Preparatory patterns of neural activity predict visual category search speed.

Rapidly detecting target object categories when objects are embedded in naturalistic scenes is facilitated by preparatory baseline signal changes. However, it is unclear as to what information most strongly predicts perceptual speed in terms of the minimal exposure duration required for accurate detection. Using novel surface-based spatiotemporal pattern classification, we found that while category-specific biases resulting from merely providing a category name can be detected in multiple cortical areas, only biases in lateral occipital complex predicted perceptual speed. These biases likely carry visual semantic information regarding multiple object categories placed in familiar scene contexts. Discriminatory voxels during the preparatory period showed congruent category-selectivity during visual stimulation.