Computer vision analysis for digital diagnostics of neurological conditions from in-clinic and home videos (BRAINWALK project)
Peter Washington, Professor
UC San Francisco
Closed. This professor is continuing with Fall 2025 apprentices on this project; no new apprentices needed for Spring 2026.
Come work with a collaboration between the the UCSF TECH Lab (PI: Peter Washington, PhD, UCSF) and the UCSF Bove Lab (PI: Riley Bove, MD, UCSF) to create computer vision-based digital diagnostics for neurological conditions through the Bove lab's BRAINWALK project, with direct mentorship on the clinical, scientific, and translational components by Dr. Bove and on the computer vision / technical components by Dr. Washington (faculty in the Berkeley-UCSF Computational Precision Health and UCSF Biological and Medical informatics PhD programs).
We will host two specific sub-projects:
(1) Classifying symptoms related to cognition and gait wellness as well as directly classifying disease status from videos recorded in-lab using multimodal deep learning. This project would lead to a publication arising from the student's work.
(2) Create a data processing pipeline for organizing and preprocessing all of the data collected from this study. This project would lead to co-authorship on several publications that end up using your pipeline.
Successful students will have a chance to continue this collaboration beyond the URAP period / enroll in URAP next year as well.
You will need to complete UCSF CITI and HIPAA training rapidly during onboarding so that you can be added to the IRB.
See these papers for details of past projects using these data from the Bove lab:
https://pmc.ncbi.nlm.nih.gov/articles/PMC11789430/
https://www.sciencedirect.com/science/article/pii/S2211034824000956?via%3Dihub
Role: The students will with with a large team to develop computer vision models to predict multiple sclerosis (MS), Parkinson's disease (PD), Alzheimer's, and other neurological conditions from in-clinic and home videos.
Attend two meetings each month: (1) BRAINWALK-wide group meetings and (2) 1-1 meeting with Dr. Washington for direct technical mentorship.
Qualifications: applied machine/deep learning; familiarity with computer vision libraries; familiarity with processing video data in, for example, Python
Hours: to be negotiated
Related website: https://bovelab.ucsf.edu/digital-medicine
Related website: https://techlab.ucsf.edu/