The persistence of global form (Part I): Stimulus inversion influences V1 fMRI activity

Abstract

The human visual system uses feedforward and feedback mechanisms to achieve and maintain the figure-ground segregation of global form. One way to study figure-ground segregation is to measure the visual persistence of global form, a type of short-term perceptual memory. Studies of persistence have shown that figure-ground segregation is maintained by the visual system for up to a few seconds following the removal of perceptual binding cues that originally led to segregation. Previous studies showed persistence-related fMRI activity in human lateral occipital cortex, but failed to show persistence-related fMRI activity in V1. If persistence were observed in V1 it would suggest that feedback from high-level visual areas influence the participation of V1 in the persistence of global form. We presented observers with outlines of faces and animals that were camouflaged until they moved; these figures were either upright or inverted. We found that both upright and inverted figures persisted for up to several seconds (measured behaviorally) after they stopped moving, but that upright figures persisted longer than inverted versions of the same stimuli. We measured persistence-related fMRI activity in several occipito-temporal regions of interest (ROIs): object-selective lateral occipital cortex (LO); face-selective occipital cortex (OFA) and the fusiform face area (FFA); and several early visual ROIs (V1, V2, V3, V4v). We observed an effect of inversion on persistence in all of our ROIs, including V1. Our findings suggest that V1 fMRI activity is influenced by the orientation of an object, even though upright and inverted versions of the same figure contain equivalent low-level stimulus information. We propose that LO and other higher-tier visual areas—including the OFA and the FFA—are part of a recurrent processing network, and that these areas facilitate visual processing in V1 via feedback. We further investigate this possibility in Part II.