This summer I participated in the Pratt Research Fellows program under Dr. Timothy Dunn and PhD student Kyungdo Kim. My project explored the application of an autoregressive hidden Markov model on behavioral data from Parkinson’s patients. These patients go through various behavioral tests such as walking or finger-tapping since the disease partially manifests as motor disorders, and I would analyze this data to identify clinically meaningful behavioral markers that can classify healthy and unhealthy subjects. This work is significant because diagnostics for Parkinson’s are heavily biased towards the clinician administering the evaluation and is primarily qualitative. With behavioral quantification and human behavior analysis (HBA) gaining traction, this work is also the first to apply such a model on human data, and provides a framework or baseline for whether it is effective. I also took Educational Psychology over the summer and traveled to Orlando for the 2024 IEEE EMBC conference, and Penn State for the Mathematical Biosciences workshop.
It was a blessing to work on a project with such a clear clinical goal and impact, and I am grateful to make enough progress to complete the project at the end of the summer. I got to see how all the data was recorded through a synced multi-camera setup, how the cameras were calibrated, and how it’s an extension on previous recordings done on mice. At the conferences, I got to meet my PI in person for the first time in three years, and noted how differently engineers vs mathematicians view digital twins and the computational biology world. In the future, I plan to immerse myself further in computational and theoretical neuroscience, with clinical applications in neurodegenerative diseases, in a PhD program.
