Major: Neuroscience Research Department: Bioinformatics Graduation Date: May 2020
Abstract: Parkinson’s disease (PD) is the second most common neurological disorder and is characterized by motor dysfunctions such as resting tremor and gait impairments. PD has no known cure, and a major impediment is the lack of a clinically adopted method to estimate the rate of disease progression. Here, we investigated baseline gait, postural stability, and clinical measures from 160 PD subjects as independent variables to predict disease progression as measured by the MDS-UPDRS Part III score at two years. Towards this, we developed statistical machine learning models to evaluate the relative predictive value of these different measures. Specifically, we used the XGBoost and Feed Forward Neural Network models and performed extensive feature selection, model optimization, and cross-validation. The best model performance was obtained using just the clinical measures in predicting the 2 year UPDRS Part III score with the XGBoost model. This model was able to explain 22% of the variance in the target and achieve a Root Mean Squared Error of 6.7 points, which is close to the minimal clinically significant difference of 5 points. We discuss what this best model learned, why the gait and postural stability measures did not have predictive value, and future directions.
What does research mean to you? Research is the application of an essentially playful and creative impulse towards confronting the mysteries of the world we live in. While it is certainly a useful activity in that it addresses important problems affecting our lives, the original impulse is one governed by the desire to craft beautiful ways of answering questions. While much emphasis is put on the knowledge research produces (and these final results comprise most of what we learn in school), the creativity and playfulness is in the method. Research, then, is the art of the method. Inspiration, experience, trial and error, and dumb luck guide both the artist and the scientist in their craft. Experiments, like works of art, can be subject to critical appraisal. A particular experiment is like a poem, and we can ask if it satisfies our aesthetic criteria: Is this design sound? Does it make sense? Is it simple and elegant? Does it get me closer to the heart of the matter?
Tell us about your journey. I’ve been fortunate enough to be able to pursue a wide range of research interests including neuroscience, biomedical engineering, and computer science since graduating high school. My very first research experience was during the summer before freshman year, when I worked full-time in Dr. Aruna Vanikar’s clinical pathology lab at the Institute of Kidney Diseases and Research Center (IKDRC) in Ahmedabad, India. Here, I worked on analyzing kidney stones using IR spectroscopy. I worked closely with Dr. Vanikar and lab technician Mr. Jignesh Dave towards analyzing a series of kidney stone patients’ data and composing a report of the results. I am grateful for both mentors’ guidance during my first encounter with data analysis and scientific writing. In freshman year I joined Dr. Mandy Maguire’s Developmental Neurolinguistics Lab and fell in love with research and the constant learning it calls for. The lab uses tools like EEG to investigate how children learn language, and I was fascinated by the whole process, from experimental design to data analysis, that could extract meaningful information about cognitive processes from the brain’s electrical activity. Dr. Maguire and Dr. Julie Schneider (a recent PhD graduate from the lab) have both been tremendous mentors who directed my thirst for learning towards meaningful projects. Last Spring I had the amazing opportunity to lead my own EEG study investigating how adults process different kinds of word relationships and I am immensely grateful for both Drs. Maguire and Schneider for supporting me in this work. In sophomore year, I joined Dr. Taylor Ware’s lab that investigates a class of materials called Liquid Crystal Elastomers towards biomedical applications. I had taken a few biomedical engineering courses up to that point and, having really enjoyed the sustained engagement with mathematics and physics, wanted to pursue the subject as deeply as I could. Joining Dr. Ware’s lab allowed me to do that by learning about a variety of advanced techniques for material characterization. Dr. Ware and grad student Cedric Ambulo were also great mentors who helped me develop as a scientist. Having this very different kind of research experience taught me to deal with and even enjoy the iterative, failure-filled process of hands-on engineering. Finally, I joined Dr. Albert Montillo’s lab at UTSW as part of the Green Fellowship (which meant full-time research). The lab is focused on developing the theory and application of machine learning towards analyzing biomedical (esp. neuroimaging) datasets. For my project, I analyzed a dataset of Parkinson’s disease patients to investigate gait and postural stability measures towards predicting disease progression. This has by far been my most intensive research experience, and I have been incredibly lucky to have mentors like Dr. Montillo and MSTP student Kevin Nguyen to guide me through it. I have learned so much during my time here, from programming in python and linux to developing and testing machine learning models and processing neuroimaging data. Throughout this, Dr. Montillo has treated me like a mature graduate student and I have been surrounded by excellent student researchers in the lab which really pulled me up and encouraged me to approach another level of ability. This has also been my closest encounter with truly translational research (distinct from the clinical/medical research at IKDRC), and it has given me a much clearer vision of the kind of work I want to pursue in the future.
Advice for Future Green Fellows
Read a lot, ask a lot of questions, and don’t get lost in the daily minutiae of your work. Try to have a consistent (perhaps weekly) moment of reflection on what you did, why you did it, and how it relates to the big picture.