Researchers have announced TRIBE v2, a next-generation predictive foundation model designed to act as a digital twin of human neural activity. The model can forecast how the brain responds to various sights, sounds, and language inputs with unprecedented speed and accuracy, offering a seventyfold increase in resolution compared to similar systems. TRIBE v2 was trained on data from more than 700 healthy volunteers exposed to diverse media, including images, podcasts, videos, and text. It enables zero-shot predictions for new subjects, languages, and tasks, outperforming standard modeling approaches.

The developers have released the model, codebase, research paper, and an interactive demo to support neuroscientists and clinical researchers. By simulating brain activity computationally, TRIBE v2 allows testing of hypotheses without requiring human subjects for every experiment. The release aims to accelerate neuroscience discovery, improve AI systems through brain-inspired insights, and advance treatments for neurological disorders.

All materials are shared under a CC BY-NC license, with the goal of fostering open research and enabling scientific and clinical breakthroughs for the broader community.