. H., House, K., Cramer, E. Y., Bracher, J., Zheng, A., Yamana, T. K., Xiong, X., Woody, S., Wang, Y., Wang., L, Walraven, R. L., Tomar, V., Sherratt, K., Sheldon,

12 PENNSCIENCE JOURNAL | SPRING 2021

impact COVID-19 has had on their communities due to several deep-seated inequities in healthcare. Unfortunately, individuals of color, of lower socioeconomic status, and of rural areas are less likely to have access to a primary care physician and have paid sick leave, potentially artificially deflating COVID-19 case and death count data and predictions in certain communities7. The CDC Model(s): Integrating Results to Learn About COVID Nationwide Given these two important levels of complexity, there are currently several COVID-19 predictions from a wide variety of research groups, hospital systems, schools of public health, and individual statisticians across the nation using data from potentially separate and extremely different patient populations. Alone, if tasked with modelling COVID-19 for all of the nation’s different cities and neighborhoods, this plethora of models, predictions, and statistics can provide a very convoluted, incoherent (but not necessarily incorrect) picture. However, when synthesized in a meaningful way, these multiple analyses may offer a more coherent and complete picture of the nation’s previous, current, and future struggles with COVID-19. It is useful to bring these individual forecasts together to both compare their predictions and understand how much certainty there is in the next few days, weeks, or months of the pandemic. The CDC did precisely this: through combining the analyses of multiple different groups, they were able to create a larger picture of the COVID-19 pandemic in the US and attain more powerful predictions of the nature of the pandemic8. Fortunately, researchers found such a summative approach predicts COVID-19 progression and cases with extremely high accuracy; over 90% of all predictions made predicted the actual number of deaths, cases, etc9. It is especially interesting and inspiring to find that a majority of the models can accurately predict the progression of this disease in the country when combined, even if many of these predictors fail to accurately predict COVID-19 statistics on their own. One final question to ponder is how we can use the experiences and lessons of today’s pandemic to both improve our understanding of biological predictive models and prepare for the next pandemic. This pandemic has made increasingly clear how mathematical models can be instrumental in both learning about the virus and keeping those we care about safe. Perhaps our collective experiences living and learning through such a perilous year has prepared us to recognize, train, and interpret these models in the pandemics of tomorrow10.

Features

The Rise of Post-Viral Chronic Neurological Disorders in COVID-19 Survivors

Written by Rebecca Nadler Designed by Jessica Hao

n January 7, 2020, the World Health Organization (WHO) identified an outbreak of pneumonia cases in Wuhan, China, as a new coronavirus. By January 30, the WHO declared the outbreak a global public health emergency with over 9,000 reported cases of COVID-191. COVID-19 begins as a mild to severe respiratory infection, initially causing flu-like symptoms. However, COVID-19 has a higher rate of transmission, hospitalization, and mortality than influenza2. Follow-up studies of COVID-19 survivors reveal a rise in those suffering from post-viral chronic illnesses, even after experiencing only mild symptoms of COVID-19. Such patients, considered to be COVID-19 “long-haulers,’’ are developing poorly understood neurological disorders as a viral response to their infections. A study of such conditions, including postural orthostatic tachycardia syndrome (POTS) and myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), may improve the current understanding of the pathophysiology of COVID-19. Although COVID-19 is often considered to be a respiratory disease, the SARS-CoV2 virus is able to affect many organ systems. The virus primarily targets the ACE2 receptors, which are present on cells throughout the body3. Additionally, the virus has the potential to harm the immune system, leading to an immunocompromised state as well as hyperinflammation. Neurological manifestations of COVID-19 have also been reported, and likely result from a combination of

neuroinvasive properties of the SARS-CoV2 virus and downstream multi-organ dysfunction4. These symptoms do not always dissipate in patients that experience a viral course. There is no clear definition for COVID-19 long haulers as of yet, but they are typically regarded as patients that experience persistent symptoms for several weeks to months after infection. In contrast to popular belief, this population is not restricted to certain demographics or case severities. Dr. Wesley Self, MD, MPH, studied 292 patients that recovered from COVID-19 who did not require hospitalization, and over one-third of his patients

“A quarter of young adults surveyed, between 18 to 34 years old, had not yet regained their health” had not returned to their usual state of health three weeks after testing positive for the virus. His study also revealed that a quarter of young adults surveyed, between 18 to 34 years old, had not yet regained their health6. A study of previously hospitalized COVID-19 patients reported fatigue (55%), shortness of breath (42%), loss of memory (34%), and concentration and sleep disorders (28% and 30.8%) after 110 days5. Given the diversity in patient symptomatology, it is critical that physicians and researchers consider the neurological implications of COVID-19, as it appears to be the most overlooked component of the disease. Cases of ischaemic stroke, GuillainBarre Syndrome, and polyneuropathy have been SPRING 2021 | PENNSCIENCE JOURNAL

ers

reported following SARS-CoV2 infection4, but a far larger proportion of COVID-19 survivors are experiencing POTS and ME/CFS. POTS is a form of dysautonomia, which is a group of disorders affecting the autonomic nervous system. It is characterized by orthostatic intolerance, or the development of cardiologic symptoms that arise when standing up from a reclining position, among other multisystemic symptoms such as fatigue, gastrointestinal issues, and chronic pain. Patients are diagnosed using a tilt-table test, with the main diagnostic criteria being a sustained heart rate increment of at least 30 beats/min within 10 minutes of standing or head-up tilt7. Dr. Mitchell Miglis, a Stanford University neurologist who specializes in autonomic nervous system disorders such as POTS, recently coauthored a case report about the emergence of classic POTS symptoms following COVID-19 infection. He and his team described a previously healthy, 26-year-old emergency department nurse who developed fatigue and tachycardia after contracting COVID-19. While the acute phase of her COVID-19 symptoms lasted for only three weeks, her tachycardia and other symptoms of autonomic impairment persisted and intensified over 5.5 months, as of when the case report was published.6

14 PENNSCIENCE JOURNAL | SPRING 2021

There are four different pathophysiological mechanisms that have been hypothesized to explain the association between COVID-19 and the rise of POTS-like symptoms and diagnoses. First, fever, anorexia, nausea, excessive nocturnal sweating, and prolonged bed rest may cause hypovolemia, or the decrease in overall blood volume, incidentally increasing cardiac outflow. Furthermore, the SARS-CoV-2 virus may be able to infect and destroy neurons that modulate heart function, thereby increasing cardiac outflow analogous to neuropathic POTS. Similarly, the SARS-CoV-2 virus could invade the brainstem and affect the part of the brain that regulates cardiovascular function, altering cardiac output. This would also explain the brain fog that some patients experience. Lastly, an autoimmunity-driven response would support the literature revealing the presence of autoimmune markers and autoantibodies in patients with POTS7. The symptoms of POTS following COVID-19 significantly overlap with

Features

those of ME/CFS. In fact, Dr. Anthony Fauci, director of the U.S. National Institute of Allergy and Infectious Diseases and the chief medical advisor to the president, recently noted that some long haulers’ symptoms like brain fog and fatigue are “highly suggestive” of ME/CFS6. ME/CFS is a debilitating chronic disease resulting in profound mental and physical fatigue, sleep disturbance, chronic pain, and intolerance to mental and physical exertion8. Interestingly, three out of four people diagnosed with ME/CFS report that it began with an infection, most commonly infectious mononucleosis or Epstein-Barr virus.6 With a disease prevalence currently reported to be 0.3 0.8%, ME/CFS is not limited to an older population or to those that experienced severe viral infection. Some experts even propose that this number is an underestimate of cases. A study of chronic COVID-19 syndrome at the Charité Fatigue Center in Berlin confirmed initial concerns that COVID-19 leads to persistent fatigue syndromes in a subset of younger individuals following only mild to moderate infection8. This has significant long-term implications for newly diagnosed patients, as the study reported that two thirds of the patients required reduced hours or were entirely unable to work 6 months after COVID-19 infection. The tragedies of COVID-19 and diminished quality of life in many survivors necessitates long-term follow up of patients and further research into their neurological state post-COVID-19 infection. COVID-19 presents a unique opportunity to study conditions such as POTS and ME/CFS at an early stage. Observational and genome-wide association studies may allow physicians and researchers to elucidate the pathway and individual susceptibility for POTS and ME/CFS. These findings could ultimately lead to a better understanding of the

REFERENCES 1.

Muccari, R., Chow, D., & Murphy, J. (2021, January 1). Coronavirus timeline: Tracking the critical moments of Covid-19. NBCNews.com. https://www.nbcnews.com/ health/health-news/coronavirus-timeline-tracking-criticalmoments-covid-19-n1154341.

Centers for Disease Control and Prevention. (2020). Symptoms of Coronavirus. Centers for Disease Control and Prevention. https://www.cdc.gov/coronavirus/2019ncov/symptoms-testing/symptoms.html.

Marshall, M. (2020). The lasting misery of coronavirus long-haulers. Nature, 585(7825), 339–341.https://doi. org/10.1038/d41586-020-02598-6

Whittaker, A., Anson, M., & Harky, A. (2020). Neurological Manifestations of COVIḊ19: A systematic review and current update. Acta Neurologica Scandinavica, 142(1), 14–22. https://doi. org/10.1111/ane.13266

Garrigues, E., Janvier, P., Kherabi, Y., Le Bot, A., Hamon, A., Gouze, H., … Nguyen, Y. (2020). Post-discharge persistent symptoms and health-related quality of life after hospitalization for COVID-19. Journal of Infection, 81(6). https://doi.org/10.1016/j. jinf.2020.08.029 6.

Rubin, R.

(2020). As Their Numbers Grow, COVID-19 “Long Haulers” Stump Experts. JAMA, 324(14), 1381. https:// doi.org/10.1001/ jama.2020.17709 7. Goldstein, D. S. (2020). The possible association between COVID-19 and postural tachycardia syndrome. Heart Rhythm. https://doi.org/10.1016/j.hrthm.2020.12.007

Written By: Sarah Pham Designed By: Caitlyn Prabowo

n i a P c i n o r t h n e C m e g a n Mand COVID-19 a

mplifying current hardships within the preexisting opioid epidemic, the COVID-19 pandemic hinders the delivery of quality chronic pain patient-care and further emphasizes the need for changes in surgery protocols to discourage opioid reliance. During this time, chronic pain patients face the double burden of managing the mental, emotional, and physical tolls that come with self-quarantine and the highly addictive use of opioids. Common associations with chronic pain include complex biopsychosocial interactions and psychological disorders. As the leading cause of disability and high economic and social burden, chronic pain compromises patients’ quality of lives through their sleep, work, and interactions with others¹. However, typical administered pain relieving medicines involving opioid analgesics can often lead to highly drug-addictive lifestyles and severe symptoms. With the crossover of the opioid epidemic and the COVID-19 pandemic, chronic pain patients risk disconnected patient-care experiences and low accessibility to biopsychosocial support.

Opioids 101 Opioids refer to a class of various compounds that bind to specific opioid receptors in the central nervous system, resulting in analgesic and narcotic effects. Opioids can act on various parts of the brain and the nervous system including the limbic system, brainstem, and spinal cord.² Specifically, opioids in 16 PENNSCIENCE JOURNAL | SPRING 2021

of pleasure, relaxation, and contentment, opioids in the brainstem influence autonomous controls and can slow breathing, stop coughing, and reduce pain sensory, and opioids in the spinal cord influence sensory reception from the body and can decrease sensations of pain.² The effect of opioids depends on the quantity, frequency, and manner in which they are taken.

Current Chronic Pain Treatments With Opioids Opioid analgesic drugs have dominated the market as the most commonly used painrelieving medications during surgical procedures. Generally, this includes ward admission, anesthesia, surgery, and recovery. While effective in reducing somatic pain in skin, muscle and soft tissues, studies have shown that opioids have little to no effect in eliminating neuropathic pain caused by damaged nerves.³ Somatic pain occurs when nociceptors (pain receptors) are activated through stimulus from temperature, vibration, swelling, or force. Effects of somatic pain generally include cramping, aching, or gnawing sensation.̇ On the other hand, neuropathic pain affects nerves which are essential to transferring information between the brain and spinal cord and other sensorimotor systems. Symptoms of neuropathic pain can include burning sensations, pin and needle

At the same time, the origins of many opioid addicts come from their experience as hospital patients where they were prescribed the drugs as analgesic agents3. What is originally used to accommodate for the surgi cal pain can easily put patients at risk of developing an opioid addiction and potentially overdosing. When administered to patients, the related side effects of opioids warrant caution by use of healthcare professionals. Namely, breathing disorders (i.e. respiratory depression) serve as the most notable opioid side effect due to its association with operations with high-incidence post-op respiratory failure and patients with obesity, sleep apnea, or chronic obstructive pulmonary disease.³ Studies also found that use of opioids can be associated with post-surgery delirium, compromised cell-mediated immunity, and increased tumor recurrence rate after surgery. In addition, opioid use can lead to immunosuppressive effects which alter the innate and acquired immune responses.³

Impact of COVID-19 on Chronic Pain Management As a result of redistributing healthcare resources to accommodate COVID-19 patients, many chronic pain patients were left in a state of distress as pain services shut down and were deemed “nonurgent.” A lack of resources for pain management combined with self-quarantine recommendations can quickly cause a downwards spiral in patients’ psychological health. The risk factors for pain morbidity and mortality have

Features

also been amplified as a result of pain patient-care disconnect. In the midst of COVID-19, fear of viral transmission deters chronic pain patients from making hospital visits. Those deciding to leave their acute pain untreated may develop chronic pain. Then, untreated chronic pain may induce immunosuppression and increase the affected person’s risk to COVID-19.¹ At the same time, the imposition of social isolation can lead to a promotion of disengaged coping strategies, thus worsening mental disorders such as depression coupled with suicidal tendencies. On the other hand, chronic pain patients with an implantable device require interventional pain procedures in order to keep up their treatment. To combat these issues, a panel of expert healthcare professionals developed recommendations for the management of pain patients based on medical factors and circumstances to mitigate the risk of viral transmission for healthcare providers and patients, conserve resources, and grant greater accessibility to pain management services.¹ Whereas patients under urgent conditions will be given faceto-face care, those with non-emergent or long term disease will be given a dedicated remote patient-care experience. Currently, telemedicine and telehealth are two communication systems which allow for athome pain treatment and the continuity of chronic pain patient-care after discharge. Telemedicine functions is a two-way communication system with visual and audio accessibility while telehealth is a modern term referring to any use of technology to assist with virtual patient care. Not only do these systems allow professionals to monitor the mental and emotional states of pain patients during isolation, but they are also used to monitor the state of patients with diabetes, chronic obstructive disease, or cardiac disease while acting as a convenient alert system for patients who utilize medical devices.¹

SPRING 2021 | PENNSCIENCE JOURNAL

Opioid Alternatives Resulting from the concern of risks from opioid use, non-opioid analgesic drugs have been developed for use in patient hospitalization. A variety of these alternatives come in the form of intravenous and oral intake such as IV Dexmedetomidine, IV Acetaminophen, IV Keterolac, IV Lidocaine, IV Magnesium, Oral Gabapentin, and Oral Pregabalin.³ Specifically, alternatives such as dezocine and buprenorphine are classified as mixed agonist/ antagonist drugs acting on opioid receptors and offer significant advantages by lowering patients’ analgesic ceiling and showing less addictiveness.³ Agonist drugs increase the opioid receptor activity while antagonist drugs inhibit it. In opioid-addicted patients, dezocine

diminishes morphine-induced dependence. Although studies show that opioid alternatives have lower addiction potential and a better side effect profile, they are currently underutilized and overlooked by anesthesia providers and surgeons due to their unconventional use in standard protocol.³ Heightened by the overlap between the opioid epidemic and the COVID-19 pandemic, the current protocol of analgesic therapy must be re-evaluated and the incorporation of alternative analgesics into perioperative pain management should be further discoursed in order to reduce the numbers of opioid-addicted patient outcomes and overdoses.

References

1. Puntillo, F., Giglio, M., Brienza, N., Viswanath, O., Urits, I., Kaye, A. D., Pergolizzi, J., Paladini, A., & Varrassi, G. (2020). Impact of COVID-19 pandemic on chronic pain management: Looking for the best way to deliver care. Best practice & research. Clinical anaesthesiology, 34(3), 529–537. doi.org/10.1016/j.bpa.2020.07.001 2. Ghelardini, C., Di Cesare Mannelli, L., & Bianchi, E. (2015). The pharmacological basis of opioids. Clinical cases in mineral and bone metabolism: the official journal of the Italian Society of Osteoporosis, Mineral Metabolism, and Skeletal Diseases, 12(3), 219–221. PMID: 26811699 3. Bohringer, C., Astorga, C., & Liu, H. (2020). The Benefits of Opioid Free Anesthesia and the Precautions Necessary When Employing It. Translational perioperative and pain medicine, 7(1), 152–157. PMID: 31712783 4. Rosenquist R.W., & Vrooman B.M. (2013). Chapter 47. chronic pain management. Butterworth J.F., IV, & Mackey D.C., & Wasnick J.D. (Eds.). Morgan & Mikhail’s Clinical Anesthesiology, 5e. McGraw-Hill. https://access medicine.mhmedical.com/ content.aspx?bookid=564&secti onid=42800580

18 PENNSCIENCE JOURNAL | SPRING 2021

The University of Pennsylvania has been one of the top research institutions in the United States for the past century. The university’s mission to grasp new information in fields such as medicine, technology, and other STEM fields has helped undergraduate, graduate and faculty members pursue research. To further encourage research in interdisciplinary fields, President Amy Gutmann recently signed the Penn 2022 Compact1. The goal of this initiative is focused on improving inclusion with the Philadelphia community, stronger innovation with faculty and students, and making a positive impact on global citizens. As of 2019, there are over 189 research centers/institutions related to Penn that are allocated over 1 billion dollars for researchers to pursue their interests2. Alongside the Penn 2022 compact, another measure to spark greater research involvement was the Penn Integrates Knowledge (PIK). Its purpose was to recruit more scholars with strong experience in interdisciplinary fields to expand Penn’s research and diversify labs3. While Penn has taken many steps to broaden its research development, the COVID-19 pandemic has significantly impacted many of the research centers. On March 17th 2020, Penn officially released a statement shutting many “nonessential labs” that did not focus on the coronavirus or immediate research4. For nearly four months, labs were completely closed until the campus constructed new COVID-19 protocols on how research could be conducted safely. This included an emphasis on social distancing, strict guidelines on what researchers could do outside the lab, and other measures to prevent the virus from spreading4. As the vaccine and better awareness have helped reduce the number of positive cases, Penn has taken several

How the COVID-19 Pandemic has Impacted Penn Research Written by Ashwin Sannecy Designed by Amara Okafor

Active Members in Penn Talk About How the Campus shutfown has affected their team’s research the past year steps to reach a full recovery for faculty and student research, such as restarting the PURM (Penn Undergraduate Research Mentoring) program this 2021 summer in-person. However, current research has still not reached complete normality and it’s important to see how different Penn members have been affected this past year as well. For the fall and spring semesters of the 2020-2021 academic year, faculty members have been instructed to stay at home and communicate with their research labs virtually. Furthermore, Penn had set a time restriction on staying on-campus for research-related activities. To determine how specific labs are impacted, I conducted an interview with Dr. Feng Gai, a professor in the Penn Chemistry Department who has been actively involved in research for several years. His current research investigates how proteins fold from their denatured state to their native functional conformations through folding mechanisms. He

Features

uses spectroscopy to study the early folding events and folding intermediates for specific proteins to understand their pathways. Delving deeper into his current research projects, Dr. Gai’s lab is studying how quickly excited protein folding will cause different folding pathways. Furthermore, this will help understand how different intermediate states are formed and how the structure of the protein varies through each state. His team’s recent works involve the study of the helix-coil transition, helix-helix interaction, and ß-sheet formation5. Dr. Gai’s other research interests include studying how peptides interact with membranes for protein aggregation (disordered proteins clumping to cause various diseases)6. While talking to him about this topic, Dr.Gai mentioned how it is a novel field in biology and is being studied for factors that have induced this process (hydrophobic interactions, mutations, etc.). After I learned what Dr. Gai’s research was on, it was important to see how the COVID-19 pandemic has affected his lab’s success. One thing he stressed was that the impact on the lab depended on what the lab focused on. A lab that involved more experimental research, such as instrumentation, benches or chemicals, was much more impacted than labs that had more computational analysis. Furthermore, the size of the lab was also a large factor in how the lab could operate. Smaller groups, such as Dr. Gai’s lab, could socially distance themselves while still wearing PPE and doing their research (in accordance with Penn’s COVID-19 protocols). Whereas larger research labs that

SPRING 2021 | PENNSCIENCE JOURNAL

included over 10-15 individuals had to designate specific times for researchers to work in their research area, making it harder to communicate. For instance, Dr Gai’s lab communicated more on Zoom than in person. As Penn’s campus opens for the fall semester, Dr. Gai is hopeful that his and others’ labs can return back to a normal state and continue productive research. I also interviewed two graduate students at Penn to see how their team’s research has been impacted. The first student was Phoebe Askelson, a physical chemist working with Dr. Jessica M. Anna in the Penn Chemistry Department. Their group’s research is focused on laser spectroscopy and analyzing how ultrafast nonlinear spectroscopy can drive synthetic light harvesting systems and photocatalysts7. This includes photosystem 1 and chlorophyll that are used in photosynthesis and also inorganic materials containing metals, dipyrrin complexes and laser dyes7. By using spectroscopy, the lab is able to understand the photosynthetic processes and apply it to real-life applications to enhance reactivity. When talking to her about how the pandemic has affected her lab, Phoebe mentioned how many of her group’s projects had to be pushed back due to the shutdown. The lab maintained their research goals, as it continuously expanded with unique molecules to research,

but needed more time to finish. Additionally, the social-distancing

Dr, Gai’s lab applies laser induced temperture-jump and photochemical triggering to rapidly initiate folding/unfolding reactions, which they can monitor by probing various infrared absorption reporters that can measure either global or local dynamics.

protocols set by Penn made it hard for newer undergraduates to learn how to work with the machinery. While research was slowed down from March to July, the team’s lab unfortunately was flooded. Since researchers didn’t have access to their lab during the shutdown, there were significant damages to their equipment and laboratory, and research goals that were pushed back. To end the interview, Phoebe voiced how she is excited to research new applications for different light-harvesting mechanisms and stressed the importance of maintaining a healthy work-life balance.

20 PENNSCIENCE JOURNAL | SPRING 2021

The other researcher I interviewed was Jeffrey McNeill, a graduate student pursuing chemistry who was doing research in the Mallouk Group. This lab’s research was based on material science and studied how ultrasound can be used to transport cells in the body. Using specific frequency, Jeffrey’s team was able to create a microbubble that could “pull” and “push” individual cells without causing apoptosis (induced cell death)8. This discovery has large applications in medical treatment with cancer and other diseases where the technology can be used to separate harmful cells from the body’s natural cells9. Similar to Phoebe, the Mallouk Group was also significantly impacted by the pandemic this past year. Their group consists of approximately 24 members, which made it hard to socially distance. This meant that

The Jessica M Anna lab’s research is focused on laser spectroscopy and analyzing how ultrafast nonlinear epctroscopy can drive synthetic light harvesting systems and photocatalysts

Features

Pictured Above are the Interviewees: Dr. Feng Gai (Left), Phoebe Askelson (Center), Jeff McNeill (Right)

the team had to split into different time slots, making it hard to share equipment and the research facilities. Even though the social distancing protocols have adjusted over the past few months, the group has had to change their research methods. Normally, the Mallouk Group would work with other

collaborators, such as universities or outside facilities, to use the cells that they incubated to test on the nanorobots (microbubbles) that they designed, but this has stopped due to restrictions. However, Jeffrey stressed that their group has used the pandemic as an opportunity to innovate and find new pathways to research this topic. Overall, the impact of the pandemic on research at Penn is evident. All members of the campus, including graduate students and faculty members, have had to adjust how their group conducted research in order to stay aligned with safety protocols. As the pandemic has slowly decreased over the past few months, it is expected that more labs will continue to open, and Penn research can reach normality. For now, Penn members continue to be passionate about their research and are able to adapt to meet safety protocols.

Using speciﬁc frequency, Jeff’s team was able to create a microbubble that could “pull” and “push” individual cells without causing apoptosis (induced cell death

References

1. University of Pennsylvania. (n.d.). Penn Compact 2022. Penn Office of the President. https://president.upenn.edu/ penn-compact. 2. Research & innovation. (n.d.). https://www.upenn.edu/research-andinnovation. 3. University of Pennsylvania. (n.d.). Professors Integrating Knowledge. PIK. https://pikprofessors.upenn.edu/aboutpik. 4. Brockmeier, E. K. (2020, March 16). On-campus research to be limited due to COVID-19. Penn Today. https:// penntoday.upenn.edu/news/campusresearch-be-limited-due-covid-19. 5. University of Pennsylvania. (n.d.). Feng Gai. University of Pennsylvania School of Arts & Sciences Department of Germanic Languages and Literatures. https://www. chem.upenn.edu/profile/feng-gai. 6. University of Pennsylvania. (n.d.). GaiLab. Gai Group - Research. http:// gaigroup.chem.upenn.edu/research.html. 7. University of Pennsylvania. (n.d.). The Anna Group. Google Sites. https://sites. google.com/sas.upenn.edu/the-annagroup/home. 8. Mallouk-Lab. (2019, April 1). web.sas. upenn.edu. https://web.sas.upenn.edu/ mallouk-lab/. 9. PennSAS. (2021). 2021 Penn Grad Talks: Jeffrey McNeill. Vimeo. https:// vimeo.com/showcase/8153831/ video/514102223.

SPRING 2021 | PENNSCIENCE JOURNAL

An Uncertain Future for Animal Testing As of late March, over 140 million doses of the

coronavirus vaccine have been administered in the United States1. The success of new COVID-19 vaccines has come as a great relief after months of pandemic lockdown. However, the development of these vaccines did not occur without renewed controversy over animal testing. Clinical studies conducted by major pharmaceutical companies, like Pfizer and Moderna, included procedures in which researchers infected animals with the coronavirus. In order to draw conclusions about our current practice of animal testing, it is important to understand the benefits and drawbacks of this practice. Animal testing has allowed humans to make great medical advances such as the diphtheria and polio vaccines, antibiotics, and modern anaesthetics2. These 20th century developments have saved countless lives to date. This is why many argue that animal testing is a necessary procedure for developing such cures and treatments. One promising example of animal testing used for current research is in the fight against Hutchinson–Gilford progeria syndrome. Progeria is a fatal genetic condition that rapidly accelerates aging in children3. However, scientists have recently found a way to use DNA editing techniques to prolong the lives of mice with the same genetic variation associated with progeria4. Researchers injected mice with a gene editor, which was astoundingly able to restore the normal DNA sequence in organ cells5. This discovery was made possible by animal testing and now provides hope for those faced with this rare condition. In breakthroughs like these, animal testing provides the scientific community with a method of determining whether medicine is safe for human use. Dr. Peter Hotez, Dean of the National School of 22 PENNSCIENCE JOURNAL | SPRING 2021

Written by: Jonathan Tran Designed by: Phuong Ngo

Tropical Medicine at Baylor College, claims that this was especially important in COVID-19 vaccine development. Despite public pressure to accelerate the development process, Dr. Hotez was adamant that animal testing was conducted. He cites the reason being to avoid vaccine enhancement, a phenomenon in which a vaccine can worsen a disease’s effect6. Past failed attempts to develop HIV vaccines have exhibited vaccine enhancement in medical trials7. However, in modern medicine, vaccine enhancement has largely been prevented by conducting extensive testing. Being able to avoid the catastrophic consequences of supporting a disease instead of preventing it is a major reason why researchers continue to conduct animal experiments. Animal testing is routinely conducted to develop not only medicine, but also commercial products like cosmetics and pesticides. Unfortunately, many of these procedures cause test subjects suffering. For the 2018 fiscal year, the US Department of Agriculture reports that upwards of 30,000 animals were used in research that involved pain8. To determine the effectiveness of products like painkillers, it is not uncommon for animals to be purposely wounded in order to study healing processes and pain remedies9. What makes matters worse is that animal testing is not always a great measure of how products can affect humans. This can have devastating effects. Thalidomide, a drug that was tested on animals before distribution, caused approximately 10,000 babies to be born with serious deformities10. Researchers had conducted tests in which they attempted to give lethal doses of thalidomide to animals, but having found it was nearly impossible to do so, they falsely concluded it would be harmless to humans. Physiological and behavioral differences in how humans and animals react to their surroundings can greatly influence experimental results. This variance is

exacerbated by the laboratory conditions that animals are kept in. Because animals are not always a great indicator of how humans will react to different products, animal rights activists argue that the suffering endured by test subjects is pointless and inhumane. In fact, many major figures in the health industry agree that it is time to move on from animal testing. According to Dr. Elias Zerhouni, former U.S. National Institutes of Health director, it is important that scientists begin to “refocus and adapt new methodologies for use in humans to understand disease11.” Possible alternatives to the current practice of animal testing include in vitro testing and computer modeling. In vitro testing is when trials are run on organism samples, such a blood or tissue, that have been removed from the body12. It has been found that modern in vitro tests can determine the biological effects of test compounds as well as, and at times, better than, studies on complete animals11. Although animal testing as a whole will be difficult to phase out, researchers are also now turning their attention toward developing models of organisms that are less costly and more efficient than our current mammal test subjects13. Some species of interest include fruit flies and zebrafish. Another increasingly common computer-based technique is quantitative structure-activity relationships (QSARs). QSARs are able to make strong approximations about the probability that a substance is dangerous based on existing knowledge of similar substances and human biology14. These are all examples of promising developments that may pave the way for researchers to finally phase out animal testing. As we continue to evaluate possible futures in animal testing, it is important to also acknowledge the economics of such decisions. The pharmaceutical industry, like all industries, has monetary motivations to make certain choices. Past surveys of American consumers have found that 67% of participants believe that animals should not be used as test subjects for products like cosmetics and soap. Moreover, 60% say that they are more inclined to purchase products that have not been tested on animals15. Public opinion highlights a potential

economic incentive for companies to explore alternatives to animal testing.

Features Features

Future innovations in the scientific community may make this decision easier to make, but for now, we are left to balance the scientific, moral, and economic implications of animal testing and its alternatives. The coronavirus has been a devastating example of the importance of medicine in the modern world. As the world population continues to climb, outbreaks of disease are becoming increasingly dangerous and widespread. Being able to make educated decisions about how we choose to test life-saving medicine will continue to be a conversation of paramount importance.

References

1. Holder, J. (2021, January 29). Tracking coronavirus vaccinations around the world. Retrieved March 29, 2021, from https://www.nytimes.com/interactive/2021/ world/covid-vaccinations-tracker.html?auth=login-email&login=email 2. Fisher, P. (2013, September 15). Why we should accept animal testing. Retrieved March 22, 2021, from https://www.huffingtonpost.co.uk/professorelizabeth-fisher/why-we-should-accept-anim_b_3608923.html 3. Progeria Research Foundation. (2019, May 09). About progeria. Retrieved March 22, 2021, from https://www.progeriaresearch.org/about-progeria/ 4. Koblan, L., Erdos, M., Wilson, C., Cabral, W., Levy, J., Xiong, Z., . . . Liu, D. (2021, January 06). In vivo base editing rescues Hutchinson–Gilford progeria syndrome in mice. Retrieved March 22, 2021, from https://www.nature.com/articles/s41586-02003086-7 5. Dna-editing method shows promise to treat mouse model of progeria. (2021, January 06). Retrieved March 22, 2021, from https://www.sciencedaily.com/ releases/2021/01/210106133046.htm 6. Steenhuysen, J. (2020, March 11). As pressure for coronavirus vaccine mounts, scientists debate risks of accelerated testing. Retrieved March 22, 2021, from https:// www.reuters.com/article/uk-health-coronavirus-vaccines-insight-idUKKBN20Y1I1 7. Huisman W, Martina BE, Rimmelzwaan GF, Gruters RA, Osterhaus AD. (2008, November 18). Induced enhancement of viral infections. Retrieved March 22, 2021, from https://pubmed.ncbi.nlm.nih.gov/19022319/ 8. Animal and Plant Health Inspection Service. (2020, January 7). Annual Report Animal Usage by Fiscal Year (Rep.). Retrieved https://www.aphis.usda.gov/ animal_welfare/annual-reports/Annual-Report-Summaries-State-Pain-FY18.pdf 9. Animal testing. (2021, January 05). Retrieved March 22, 2021, from https:// www.hsi.org/issues/animal-testing/ 10. Thalidomide. (2019, December 11). Retrieved March 22, 2021, from https://www.sciencemuseum.org.uk/objects-and-stories/medicine/ thalidomide 11. UC San Diego. (n.d.). The humanoid center of Research Excellence. Retrieved March 22, 2021, from https:// medschool.ucsd.edu/research/inm/humanoid/Pages/default. aspx 12. Center for Devices and Radiological Health. (n.d.). In vitro diagnostics. Retrieved March 22, 2021, from https://www.fda.gov/medical-devices/products-andmedical-procedures/vitro-diagnostics 13. Encyclopædia Britannica. (n.d.). Retrieved March 22, 2021, from https://www.britannica.com/explore/savingearth/scientificalternatives-to-animal-testing-a-progress-report 14. PETA. (2021, March 15). In vitro methods and more animal testing alternatives. Retrieved March 22, 2021, from https://www. peta.org/issues/animals-used-for-experimentation/alternatives-animaltesting/#:~:text=%20Alternatives%20to%20Animal%20Testing%20%20 1%20In,%E2%80%9Cmicrodosing%E2%80%9D%20can%20provide%20 vital%20information%20on...%20More 15. American Anti-Vivisection Society. (n.d.). Testing. Retrieved March 22, 2021, from https://aavs.org/animals-science/how-animals-are-used/testing/ ebwrbwrtbwrtb

SPRING 2021 | PENNSCIENCE JOURNAL

Extracellular Matrix Application in Regenerative Medicine Written By: Michelle Paolicelli Designed By: Bianca Vama

t is well-known that salamanders are able to regrow their tails completely after they have been severed and that starfish can similarly regrow their arms. Even more incredible is the ability that flatworms possess to regenerate their bodies after an injury of almost any severity. These seemingly supernatural abilities offer animals a huge competitive advantage and the ability to survive in their environments. Humans do not possess the same innate abilities as salamanders, starfish, and flatworms to regrow lost body parts, but what if there was a way to regrow a human finger or even a whole organ from scratch? The field of regenerative medicine explores this exciting possibility in the hopes of improving the quality and length of life for human patients. Salamanders are able to regrow their tails and other body parts because they have an entire class of stem cells not found in mammals that can differentiate into tissues and molecules1. Due to this, it is not feasible to simply mimic the

24 PENNSCIENCE JOURNAL | SPRING 2021

mechanism by which salamanders regenerate tissue in humans. Instead, researchers have turned to the extracellular matrix (ECM). The ECM is present in all tissues and organs and, as the name suggests, is external to the cells themselves. It contains collagen, enzymes, and various proteins and provides structural and biochemical support to the adjacent cells. More specifically, the ECM refers to the collective network of fibrillar proteins, proteoglycans, and glycosaminoglycans (GAGs) that make up the cell’s direct environment. The relative make-up of cells and ECM determine the characteristics of different tissues and bones. In general, more cartilaginous tissue contains a greater amount of hydrated ECM compared to bone where the ECM is highly mineralized2. The ECM provides a favorable scaffold for cell and tissue proliferation and its location on the outer surface of cells means that it acts as a sensor between the inner cell and surrounding environment. Through this function, it can signal cell growth and gene expression2. The role of the ECM in tissue engineering has become to tailor the phenotype of the resultant cells. By offering an in vivo, tailored environment for cell growth, the ECM opens new avenues for the field of regenerative medicine3. Although the regular use of ECM for limb regeneration in the clinical setting is still many years away, there have been recorded cases of its successful use in regenerating lost fingers. One specific case is that of Lee Spievak. After catching his finger in the propeller of a model airplane, the tip of his finger was severely damaged and subsequently removed by doctors. In most cases, a patient in Spievak’s shoes would let the injury heal and have to learn to adapt to the loss. However, Spievak was in the unique position: his brother had access to a lab experimenting with ECM applications in regenerative medicine. The ECM from the University of Pittsburgh lab was in powdered form and made from the lining of a pig’s bladder4. Although not harvested from humans, pig cells are similar enough to our cells that the ECM powder is still biocompatible. This means that it is not rejected by the human body and does not cause any harm. Spievak applied the powder from the lab to his severed finger every day

for about a month. Without the use of the ECM powder, the wound would have closed and scarred over. With the intervention of the ECM powder, new tissue cells grew instead. The result was that Spievak’s entire finger regrew in the span of only a few weeks. Incredibly, his fingernail also regrew after a short time. This example, although rare and unconventional, exhibits the amazing ability of the ECM, human or not, to influence successful cell phenotyping and growth4. A more lofty goal for ECM application is the ex vivo creation of tissues and organs. By engineering the ECM in the lab setting based on natural repair processes, it is possible to grow skin in an environment completely external to the human body. This has the potential to completely revolutionize burn treatments and other procedures that require skin grafts. Additionally, the ECM may even allow for the proliferation of extremely specialized cells such as those found in the liver and heart. If this method is researched and tested further it will have vast implications on the organ transplant process, potentially saving thousands of lives each year. By fine tuning the constituent parts that make up the ECM and engineering a cellular environment in which tissue proliferates, regenerative medicine has the potential to enter a new age. Another avenue of research involving the ECM is 3D bioprinting. By isolating, or decellularizing the ECM, specially engineered scaffolds can be deposited by 3D printers. This creates an opportunity to recreate the environment needed for cells to survive and grow in a controllable and accessible way. The versatility and commonality of the ECM throughout all body systems means that bioprinting offers numerous potential applications in the future5. As previously mentioned, ex vivo proliferation of cells using the ECM as a scaffold is also an emerging practice. This differs from 3D bioprinting in that ex vivo proliferation of cells requires significant time for sufficient growth while bioprinting extrudes the ECM in a desired geometry in a continuous manner6. Although the ECM is still not fully understood, and many of its applications are yet to be explored, early research and cases like Lee Spievak showcase the power of the human body to adopt foreign ECM. By taking compounds and substances already found in our bodies and engineering them for another purpose, medical care is greatly enhanced. It is clear that it is important

to innovate both on artificial man-made products Features Features and devices as well as naturally occurring substances. The ubiquitousness of the ECM throughout the body places it in a unique position to have a wide range of applications. The examples of the salamander, the starfish, and the flatworm show us what is possible, and the ECM offers a path to get there. Humans may never be able to regener ate their own limbs and organs, but through tissue engineering applications of the extracellular matrix, science fiction can become reality.

References 1. University of Pittsburgh Schools of the Health Sciences. (2018, August 13). When it comes to regrowing tails, neural stem cells are the key. ScienceDaily.www.sciencedaily.com/ releases/2018/08/180813160522.htm 2. Joven, A., Elewa, A., & Simon, A. (2019). Model systems for regeneration: Salamanders. Development, 146(14). doi.org/10.1242/ dev.167700 3. Yi, S., Ding, F., Gong, L., & Gu, X. (2017). Extracellular Matrix Scaffolds for Tissue Engineering and Regenerative Medicine. Current stem cell research & therapy, 12(3), 233–246. doi. org/10.2174/1574888X11666160905092513 4. Ransford, M. (2008, May 01). Man regenerates finger. Popular Science. https://www.popsci.com/ scitech/article/2008-05/man-regenerates-finger/ 5. Dzobo, K., Motaung, K., & Adesida, A. (2019). Recent Trends in Decellularized Extracellular Matrix Bioinks for 3D Printing: An Updated Review. International journal of molecular sciences, 20(18), 4628. doi.org/10.3390/ ijms20184628 6. Hussey, G.S., Dziki, J.L. & Badylak, S.F. (2018) Extracellular matrix-based materials for regenerative medicine. Nat Rev Mater 3, 159– 173. doi.org/10.1038/s41578-018-0023-x 7. Fernades, H. (2009). Extracellular matrix and tissue engineering application. Journal of Materials Chemistry, 19, 5474-5484. doi.org/10.1039/ B822177D SPRING 2021 | PENNSCIENCE JOURNALB Spring 2021 | PENNSCIENCE JOURNAL

Stem Cell Therapeutics The Solution of Tomorrow Written by: Harish Bayana

debates, along with other scientific factors related to its development. Despite this, scientists have recently managed to overcome many of these technical setbacks, and have set forth on a journey that will likely redefine the future of stem cells and their applications in the biomedical field.

Major Breakthroughs: iPSCs and Directed Differentiation

As stem cells mature, they follow a path of decreasing developmental potency, with totipotent cells being able to develop into nearly any cell in the body, while unipotent cells can only specialize within a single lineage of cell types

or decades, scientists have been captivated by the regenerative abilities of certain organisms that allow them to completely regenerate damaged or missing tissues and organs with a high degree of precision and functionality. Yet this seems to have been lost in evolution, with more complex organisms lacking such a system. As such, scientists have since focused on the mechanisms behind these regenerative abilities: stem cells. Every pivotal discovery in the field of stem cell research has seemingly brought about more expectations to completely revolutionize the future of regenerative medicine. This has allowed stem cell research to become publicly regarded as one of the most innovative avenues of study for the future. However, much of the history of this field has seen these optimistic views undermined by political, ethical, and religious 26 PENNSCIENCE JOURNAL | SPRING 2021

Of the stem cell lineages shown in the graphic, totipotent and pluripotent stem cells have the highest degree of developmental potency, meaning they can differentiate into nearly any type of cell in the body. As such, these cells would allow for the broadest extension of stem cell research and practical applications compared to other cell types, such as the much less potent adult stem cell. However, these cell types raise a major ethical dilemma, in that the source of these multipotent and totipotent cells must be an embryo. It is only from this embryo that human embryonic stem cells (hESCs) can be artificially cultivated in laboratory cultures. Therefore, research involving hESCs has been heavily restricted - for the majority of the lifespan of this field, stem cell research was limited to studying only adult stem cells1. However, in 2006, scientists were able to regenerate this pluripotency in weaker stem cells, marking the discovery of induced pluripotent stem cells (iPSCs). This development provided a way for scientists to circumvent the major ethical hurdles that accompanied embryonic stem cells by avoiding the need to use embryos while retaining the myriad of advantages that came with the pluripotency of embryonic stem cells. On top of this, iPSCs are self-renewing, meaning

one cell can provide a nearly limitless supply of human pluripotent stem cells. Altogether, this breakthrough helped to expand the potential for the application of stem cells in a number of fields, such as regenerative medicine, drug discovery, and disease modeling.

Despite this revolutionary finding, there still remained one major issue that needed to be addressed: without any way to actively control the process of differentiation, scientists had relatively little control over what the stem cells do, making them virtually useless in terms of their applications in health and medicine. To overcome this issue, researchers developed directed differentiation protocols, in which they are able to mimic the signals that cells receive as they undergo successive stages of specialization. By manipulating the environment and biochemical signals around the stem cells and understanding the mechanisms through which specific differentiation processes take place, scientists have been able to create differentiated cultures of stem cells2.

Applications of Stem Cells Another major issue that needs to be addressed is the fact that these therapies and research are negatively affected by a number of donor variables (i.e. age, health status) that compromise the efficacy of those stem cells. This makes standardizing the quality of iPSCs imperative to this field. To combat this issue, a growing infrastructure of stem cell banks has recently been developed. These banks will implement a foundation for many of the logistical factors that currently restrict this field by providing a more efficient means of generating, depositing, and delivering iPSCs to institutes. They will also help to alleviate the aforementioned issue of the donor variables by means of iPSC quality control tests, which ensure that all cells produced meet certain quality standards3,4. With all of these developments, stem cells have recently seen a significant increase in research in relation to their applications. One particular avenue of study, stem cell therapy, has an enormous range of applications, especially in relation to medicine and public health. These therapeutics aim to regenerate damaged or dead

Features

tissues by using the body’s own stem cells and growth factors. This treatment holds great potential for the future of treating diseases. This is especially important in combating serious medical conditions that are a result of improper differentiation or cell division, such as cancer or other birth defects. A number of stem cell therapies are currently being developed, including treatments for spinal cord injury, heart failure, eye degeneration, tendon ruptures, and type 1 diabetes2,5. These therapies can also be used to treat more common problems; for instance, arthroplasty (the regeneration of the function of a joint) has the potential to be greatly improved with a more complete understanding of stem cells6. Additionally, stem cells can help accelerate the development of drugs through means of improved toxicity testing. Since scientists are now able to artificially develop organs in the form of organoids, they are also thereby able to test experimental drugs or other therapies in these systems. This has enabled testing for undesirable effects of medications without needing to potentially harm people in the process. This can help to greatly improve the rate at which these treatments are being tested and developed, thereby saving both time and money in developmental costs. This procedure is currently being implemented in treatments that involve the heart or the liver, and soon scientists hope to be able to expand this to other organs as well7.

Limitations of Stem Cell Therapeutics Directed differentiation and the many stem cell associated therapeutics are becoming increasingly efficient in terms of both time and resources, allowing this field to expand at an unparalleled rate. However, one major issue that remains is the fact that many of these applications have not been implemented in vivo. Scientists are still working to develop culture conditions that will SPRING 2021 | PENNSCIENCE JOURNAL

allow for the differentiation of these iPSCs into the desired functional tissues. This means that References developing fully functional tissues and organs that can be implemented within a person remains a distant 1. Gottweis, H. (2010). The Endless hESC goal. Without this, most medical applications of stem Controversy in the United States: History, cells in relation to regenerative abilities will still Context, and Prospects. Cell Stem Cell, need years of research to be viable8. 7(5), 555–558. https://doi.org/10.1016/j. stem.2010.10.008

As described earlier, stem cell banks show great promise in helping facilitate a global infrastructure for the development and delivery of stem cell cultures. However, the current state of these banks is not sufficient to keep up with global research initiatives. The iPSC quality control tests require a great deal of resources to repeatedly implement, meaning that developing these cultures will need a constant source of funding and a sufficient amount of time, hindering the goals of scientists worldwide. Thus, these banks will need to expand into more countries and will need to become more commonplace if they hope to ever be able to effectively solve this issue.

Zakrzewski, W., Dobrzẏski, M., Szymonowicz, M., & Rybak, Z. (2019). Stem cells: Past, present, and future. Stem Cell Research & Therapy, 10(1), 68. https://doi. org/10.1186/s13287-019-1165-5

Huang, C.-Y., Liu, C.-L., Ting, C.-Y., Chiu, Y.T., Cheng, Y.-C., Nicholson, M. W., & Hsieh, P. C. H. (2019). Human iPSC banking: Barriers and opportunities. Journal of Biomedical Science, 26(1), 87. https://doi.org/10.1186/s12929-019-0578-x

Challenges in Ensuring Human Pluripotent Stem Cell Quality. (n.d.). Retrieved March 29, 2021, from https:// www.stemcell.com/nature-research-roundtable-hPSCquality

Ouyang, X., Telli, M. L., & Wu, J. C. (2019). Induced Pluripotent Stem Cell-Based Cancer Vaccines.

The current state of stem cell therapies and Frontiers in Immunology, 10. https://doi.org/10.3389/ regenerative medicine is far from its potential. The fimmu.2019.01510 US Food and Drug Administration has currently 6. Wang, Y., Jin, S., Luo, D., He, D., Shi, C., Zhu, L., only approved a single stem cell treatment: the Guan, B., Li, Z., Zhang, T., Zhou, Y., Wang, C.-Y., use of hematopoietic progenitor cells in patients & Liu, Y. (2021). Functional regeneration and repair 3 with disorders that affect the production of blood . of tendons using biomimetic scaffolds loaded with This is the result of over 350 businesses and 570 recombinant periostin. Nature Communications, 12(1), clinics across the nation marketing stem cell-based 1293. https://doi.org/10.1038/s41467-021-21545-1 interventions, and a myriad of clinical trial failures for various therapeutics in the past few decades9. 7. Cressey, D. (n.d.). Stem cells take root in drug This brings to light the fact that stem cell treatments development. Nature News. https://doi.org/10.1038/ nature.2012.10713 are often seen as this ongoing revolution in the field of medicine, but in reality, there remains a great 8. Research, C. for B. E. and. (2020). Consumer Alert on deal of work that needs to be done before any Regenerative Medicine Products Including Stem Cells lasting changes can be made.

Key to the Future

and Exosomes. FDA. https://www.fda.gov/vaccinesblood-biologics/consumers-biologics/consumer-alertregenerative-medicine-products-including-stem-cells-andexosomes

As a whole, stem cell research has 9. Turner, L. (2018). The US Direct-to-Consumer seen a rapid evolution in the past few decades Marketplace for Autologous Stem Cell that has brought it to the forefront of public Interventions. Perspectives in Biology and health. As of now, expectations for this Medicine, 61(1), 7–24. https://doi.org/10.1353/ field outpace the actual implementation of pbm.2018.0024 these therapies in scenarios outside of the laboratory. However, this field has recently taken a promising turn, increasing the likelihood that these therapeutics will soon revolutionize the future of medicine. 28 PENNSCIENCE JOURNAL | SPRING 2021

Features

The Role Of AI In Medicine And The COVID Pandemic Written by Kevin Guo Designed By Amara Okafor

What is artificial intelligence? Is it just a machine that can make a billion calculations instantly? Or is it something that we ought to fear? Well, one way to define artificial intelligence is that it is a set of computer systems that are given organized data sets and analyze each data point to mitigate error. The method is improved by running through the data sets many times and optimizing data analytical approaches. Generally, these systems utilize algorithms – processes of sets and rules that are adhered to in calculations – that evaluate each data set, so programmers can assess AI’s ability to determine the correct answer8. Regardless of the way in which we define AI, it has had an importance influence in medicine and continues to do so during the COVID-19 pandemic. SPRING 2021 | PENNSCIENCE JOURNAL

Features

Before delving into AI’s function in the healthcare setting, it is important to understand the types of AI in medicine and their history. AI is usually divided into two groups: virtual and physical. Virtual AI may assist “electronic health record (EHR) systems and neural networkbased guidance in treatment decisions” while physical AI manages robots “assisting in performing surgeries, intelligent prostheses for handicapped people, and elderly care2.” One important type of AI seen often is machine learning, which is “a statistical technique for fitting models to data and to ‘learn’ by training models with data.” It is very common in many organizations in the US; in fact, a 2018 survey of 1,100 US managers showed that 63% of companies employed machine learning at some capacity3. In healthcare specifically, one of the fields where machine learning is prevalent is precision medicine – determining the likelihood of success for each treatment method based on patient characteristics and the circumstances surrounding the treatment. This field requires some sort of “supervised learning,” where the outcome of a patient’s treatment is already known based on an inputted “training dataset.”4 It is akin to predicting whether the physical flipping of a coin will result in heads or tails using several variables such as the initial tilt of the coin, its size, the initial velocity/force applied, and air resistance. AI first saw its use in the form of a program called Logic Theorist which was engineered by Allen Newell, Cliff Shaw, and Herbert Simon. Funded by the Research and Development (RAND) Corporation, the program was supposed to “mimic the problem solving skills of a human.” In a conference led by esteemed computer scientist John McCarthy, this project was discussed and the term “artificial intelligence” was coined by McCarthy himself. Even though there were no standard methods established in this field yet, there was a consensus that AI

30 PENNSCIENCE JOURNAL | SPRING 2021

was achievable. This prediction turned out to be correct, as that conference spurred the next several decades of AI research3. One example of such research in the field of medicine is the development of DXplain, a decision support system developed in 1986 that gave “a list of probable differentials based on the symptom complex” and served “as an educational tool for medical students filling the gaps not explained in standard textbooks”. Another example was Germwatcher, a relatively recent innovation that “detect[s] and investigate[s] hospital acquired infections2.” AI has many applications in the modern day healthcare system. For example, it is often used to handle patient data on a daily basis. This would allow doctors to spend more time engaging with patients. For example, a 2016 study found that “physicians spent 27% of their office day on direct clinical face time with their patients and spent 49.2% of their office day on electronic hospital records (EHRs) and desk work. When in the examination room with patients, physicians spent 52.9% of their time on EHR and other work8.” With AI to reduce doctors’ workload – especially any repetitive paperwork – significantly, this allows doctors to spend more time with patients and more accurately assess patients’ conditions and symptoms without having to worry about manually filling in too much paperwork regarding the patient’s attributes. In addition to data management, AI has had a profound impact on radiological imaging and clinical screening. From incorporating CAD (computer-assisted diagnosis) in a screening mammography to identifying abnormal exams or results, AI has substantially supported radiologists in various imaging capacities, especially in situations that require high volumes of imaging or with less human resources2. During the ongoing COVID-19 pandemic, AI’s role has been important in several different areas. For instance, a paper published in the Journal of the American Medical Informatics Association demonstrated a machine learning model’s ability to predict the prognosis of COVID-19 patients by evaluating individual characteristics. It can designate each patient “in terms of clinical states – moderate, severe or critical – as well as hospital utilization5,” which is defined as how a certain health care facility is used. One way AI could reduce the impact of COVID-19 is by minimizing in-person

contact. The Singaporean company NCS has proven this by “deploy[ing] AI-powered thermal cameras with several public healthcare providers in Singapore … These cameras can be integrated to the visitor and employee management system and contact tracing systems to facilitate monitoring and further follow-ups as required by the healthcare providers.”1 With systems such as these that can visualize potential virus carriers, health authorities are able to initiate contact tracing and prevent any potential spread of the virus without having to interact with patients directly. AI has also had an early role in modeling the COVID-19 virus and identifying any potential drug or vaccine mechanisms that could target the pathogen and potentially neutralize it. One such example is how the company BenevolentAI provided AI-derived biomedical data to predict “commercially available antiviral drugs that may target the SARS-COV2-related protease and helicase7.” Another AI model, called the “iNeo tool,” was used to help design a vaccine that contained B-cell and T-cell antibody binding sites. With these types of models, potential new strategies of “killing off” the COVID-19 virus could be investigated in the presence of potential mutations7. Not all clinicians have an optimistic viewpoint on AI. However, for the most part, doctors believe that AI will be of enormous benefit to physicians in need. Theodoros Zanos, PhD, head of the Neural and Data Science Lab and assistant professor at the Feinstein Institute for Medical Research at New Hyde Park, asserts that he does not believe that “physicians and nurses can be replaced by algorithms any time soon but they can be tremendously assisted by these methods so they can focus more on the irreplaceable aspects of their work, like human interaction5.” Oscar Marroquin, MD, chief clinical analytics officer at UPMC, shared a similar sentiment, citing that most clinicians would agree that “AI in diagnostics affords [them] the opportunity to be more efficient in how they do their work. There are tasks that machines can do well and reliably well, and we are going to take advantage of those things to increase workflow while at the same time improving accuracy and efficiency of how we deliver care5.” No matter our opinions on the subject, there is no doubt that AI is and will continue to be a major facet of our

Features lives, especially when we have to go see a doctor. From managing basic patient data to finding the right treatments based on our reported symptoms to even addressing a global pandemic, AI and the methods of developing these systems never cease to demonstrate their capacity to adapt to any situation.

REFERENCES

1. AI-powered solutions in TACKLING COVID-19 and beyond. (2020, December 15). Retrieved February 28, 2021, from https://www.healthcareitnews.com/news/asia-pacific/aipowered-solutions-tackling-covid-19-and-beyond 2. Amisha, Malik, P., Pathania, M., &amp; Rathaur, V. (2019, July). Overview of artificial intelligence in medicine. Retrieved February 28, 2021, from https://www.ncbi.nlm.nih. gov/pmc/articles/PMC6691444/ 3. Anhoya, R. (2020, April 23). The history of artificial intelligence. Retrieved March 07, 2021, from https://sitn.hms. harvard.edu/flash/2017/history-artificial-intelligence/ 4. Davenport, T., &amp; Kalakota, R. (2019, June). The potential for artificial intelligence in healthcare. Retrieved February 28, 2021, from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6616181/ 5. Dyrda, L. (2020, January 10). How ai is changing medicine, the role of physicians. Retrieved February 28, 2021, from https://www.beckershospitalreview.com/artificial-intelligence/how-ai-is-changing-medicine-the-role-of-physicians. html 6. Jercich, K. (2021, January 22). New AI model can PREDICT length of COVID-19 hospitalization. Retrieved February 28, 2021, from https://www.healthcareitnews.com/news/newai-model-can-predict-length-covid-19-hospitalization 7. Keshavarzi Arshadi, A., Webb, J., Salem, M., Cruz, E., Calad-Thomson, S., Ghadirian, N., . . . Yuan, J. (2020, July 17). Artificial intelligence FOR COVID-19 drug discovery and vaccine development. Retrieved February 28, 2021, from https://www.frontiersin.org/articles/10.3389/ frai.2020.00065/full 8. Greenfield, D. (2019, June 19). Artificial intelligence in Medicine: Applications, implications, and limitations. Retrieved February 28, 2021, from https://sitn.hms.harvard. edu/flash/2019/artificial-intelligence-in-medicine-applications-implications-and-limitations/ 9. Vaishya, R., Javaid, M., Khan, I., &amp; Haleem, A. (2020, July). Artificial intelligence (AI) applications FOR COVID-19 pandemic. Retrieved February 28, 2021, from https://www. ncbi.nlm.nih.gov/pmc/articles/PMC7195043/

SPRING 2021 | PENNSCIENCE JOURNAL

Applications of Quantum Mechanics in Magnetoencephalography Written By: Benjamin Beyer Designed By: Bianca Vama Quantum mechanics has helped us understand the human brain better than ever before. Even though quantum mechanics is a new theory dealing with the properties of small particles, its insights have spurred great progress in society. By understanding quantum mechanics, we have been able to innovate in unfathomable ways, revolutionizing computers, smartphones, lasers, GPS, and medical technology. Before the twentieth century, the prevailing theories of physics were those based on classical mechanics. Derived from the work of Newton, Hamilton, and Lagrange, these theories attempted to use the principles of macroscopic objects to describe everything around us, from the largest phenomena to the smallest. As science became more advanced over time, classical mechanics could not describe certain properties of atoms like discrete energy levels and electromagnetic radiation; thus, researchers began to search for a theory that could describe these phenomena. In the early twentieth century, the theory of quantum mechanics was devised and provided a much more probabilistic theory that better described the smallest particles. Quantum mechanics leads to many propositions that would never be possible under classical mechanics like tunneling, entanglement, and superconductivity. This technology is currently being harnessed to streamline processes in the medical field with applications like magnetoencephalography (MEG), a groundbreaking neuroimaging technique. MEG works by measuring small changes in the magnetic fields of the brain and using them to map the functions of the brain. Thus, it can be used to identify and treat brain disorders such as Parkinson’s disease. The brain is a complicated network of neurons that communicate through electrical signaling. In the late sixties, scientist David Cohen discovered neural magnetic fields within our brains created by the brain’s electrical activity1. Cohen’s findings, coupled with modern day physics, have made MEG technology a reality. To conduct these brain scans with MEG scanning, our understanding of quantum mechanics is implemented to help interpret incredibly small changes in neural magnetic fields in order to record brain activity. The scanners work by using rubidium-87 atoms energized to a specific quantum state. In this state, the magnetic fields will cause minute shifts in the angular momentum of these atoms, which can be measured and used to map the neural activity 32 PENNSCIENCE JOURNAL | SPRING 2021

of the brain. To calibrate the field, a second outside electric field is run such that it can be measured in respect to the magnetic fields of the brain2. This allows the MEG scan to be a comprehensive scan that provides medical professionals detailed insight into the functioning of a patient’s brain. MEG technology holds major advantages over other brain-imaging technologies like electroencephalography (EEG), which is the measuring of the brain’s electric pulses and fields to form a holistic scan of the brain. Although it is expensive, unlike EEG scans, MEG scans do not face the obstacle of distortion as the signals pass through patients’ heads3. This technology has not reached its true potential, and the capabilities of MEG scans are already better at producing these precise scans4. Therefore, the optimization of MEG technology is important so that it can be used more widely and revolutionize brain scanning. While MEG is great for making thorough scans of the brain, the technology has encountered a few limitations. The magnetic field of the Earth is much stronger than those of the brain, rendering the data from the MEG scan unreadable. This is usually mitigated by limiting the patient’s range of motion to be within a millimeter such that the magnetic field of the Earth is known and can be removed, producing a readable scan. This requires that the patient be incredibly still, which has been a glaring downside of the technology; however, recent innovations such as Helmholtz cells have helped mitigate this problem. The use of Helmholtz cells --powerful pieces of electronic equipment that are able to stabilize electric fields -- in tandem with the MEG scan have allowed the researchers to effectively cancel out outside magnetic fields such that the patient can have a wider range of motion5. Another large hurdle to overcome is that the MEG scanning process requires a vast amount of processing power in order to interpret its many data points. Bayesian algorithms that implement the statistical methods relating to Bayes’ Theorem, an important theorem of probability, are employed to find a structure for the brain that fits the data collected. In this context, Bayesian algorithms use prior assumptions along with the data to attempt to assemble a scan from the data of the brain’s magnetic fields6. In order to progress MEG scanning, these algorithms

Features

must be improved to increase the accuracy of the scans. While these issues must be resolved to improve the capabilities of MEG scanning, MEG scanning remains a powerful tool for brain imaging. Looking towards the future, MEG scanning has the capabilities to greatly improve our efforts to image brains and brain activity. This technology has potential for widespread applications across medicine such as the recognition of which parts of the brain contribute to the operation of different bodily systems. Furthermore, it has shown promise in identifying autism, Alzheimer’s, and other psychiatric diseases7. This is done by comparing scans of unaffected brains to those of patients to determine if the brain’s activity is being altered by a psychiatric disorder. While MEG scanning has not been optimized for widespread use, it has been approved for preoperative brain scanning for surgeries pertaining to disorders like epilepsy4. As we learn more about MEG scanning and better implement it into our hospitals and clinics, our brain imaging capabilities will be much improved, and we will have the ability to further understand and treat a wide variety of brain disorders. With the potential for high resolution scans that have many advantages over other scanning methods, MEG scanning is poised to be further implemented into the public health sector, and the applications of quantum mechanics continue to help innovate in many facets of society. 1.

References:

Cohen, D. (1968). Magnetoencephalography: Evidence of magnetic fields produced by alpha-rhythm currents. Science, 161(3843), 784-786. doi:10.1126/science.161.3843.784 2. Boto, E., Holmes, N., Leggett, J., Roberts, G., Shah, V., Meyer, S. S., et al. (2018). Moving magnetoencephalography towards real-world applications with a wearable system Springer Science and Business Media LLC. doi:10.1038/nature26147 3.

Zhang, R., Xiao, W., Ding, Y., Feng, Y., Peng, X., Shen, L., et al. (2020). Recording brain activities in unshielded earth’s field with optically pumped atomic magnetometers. Science Advances, 6(24), eaba8792. doi:10.1126/sciadv.aba8792 4. Singh, S. P., & Singh, S. P. (2014). Magnetoencephalography: Basic principles full text 5.

Coleman, H., & Brookes, M. (2021). Quantum physics gives brain-sensing MEG scanners a boost. https://physicsworld.com/a/quantum-physics-gives-brain-sensing-meg-scanners-a-boost/ 6.

Proudfoot, M., Woolrich, M. W., Nobre, A. C., & Turner, M. R. (2014). Magnetoencephalography BMJ. doi:10.1136/practneurol-2013-000768

7. Having a MEG scan. (2021). https://www.alzheimers.org.uk/research/take-part-research/ meg-scan

SPRING 2021 | PENNSCIENCE JOURNAL

The Rheological Behavior of Firn: Experimental Observations of Dislocation Creep via Grain Boundary Sliding By: D. R. Furman1 Advisor: D. L. Goldsby1 Department of Earth and Environmental Science, University of Pennsylvania, Philadelphia, PA 19104 Corresponding author: Daniel R. Furman (dfurman@sas.upenn.edu)

Acknowledgements: This work was conducted with the materials and resources in the Penn Experimental Geophysics Laboratory and funded through Penn CURF. We are grateful to Prof. Richard B. Alley for discussions on the near-surface cryosphere, and for inputs from Andrew J. Cross and doctoral student Travis Hager, which greatly improved the experiments, and the content and clarity of the research. All data and analyses required to replicate the numbers, figures, and models within are shared open access at the following link: https://github.com/daniel-furman/ice-densification-research.

Key Points

Abstract

Firn densifies through a number of processes at the near-surface of ice sheets and glaciers, with diffusion creep and dislocation creep previously identified as operative mechanisms. Here, we performed a series of compaction experiments on ice powder samples synthesized with differing grain size, characterized by ultra-fine (~ 5 µm), fine (~ 17 µm), medium (~ 187 µm), and coarse (~ 550 µm) radii. Mechanical tests were performed at constant stress (0.3 – 1.4 MPa), at a constant temperature (233 K), and between 80 to 90% relative density. Steady-state creep rates were analyzed via a flow law (Eq. 1) with a power-law relationship between the densification rate (E), applied stress (o), and grain diameter . The creep rates were found to be dependent on grain size in fine- and ultra-fineKey Words grained samples, with a stress exponent n ~ 1.6 and a grain size exponent p ~ 0.9, and independent of • Ice sheets, glaciers, firn densification, grain size in medium- and coarse-grained samples, microstructure, grain size sensitivity, with n ~ 3.7 and p ~ 0. We show that these data dislocation creep, disGBS, the field boundary represent two creep mechanisms, disGBS and hypothesis, constant-stress creep tests dislocation creep respectively, as is observed in the flow behavior of solid ice in terrestrial and planetary ice bodies.

• Constant-stress laboratory experiments were performed on H2O ice powder samples with roughly uniform grain sizes varying from 5 to 550 micrometers (µm) in radius. • Two rheologically-distinct creep regimes emerged, characterized by their stress dependence n and grain size sensitivity p: dislocation creep (n ~ 3.7, p ~ 0) and disGBS (n ~ 1.6, p ~ 0.9). • Extrapolation of flow laws for disGBS and other creep mechanisms show that disGBS is the rate-limiting mechanism for natural conditions, such as in glaciers and ice sheets.

34 PENNSCIENCE JOURNAL | SPRING 2021

Plain Language Summary Vast deposits of partially dense ice, or firn, form layers in the near-surface of glaciers and ice sheets on earth and in space. To determine the flow properties for these ice bodies, we produced and then deformed samples of ice powder in the laboratory, a controlled setting wherein we isolated (and varied) particular physical variables. Experiments were conducted in the freezer and took anywhere between two hours to fourteen days in total, depending on the rate of deformation. We found that the ice compaction rates were independent of grain size at coarser grain sizes. However, among finer grain sizes, we discovered a mechanism directly sensitive to grain size, meaning that samples composed of ultra-fine grains flowed more rapidly than samples of fine grains. Grain size sensitivity for firn should be further tested and, upon future consensus review, included in glaciological models.

1. Introduction Firn is defined as polar snow surviving longer than a single melt season, a transition state between snow and solid ice (Baker, 2019). In general, glaciers and ice sheets increase their mass through an accumulation of snow at the surface, which seasonally stratifies into firn layers at depths of up to ~80 m. The firn column requires anywhere from hundred(s) to several thousands of years to fully densify in terrestrial settings (Ebinuma and Maeno, 1987, Schwander et al., 1997). Near-surface densification is an important process with implications for understanding the cryosphere’s climate records. Trapped gases hundreds of thousands of years into the past are contained within glaciers and ice sheets; however, dating uncertainties arise from the time lag between younger atmospheric gases that permeate the firn down to the bubble close off depth and the older age of the ice at the close off depth (Faria et al., 2010, Adolph & Albert 2014). Glaciologists typically turn to models to date ice records by determining rates of densification,

Research

backing out the time it took for pores to become disconnected from one another and from the atmosphere (Wilkinson 1988). Densification models are also critical for determining the overall mass of ice bodies, particularly from measurements of surface elevation from satellites, which are influenced by the density structures of near-surface ice in the conversion of volume to mass (Li et al., 2011, Lenaerts et al., 2019, Keenan et al., 2020). Ice deforms through steady-state viscous flow, or creep, in low deviatoric stress environments common to the cryosphere (Goldsby 2006). Ice creeps at rates that far exceed its theoretical strength, a phenomenon linked to the presence of lattice defects – dislocations – in the ice. Dislocation creep mechanisms refer to modes of deformation that move dislocations through the material, dominant in ice, metals, alloys, and other materials. Steady-state creep is described by constitutive equations (flow laws), which are relationships between the creep rate (E ̇) and the applied stress (o) (with ). In expanded form, the flow law includes the environmental conditions and material properties of influence on creep rate (e.g., Eq. 1), such as temperature and grain size (e.g., Goldsby 2006). Ice’s material properties include its grain size (Figure 1b), particle-pore geometry, crystallographic preferred orientation, and impurity content (Freitag et al., 2013, Faria et al., 2014), all of which may influence firn’s creep rate. The flow law includes quantifiable parameters that encode each material’s unique response to the deformation conditions, such as the power law exponents for stress (n) and grain size (p). These parameters are determined directly from laboratory data, with rheological testing as described herein and, for example, by Goldsby and Kohlstedt (2001) and Durham et al. (2001). In contrast, empirical models for firn densification (e.g., Herron, and Langway, 1980) SPRING 2021 | PENNSCIENCE JOURNAL

Figure 1. (a) Map of Antarctica’s average annual surface temperature, from Breant et al. (2017). Notice that our 233 K experimental temperature (-40 oC) is a common temperature in the interior regions. Density-stress profiles from Byrd, Dome C, Vostok, Mizuho (NE Antarctica), and GISP-2 (Greenland) are explored herein. (b) Firn from the EDML site at 70 m depth, from Faria et al. (2014) (scale bar = 1 mm). The pore space and grain boundaries are both in black. Dynamic recrystallization processes may have nucleated the smallest grains (r ~ 0.1 mm).

overlook the constitutive physical relationships and are based instead on compaction data from field measurements. This class of models is highly tuned to its training data, and, while they perform well when extrapolated to similar conditions, empirical models are hindered in extrapolation to novel conditions (Spencer and Alley, 2001, Schultz et al., 2021). This leads to the hypothesis that a physics-based modeling approach is preferred for many glaciological applications (Wilkinson 1988, Keenan et al., 2020), such as dating trapped gas records compacted under conditions absent from the field today, so long as the flow law accurately captures the operative mechanisms at play.

for terrestrial ice sheets at the intermediate stage vary from 1e−11 to 1e−12 (s-1), as measured by the rate of change in density (p/pice) (Wilkinson and Ashby, 1975). These values were resolved from five density-stress field profiles spanning the terrestrial cryosphere from Antarctica to Greenland (Ebinuma and Maeno, 1987) (Figure 1a). Overall, cumulative strains often reach tens of percent in the firn column, much larger than those characteristic of solid ice creep (Faria et al., 2014, pers. comm., R. Alley, 2020). The firn network develops stress localizations as a result, which in turn activates a number of dynamic microstructural processes (Figure 1b) (Kipfstuhl et al., 2009, Faria et al., 2014).

Firn densification occurs in ice with three distinct geometries, with the latter two dominated by steady-state creep. The primary stage consists of frictional grain rearrangement, vapor phase and surface diffusion, as well as melting and refreezing (Maeno and Ebinuma, 1983, Alley, 1987, Schultz et al., 2021). Steadystate creep becomes rate-limiting for both the intermediate and final stage geometries, meaning that creep dominates the overall rate of deformation. For the intermediate stage, the relative density (pr) varies between 80 and 90% of the density of solid ice (Maeno and Ebinuma, 1983). The pore network remains interconnected until pore closure, i.e., at the boundary between the intermediate and final stages. Creep rates

New experiments exploring firn creep are warranted for many reasons. Contributions from grain size-sensitive and grain sizeinsensitive creep mechanisms are common to natural settings (DeBresser et al., 1998), yet most previous firn flow laws lack this feature (Maeno and Ebinuma, 1983, Wilkinson, 1988, Meussen et al., 1999). For example, disGBS (dislocation-accommodated grain boundary sliding) and dislocation creep both contribute to olivine’s densification rate in crustal magma systems (Yao et al., 2019). Previous firn flow laws simply consider one mode of grain sizeinsensitive creep (i.e., classical dislocation creep), with the Glen law n = 3 stress exponent (Glen 1952, 1955). These models overlook the grain

36 PENNSCIENCE JOURNAL | SPRING 2021

size sensitivity observed between firn’s stratified layers in the field (e.g., Alley, 1982) and the grain size sensitivity observed in solid ice creep (Goldsby and Kohlstedt, 2001, Durham et al., 2001). Overall, this study was predicated on the possibility that we might observe new forms of grain size-sensitive compaction by testing a range of fine-grained samples.

2. Methods 2.1 Starting Material Powdered ice of nearly uniform grain size was pressed into a dead-weight-loaded compaction die, to the onset of intermediate stage densification (pr = 0.8). Ultra-fine-grained samples of ~5 µm grain radius (Qi et al. 2018) were synthesized by spraying a mist of distilled water through an aerating nozzle into a bath of liquid nitrogen, which was eventually evaporated from the accumulated grains. Fine-, medium-, and coarse-grained samples of ~17, 187, and 550 µm radii were synthesized by crushing distilledwater ice cubes in a kitchen blender. The resulting powder was then sieved to the specified grain size, while occasionally clearing clogged gaps in the mesh with an air gun and by pouring liquid nitrogen over the sieve (Goldsby and Kohlstedt, 2001). The powders were funneled into the die (with a 1-inch diameter) and pressed to an initial porosity near the onset of the intermediate stage, ~80% relative density.

2.2 Compaction Tests The compaction die (Figure 2a/b) was loaded with weights to a constant applied stress. However, our experiments were not truly constant stress, as the contact area of grains supporting the load increased during densification. Displacement data were recorded with a DC-powered linear variable displacement transducer (LVDT) at a sampling rate of 0.001 Hz. Time-series densification curves were determined from the volume of the sample, which was measured continuously, and the sample mass, which was determined after testing. The data revealed a transient response followed by steady-state creep; for the latter,

Research

densification occurred at a constant creep rate (i.e., the sample compacted linearly with time). Densification rates for steady-state creep (p/ pice) (s-1) were resolved from the time-series slope, calculated in aggregation to match the densification curve (Appendix Jupyter Notebook, Section 1). Rates corresponded to relatively small changes in density; therefore, the mean relative density was considered representative for the measurement. In addition, interconnected air pockets were assumed to support a negligible load, and pore pressures were ignored due to the channel interconnectivity. Thus, the compressing load was transformed to the applied stress (o/pr), which ranged from 0.3 to 1.4 MPa, to correct for the area of pore contacts with the piston surface. Some tests were performed under a single applied stress, yielding one rate measurement, while other experiments, particularly those with relatively small rates of deformation, included steps in applied stress. The die sat in a relatively large freezer held at a fixed temperature of 233 K for each experiment, representative of Antarctic firn temperatures (Figure 1a). The temperature variation for our tests was less than one degree Celsius. Variations in temperature may have exceeded the one-degree limit solely at the beginning and end of an experiment, when the top of the freezer was open; thus, these portions were clipped from the densification curve outputs. For each experiment, the bore was first cleaned with ethanol and then lubricated with molybdenum disulfide spray, minimizing friction between the sample and the die. In addition, the piston was inspected for smoothness and polished with fine sandpaper to remove any burrs. It is also worth noting that any lateral forces resulting from the sample’s expansion were minimized by performing experiments at relatively low applied stress. SPRING 2021 | PENNSCIENCE JOURNAL

Figure 2. (a) Photograph of dead-weight compaction die used for creep tests and (b) a corresponding diagram. (c) Photograph of a coarse-grained sample, post-compaction (white scale bar = 0.5 inches). Close inspection of the surface reveals three individual grains (shaded and outlined in black).

3. Results We measured sixteen total creep rates across ultra-fine, fine, medium, and coarse series (grain radius ~ 5, 17, 187, 550 µm), color-coded onto Figure 3’s data. The steady-state creep results reveal two distinct mechanisms, with the slope of the linear regression corresponding to the stress exponent (n)( ) across log-log Ė-o space. The n exponents for the fine and ultra-fine series of samples (n = 1.57 ± 0.22, n = 1.68 ± 0.45) suggest the operation of a common mechanism, with an average stress exponent of n = 1.63 ± 0.34. The rate differences between the ultra-fine- and finegrained series indicate the mechanism’s grain size sensitivity. A distinct n exponent emerged for the medium-grained samples (n = 3.74 ± 1.02), with the overlapping data indicative of this mechanism’s grain size insensitivity. Uncertainties for the n exponents were determined from the 95% confidence intervals of the linear regression. The mean relative density for the rate measurements were kept nearly consistent within each series (pr = 0.831 ± 5.6e-3, 0.818 ± 7.1e-3, 0.815 ± 3.3e-3), essential as the density is a variable that influences the creep rate (Eq. 1). These mean and 95% confidence intervals were estimated from bootstrap resampling of the relative density measurements. Experiments were conducted at a

38 PENNSCIENCE JOURNAL | SPRING 2021

relatively low temperature (233 K) to avoid pre-melt along grain boundaries (Goldsby, 2006) and grain growth. The creep rates were analyzed with the intermediate stage flow law for viscous creep (Wilkinson and Ashby, 1975, Maeno and Ebinuma, 1983), modified to include the grain-size dependence observed in our compaction experiments (Eq. 1). The densification flow law includes the creep rate Ė, material constant A, applied stress o, grain size diameter d, temperature T, activation energy Q, gas constant R, n and p exponents, and relative density pr. The power law relationship between creep rate and grain size was added to the original model derived by Wilkinson and Ashby (1975), with the material constant A cancelling the newly introduced units. When p is non-zero, the grain size power law alters the numerical behavior of the model, for example, subtly weakening the dependence of creep rate on density. Further research is warranted to encapsulate the grain size dependence in our compaction data without altering the numerical behavior of the original flow law.

Research

Figure 3. Creep rate versus applied stress, in log-log transformed space. The data constrain our flow law (grey lines) by resolving the rate-limiting mechanism. The stress exponents (n ~ 1.6, n ~ 1.7, n ~ 3.7) are taken from the slope of the linear regression (dashed lines).

4. Discussion A disGBS regime and dislocation creep regime were observed for fine and medium-grained samples., respectively, a result consistent with flow laws for solid ice creep (Goldsby, 2006). This leads us to the hypothesis that both mechanisms contribute to the overall creep in natural settings (Behn et al., 2021), particularly for firn columns with grain size distributions significantly deviated around their mean.

4.1 An Updated Framework for Firn Creep

These values match well with those for disGBS in fully dense ice, where n = 1.8 and p = 1.4 (Goldsby and Kohlstedt, 2001, Durham et al., 2001). The disGBS mechanism occurs through grain boundary sliding accommodated by the passage of dislocations along basal slip systems (Goldsby, 2006). The close match between these values of the stress exponent in our tests and for creep of solid ice deforming by disGBS strongly suggests that the mechanism is the same in both cases, as the stress exponent is considered a fundamental property of the steady-state creep mechanism (Wilkinson and Ashby, 1975). Solid ice’s p = 1.4 grain size exponent, in contrast, is significantly different from our results for firn densification, over 50% higher than the value of p ~ 0.9 in our experiments. This is perhaps due to the differences in firn’s microstructure compared with solid ice; particularly, in firn, grains have much more freedom to move than in solid ice. The firn grains’ proclivity to move perhaps makes it more difficult to develop pinned triple grain junctions compared to solid ice, a key factor for disGBS creep (Goldsby and Kohlstedt, 2001). We assume the activation energy is equivalent to disGBS in solid ice (Q = 49 kJ/mol) (Goldsby, 2006), as was assumed for dislocation creep in the previous firn flow laws (Ebinuma and Maeno, 1983, Wilkinson 1988).

The data in the medium- to coarse-grained regime can be fit with a stress exponent of n = 3.74 ± 1.02, and a pre-exponential term A ~ 1.48e5, with the activation energy set to the solid-ice equivalent (Q = 60 kJ/m) (Goldsby, 2006). Dislocation creep exhibits a negligible grain size dependence (p ~ 0), evidenced by the overlap between medium- and coarse-grained data (Figure 3). Additionally, the stress exponent for the medium-grained mechanism may more closely match dislocation creep (n = 4) in solid ice (Goldsby and Kohlstedt, 2001, Durham et al., 2001), with bootstrap resampling among the medium-grained series yielding an n ~ 4 estimate of the mean. In contrast, the finegrained mechanism fits a stress exponent n = 1.63 ± The experimental grain size sensitivity was 0.34, a grain size constant A ~ 0.44. clearly distinct from diffusion creep mechanisms, which exhibit differing flow law parameters (n = 1, p

SPRING 2021 | PENNSCIENCE JOURNAL

= 2/3) (Nabarro, 1948, Herring, 1950, Coble, 1963, Coble, 1970). While diffusion creep is dominated by other mechanisms in most natural settings (Wilkinson 1988), it is nevertheless still operative during deformation (Duval et al., 1983, Burr et al., 2019). However, we chose to ignore diffusion creep here, as its flow law contains uncertainty due to a lack of confirmation in laboratory experiments. In addition, other grain size-sensitive mechanisms may contribute to firn creep, with their identification subject to future testing across expanded laboratory conditions.

4.2 Extrapolation to Natural Settings We first extrapolated the flow law to the full range of stresses and grain sizes in terrestrial ice sheets and glaciers (Figure 4a). The disGBS mechanism is rate-limiting with decreasing stress and at finer grain size, and, conversely, dislocation creep is rate-limiting with increasing stress and at coarser grain size. The flow law predictions suffered from at least an order of magnitude overprediction error, suggesting that its application to glaciology is at present premature. This overprediction error is easily observed in Figure 4a’s creep rate contours (dashed grey lines); in particular, the range of densification rates for terrestrial ice sheets (1e−11 to 1e−12 ) correspond to creep rate contours that pass outside of the stress and grain size range for ice sheets (blue box). Secondly, flow was extrapolated directly to the temperature, stress, and density data resolved from Ebinuma and Maeno (1987). The disGBS mechanism rate-limits by approximately half an order of magnitude in the coarse-grained limit (Appendix Jupyter Notebook, Section 6). However, dislocation creep rates were a closer match to the natural densification rates, a contradictory result warranting future inquiry. The model’s compatibility with natural firn microstructures is a key advantage of our flow law relative to previous results. Firn’s inhomogeneous strain state is well known for developing localized stress concentrations and dynamic recrystallization processes (e.g., grain subdivision, migration recrystallization, and rotation recrystallization) (Faria et al., 2014). It is also important to note that disGBS and dislocation creep are both capable of developing the dislocation structures observed in firn columns. For example, these structures were observed within fine-grained solid ice deformed via disGBS in the laboratory (Goldsby and Kohlstedt, 1997a & 1997b). The disGBS mechanism may also 40 PENNSCIENCE JOURNAL | SPRING 2021

accommodate an impurity dependence for firn creep (Freitag et al., 2013, Breant et al., 2017), akin to the minor impurity dependence observed on ice creep among the Antarctic Meserve Glacier’s basal layers (Cuffey et al., 2000a). The flow law is also applicable to icy planetary bodies in the solar system. In these settings, cold temperatures limit grain growth, as well as the overall creep rate. For example, the icy moons of Jupiter exhibit near-surface temperatures of ~100 K, and rates of densification are further hindered by hydrate sulfate salts (Durham et al., 2001, Durham et al., 2010). Planetary ice sheets also contain nearsurface firn densification, such as among the dry ice-rich layered polar deposits at the Martian poles (Arthern et al., 2000, Cassanelli et al., 2015). The viscosity, shape, and age of the Martian caps all indicate the presence of H2O ice layers alongside layers of CO2 ice. Our experiments suggest that disGBS is the dominant mechanism for H2O firn creep on Mars (Figure 4b), at the average surface temperature of the polar ice caps, ~165 K (Nye, 2000). Clathrate hydrates, which are also present here, likely have lower creep rates (Durham et al., 2010) than those for H2O firn. Experimental errors may have included unwanted contributions from transient creep to steady-state creep (e.g., Durham et al., 2001) and the unconstrained confidence intervals on our sample’s grain size distributions. The latter was due to the delay of Scanning Electron Microscope imaging, which also resulted in the unconstrained confidence intervals for disGBS A and p flow law parameters. If we assume the ultra-fine-grained series experienced grain growth to an average radius of 8 ̇m, the flow law’s overprediction errors are drastically reduced and more closely match natural densification rates. Under this scenario, dislocation creep becomes rate-limiting across roughly half of the natural conditions, a picture more consistent with the field boundary hypothesis (DeBresser et al., 1998). The field boundary hypothesis posits that grain sizesensitive and grain size-insensitive creep mechanisms contribute in roughly equal proportion to the total rate in natural settings. As such, our multimechanism flow law should, in practice, predict the creep rate as a function of the grain size distribution at a given site, assuming detailed data on grain size distributions are known. Modeling errors may have included an overestimation of the applied stress, as the macroscopic values derived from the ice sheet’s density-stress profiles were likely born only by a

Research

Figure 4. (a) Mechanism map for intermediate stage densification (pr = 0.85) in the terrestrial cryosphere (233 K), plotted across grain size and stress, after Goldsby (2006). The dashed lines are creep rate contours (ṗ/pice )(s-1) and the blue line is the proposed mechanism regime change (b) Mechanism map for intermediate stage densification (pr = 0.85) in the Martian polar layered deposits (for H2O firn) (165 K), with its rate contours and regime change also plotted. Here, firn is expected to creep under lower stresses and grain sizes relative to (a) (Durham et al., 2010), rendering disGBS as predominately rate-limiting. neighboring pore channels (Wilkinson, 1988, Faria et al., 2014). In addition, the stress may decrease in response to variations in pore pressure, which were ignored herein due to the channel interconnectivity during the intermediate stage (Maeno and Ebinuma, 1983, Wilkinson, 1988). Lastly, disGBS may invoke more intricate physical considerations of firn’s microstructural state, such as limitations for disGBS for finer grains directly neighboring coarser grains.

5. Summary We developed a physics-based flow law for firn creep with parameters directly constrained from our laboratory data on ice powder densification. We identified two distinct mechanisms in the laboratory, with classical grain size-insensitive dislocation creep observed for the medium- and coarse-grained samples, and grain size-sensitive disGBS observed from fine and ultra-fine-grained samples. Upon extrapolation to natural settings, the flaw law predicts dislocation creep and disGBS as rate-limiting mechanisms for higher and lower stress conditions, respectively. Our two-mechanism flow law is similar in form to that for solid ice rheology, unifying the steady-state creep framework for ice sheets and glaciers from near-surface firn densification to ice flow at depth.

References

Adolph, A. C., Albert, M. R.: Gas diffusivity and permeability through the firn column at Summit, Greenland: measurements and Comparison to microstructural properties. The Cryosphere, 8, 319-328, https://doi.org/10.5194/tc-8-319-2014, 2014. Alley, R. B.: Polar Firn Densification and Grain Growth. Annals of Glaciology, 3, 7-11, https://doi.org/10.3189/ S0260305500002433, 1982. Alley, R.B.: Firn densification by grain boundary sliding: a first model, Journal de Physique, https://doi.org/10.1051/ jphyscol:1987135, 1987. Alley, R.B.: Transformations in Polar Firn. PhD Thesis U. of Wisconsin-Madison D. of Geology, p. 88, 1987. Arthern, R. J., Winebrenner, D. P., & Waddington, E. D.: Densification of Water Ice Deposits on the Residual North Polar Cap of Mars. Icarus, 133, 367-381, https://doi.org/10.1006/ icar.1999.6308, 2000. Baker I.: Microstructural characterization of snow, firn and ice. Phil. Trans. Royal Society A, A377, https://doi.org/10.1098/ rsta.2018.0162, 2019. Behn, M. D., Goldsby, D. L., and Hirth, G.: The role of grain-size evolution on the rheology of ice: Implications for reconciling laboratory creep data and the Glen flow law, The Cryosphere Discuss. [preprint], https://doi.org/10.5194/tc-2020295, in review, 2020. Breant, C., Martinerie, P., Orsi, A., Arnaud, L. & Landais, A.: Modelling firn thickness evolution during the last deglaciation: constraints on sensitivity to temperature and impurities, Climate of the Past, 13, 833–853, https://doi.org/10.5194/cp-13-833-2017, 2017. Burr, A., P. Lhuissiera, C. L. Martina & A. Philip.: In situ X-ray tomography densification of firn: The role of mechanics and diffusion processes. Acta Materialia, 167, 210-220, https://doi. org/10.1016/j.actamat.2019.01.053, 2019.

SPRING 2021 | PENNSCIENCE JOURNAL

Cassanelli, J. P, and Head, J. W.: Firn densification in a Late Noachian “icy highlands” Mars: Implications for ice sheet evolution and thermal response. Icarus, 253, 243-255, https:// doi.org/10.1016/j.icarus.2015.03.004, 2015 Coble, R.L.: A model for boundary diffusion controlled creep in polycrystalline materials, J. of Appl. Physics, 34, 1679-1682, https://doi.org/10.1063/1.1702656, 1963. Coble, R.L.: Diffusion Models for Hot Pressing with Surface Energy and Pressure Effects as Driving Forces, J. of Appl. Physics, 41, https://doi.org/10.1063/1.1658543, 1970. Cuffey, K., Conway, H., Gades, Anthony, Hallet, B., Raymond, C., & Whitlow, S.: Deformation properties of subfreezing glacier ice: Role of crystal size, chemical impurities, and rock particles inferred from in situ measurements, J. of Geophysical Research, 105, https://doi.org/10.1029/2000JB900271, 2000. De Bresser, J.H.P., C.J. Peach, J.P.J. Reijs & C.J. Spiers: On dynamic recrystallization during solid state flow: Effects of stress and temperature. Geophys. Res. Letters, 25, 3457–3460, 1998. Durham, W. B., L. A. Stern, & S. H. Kirb.: Rheology of ice I at low stress and elevated confining pressure. J. of Geophys. R., 106 (11), 11031-11042, https://doi.org/10.1029/2000JB900446, 2001. Durham, W.B., Ballesteros, O. & Goldsby, D.L.: Rheological and Thermal Properties of Icy Materials. Space Sci review. 153, 273-298, https://doi.org/10.1007/s11214-009-9619-1, 2010. Duval, P., Ashby, M.F., & Anderman, I.: Rate-controlling processes in the creep of polycrystalline ice. J. Phys. Chem. 87 (21), 4066–4074, https://doi.org/10.1021/j100244a014, 1983. Ebinuma, T. & Maeno, N.: Experimental Studies on densification and pressure sintering of ice. Annals of Glaciology, https://doi.org/10.3189/1985AoG6-1-83-86,1985. Ebinuma, T. & Maeno, N.: Particle Rearrangement and Dislocation Creep in a Snow-Densification Process. Journal de Physique Colloques, 48, 263-269, https://doi.org/10.1051/ jphyscol:1987137, 1987. Faria, H. F., Freitag, J., & Kipfstuhl, S.: Polar ice structure and the integrity of ice-core paleoclimate records. Quaternary Science Reviews, 29, 338-351, https://doi.org/10.1016/j. quascirev.2009.10.016, 2010. Faria, S. H., Weikusat, I., & Azuma, N.: The microstructure of polar ice. Part II: State of the art, J. of Structural Geology 61, 21-49, https://doi.org/10.1016/j.jsg.2013.11.003, 2014. Freitag, Kipfstuhl, Laepple, & Wilhelms: Impurity controlled densification: a new model for stratified polar firn. J of Glaciology. https://doi.org/10.3189/2013JoG13J042, 2013. Frost, H. J., & Ashby, M. F.: Deformation-mechanism maps: The plasticity and creep of metals and ceramics. Oxford: Pergamon Press. 1982. Glen, J.W.: Experiments on the deformation of ice, J. of Glaciology, 2, 111- 114, https://doi.org/10.3189/ S0022143000034067, 1952. Glen, J.W.: The creep of polycrystalline ice, Proc. R. Soc. London, 228, 519-538, https://doi.org/10.1098/rspa.1955.0066, 1955. Goldsby, D.L., & D.L. Kohlstedt: Grain boundary sliding in fine-grained Ice I. Scr. Mater. 37, 1399–1406, https://doi. org/10.1016/S1359-6462(97)00246-7, 1997a. Goldsby, D.L., & D.L. Kohlstedt: Flow of ice I by dislocation, grain boundary sliding, and diffusion processes, in Proceedings

42 PENNSCIENCE JOURNAL | SPRING 2021

of the 28th Annual Lunar and Planetary Science Conference (Lunar and Planetary Institute, Houston, 1997b), 429–430, 1997b. Goldsby, D.L. & D.L. Kohlstedt: Superplastic deformation of ice: Experimental observations, J. of Geophysical R., 106 (11), 17-30, https://doi.org/10.1029/2000JB900336, 2001. Goldsby, D.L.: Superplastic flow of ice relevant to glacier and ice-sheet mechanics, Glacier Science and Environmental Change, Blackwell Publishing, 2006. Herring, C.: Diffusional viscosity of a polycrystalline solid, J. Appl. Physics, 21,437-444, https://doi.org/10.1063/1.1699681, 1950. Herron, M. M. & C. C. Langway, Jr.: Firn densification: an empirical model. J. of Glaciology, 25 (93), 373 - 385, https://doi. org/10.3189/S0022143000015239, 1980. Keenan, E., Wever, N., Dattler, M., Lenaerts, J. T. M., Medley, B., Kuipers Munneke, P., and Reijmer, C.: Physics-based SNOWPACK model improves representation of near-surface Antarctic snow and firn density, The Cryosphere, 15, 1065– 1085, https://doi.org/10.5194/tc-15-1065-2021, 2021. Kipfstuhl, S., S. H. Faria, N. Azuma, J. Freitag, I. Hamann, P. Kaufmann, H. Miller, K. Weiler, & F. Wilhelms: Evidence of dynamic recrystallization in polar firn, J. Geophys. Res., 114, B05204, https://doi.org/10.1029/2008JB005583, 2009. Maeno, N. & Ebinuma, T.: Pressure sintering of ice and its implication to the densification of snow at Polar glaciers and ice sheets, J. Phys. Chem., 87 (41), 3–10, https://doi.org/10.1021/ j100244a023, 1983. Meussen, B. Mahrenholtz, & O. Oerter, H.: Creep of Polar Firn. Cold Regions Sci. and Tech. 29, 177-200, https://doi. org/10.1016/S0165-232X(99)00018-X, 1999. Nabarro, F.R.N., Deformation of crystals by the motion of single ions, in Report of a Conference on Strengths of Solids (Bristol), 75- 90, Phys. Soc., London, 1948. Nye, J. F.: A flow law model for the polar caps of Mars, J. of Glaciology, 46, 154, 438-444, https://doi. org/10.3189/172756500781833151, 2000. Qi, C., Stern, L. A., Pathare, A., Durham, W. B., & Goldsby, D. L.: Inhibition of Grain Boundary Sliding in Fine-Grained Ice by Intergranular Particles: Implications for Planetary Ice Masses, Geophys. R. Letters, 45, 12757–12765, https://doi. org/10.1029/2018GL080228, 2018. Schultz, T., Müller, R., Gross, D., and Humbert, A.: On the Contribution of Grain Boundary Sliding to Firn Densification – an Assessment using an Optimisation Approach, The Cryosphere Discuss. [preprint], https://doi.org/10.5194/tc-202175, in review, 2021. Spencer, M. K. & Alley, R. B.: Preliminary Firn Densification Model with 38-site Dataset. J. of Glaciology, 47 (159), 671-676, https://doi.org/10.3189/172756501781831765, 2001. Wilkinson, D. S. & M. F. Ashby.: Pressure sintering by power law creep. Acta Metall., 23, 1277-1286, https://doi. org/10.1016/0001-6160(75)90136-4, 1975. Wilkinson, D.S. A pressure-sintering model for the densification of polar firn and glacier ice. J. of Glaciology, 34 (116), https:// doi.org/10.3189/S0022143000009047, 1988. Yao, Z., Qin, K., Wang, Q., & Xue, S., Weak Ḃtype olivine fabric induced by fast compaction of crystal mush in a crustal magma reservoir. J. of Geophys. Research: Solid Earth, 124,

3530–3556, https://doi.org/10.1029/2018JB016728, 2019.

Interested in joining PennScience? We Want You! Contact www.pennscience.org PennScience is sponsored by the Science and Technology Wing at the University of Pennsylvania