MEGA — reading memory from the eyes

A no-report paradigm that quantifies episodic memory from anticipatory eye gaze, published in Nature Communications Psychology.

MEGA (Memory Episode Gaze Anticipation) is the research program behind my MSc at Tel Aviv University’s Yuval Nir Lab: a method to measure whether someone remembers an event without asking them — by tracking where their eyes go.

The idea

When people re-watch a movie clip, their gaze starts drifting toward the location of a salient event seconds before it appears — but only if they remember it. That anticipatory gaze is a readable memory trace:

“Even when subjects said they didn’t remember, their gaze direction showed they did. This means that sometimes people remember, but can’t say that they remember.”

The machine-learning core

The paradigm is as much an ML problem as a neuroscience one: raw eye-tracking streams go through signal processing and feature extraction, and a classifier identifies — from single-trial gaze features — whether a given viewing reflects memory for the event.

Why it matters

Memory assessment today mostly requires verbal reports. MEGA opens the door to measuring memory in populations that can’t provide them — patients with brain injuries, pre-verbal infants, even animals — with applications from early dementia screening to basic memory research.

Outputs