Unsupervised Feature Learning Improves Prediction of Human Brain Activity in Response to Natural Images
The parameter tuning varied across the voxels and had a bias for high spatial frequencies and oblique orientations. (A) Two-dimensional Gaussian functions that were fit to the responses of three representative voxels to point stimuli at different locations. (B) Responses of three representative voxels to sine-wave gratings that spanned a range of orientations and spatial frequencies. (C) Mean responses across the voxels to sine-wave gratings that spanned a range of orientations and spatial frequencies.