Thesis Working Title: Artificial Intelligence motion capture for the evaluation and monitoring of Parkinson’s disease

Parkinson’s disease (PD) is the fastest growing neurological disorder in the world. The symptoms are diverse and debilitating. The current gold-standard for assessment and monitoring in PD is expert clinician rating with an assessment tool known as the MDS-UPDRS. My research seeks to demonstrate that artificial intelligence (AI) motion capture, through deep learning, performs as well as, or better than, conventional expert clinical assessment. A fundamental motivation for this research is a need to expand the scope of research and clinical access to neurological assessment. As such, the network we are using (DeepLabCut) is a free and open-sourced software, with video captured on an everyday smartphone. Upon validation of the performance of our trained network, our software will be used to quantify movement aberrations in PD, and document additional movement phenomena. We have a particular interest in a motor overflow phenomenon know as synkinesis, in which the large effort required for movement in PD causes signal “overflow” into inappropriate brain regions, resulting in involuntary movements.