#fMRI03/09/2025
How DINOv3 Reveals Brainlike Visual Representations in Space and Time
'Study finds DINOv3 models trained on billions of natural images align with human cortical responses across space and time, driven by model size, training amount, and image type'