A new artificial intelligence model can predict human brain activity across vision, hearing, and language more accurately than ever before. Researchers from Meta AI and academic institutions developed TRIBE v2, a tri-modal foundation model trained on over 1,000 hours of fMRI data from 720 subjects. When tested on novel stimuli and tasks, it outperformed traditional linear encoding models by several-fold, offering a unified framework for cognitive neuroscience.
The Research
Led by Stéphane d'Ascoli and colleagues, the study (arXiv:2605.04326) introduces TRIBE v2, a model that processes video, audio, and text simultaneously. The team compiled a dataset of over 1,000 hours of high-resolution fMRI recordings from 720 participants as they watched movies, listened to sounds, and read text. The model was trained to predict brain activity from these stimuli. In validation experiments, TRIBE v2 accurately predicted brain responses to entirely new stimuli and tasks — not just those it was trained on. It also replicated classic findings from decades of visual and neuro-linguistic research, demonstrating that it captures fundamental cognitive processes. By interpreting the model's internal representations, the researchers mapped fine-grained multisensory integration in the brain, showing where and how different senses combine.
Why It Matters
This research matters because it bridges AI and neuroscience, enabling in silico experiments — testing theories on a computer model instead of in the lab. For anyone interested in their own cognition, this suggests that AI can help decode how we process complex real-world stimuli. It also raises the possibility of personalized brain models that could predict how an individual's brain responds to different types of information, potentially guiding educational or training approaches.
What You Can Do
While you can't run TRIBE v2 at home, you can explore how your own brain processes multimodal information by taking cognitive tests that measure memory, attention, and reasoning across different domains. Brain training exercises that combine visual, auditory, and language tasks may help improve your cognitive flexibility. For a start, try our free adaptive IQ test to gauge your general cognitive abilities, then explore evidence-based training modules.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.
