fairfolio
Saved Projects

NeuroTrace:A Deep Learning Framework for Monitoring Multiple Sclerosis Progression from Oculomotor and Speech Biomarkers

CSEF · 2026 Medicine & Physiology (Track 2) (Senior Division)

Medha Choudhry

Overview

Multiple Sclerosis (MS) is a chronic autoimmune neurological disorder affecting 2.9 million people worldwide. For MS, an incurable disease, the prevention of progression (from RRMS to PPMS) is crucial to preserve the quality of life. Patients with MS are subject to different disease courses and require individualized treatment; clinicians need to respond quickly to changes in therapy plans and adjust them accordingly. Current monitoring tools include clinical assessments, such as the Expanded Disability Status Scale (EDSS) and MRI scans, which are expensive, time-consuming, and poorly suited to capturing daily neurological changes. This project introduces NeuroTrace, a multimodal AI framework for monitoring MS progression through two digital biomarkers: speech and eye movements. For both modalities, NeuroTrace first classifies the disorder type, then tracks its severity over time. For the audio pipeline, waveforms from the TORGO and UA-Speech datasets (29,313 samples from 43 speakers) were converted into mel-spectrograms and analyzed using a SmallCNN + BiGRU model with leave-one-speaker-out cross-validation. This binary dysarthria classifier achieved an accuracy of 90.4% and an 89.8% weighted F1 score. A Ridge Regression model predicted speech intelligibility with a 12.7% mean absolute error and a Spearman correlation of 0.887, providing a method for tracking the worsening or improvement of one’s dysarthria. For the vision pipeline, 840 clinical eye exam videos from the University of Utah were segmented using two multimodal LLMs. A DenseNet-121 + BiGRU architecture achieved a Weighted F1 score of 80.9% and accuracy of 80.2%. In parallel, the videos were sent to ConVNG, which returned pupil traces that were used to extract oculomotor features, including fixation dispersion and drift velocity. Together, these results support the feasibility of a low-cost, accessible system for longitudinal monitoring of MS progression.

Competition history

  • CSEF 2026 Medicine & Physiology (Track 2) (Senior Division) · Entry S-21-06

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: California Science & Engineering Fair public projects

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google