Back to projects

Rivalry · Autism · EEG · Computational neuroscience

Slower binocular rivalry in the autistic brain

Co-authored EEG study showing slower neural and behavioral rivalry dynamics in autism, with neural signals predicting symptom severity and classifying diagnostic status.

Methods

Frequency-tagged rivalry · EEG / SSVEP

Classification

86.5% LOO accuracy

Venue

Current Biology

Overview

This project used frequency-tagged binocular rivalry and EEG/SSVEP to measure rivalry dynamics directly from visual cortex, without relying only on behavioral reports.

The study showed that binocular rivalry was slower in autism both behaviorally and neurally. We derived a Neural Rivalry Index from occipital EEG signals, found that it strongly predicted participants’ reported switch rates, and showed that this neural marker also predicted autism symptom severity. Using neural features alone, we could classify diagnostic status with 86.5% accuracy in a leave-one-out analysis.

Why it matters

This paper provided a non-verbal neural measure of a perceptual marker linked to autism, helping connect low-level visual dynamics to clinically meaningful variation. It also showed how carefully designed neural signals can be turned into interpretable individual-differences measures.

My role

Co-author on the investigation team, with explicit contributions to visualization and experimental work.