Despite recent advances in brain imaging, the brain’s representation of the visual world is still largely unknown. Using a new artificial intelligence technique – deep learning – to decode brain activity recorded in functional imaging, CerCo’s Rufin VanRullen and Leila Reddy can reconstruct images seen – or imagined – by subjects. This method, published in Nature Communications Biology, opens the door to new explorations of the brain.

Reconstructing faces from fMRI patterns using deep generative neural networks Rufin VanRullen & Leila Reddy Nature Communications Biology