TRIBE v2 (Temporal Response In Brain Encoding)

Meta AI's model that predicts fMRI cortical responses to video stimuli, used as the engine behind VidCognition.

TRIBE v2 (Temporal Response In Brain Encoding, version 2) is an AI model developed by Meta AI Research that predicts how the human visual cortex responds to video stimuli — second by second — using fMRI brain scan data as its training signal.

TRIBE v2 was trained on 7T fMRI data from human subjects watching natural video content. It learned to predict the cortical activity pattern — which brain regions activate and at what intensity — for any new video input. The model achieves approximately 92% correlation with actual fMRI measurements.

The practical implication: TRIBE v2 can predict brain engagement for any video without requiring a human in an MRI scanner. This makes enterprise-grade neuromarketing insights accessible outside of laboratory settings.

VidCognition is built on TRIBE v2, making it the first creator-accessible tool to use this model for short-form video optimization.

Try it with VidCognition