Figure-ground signals in early and object specific visual areas: A combined fMRI, EEG and TMS study

Abstract

Two processes can be discriminated when distinguishing a figure from its background: boundary detection and surface segregation. The neural origin and temporal dynamics of these two processes is still much disputed. In monkey V1 neural correlates of surface boundary detection have been found, but there is still debate about whether scene segmentation signals can be found in human V1. Here we studied the neural correlates of these two processes using texture and motion defined stimuli that differentiate between surface segregation and edge detection. By this means we were able to generalize figure-ground segregation results across different types of stimuli. In several ROI's (V1, V2, LO and V5), BOLD-MRI was measured (n=13) to establish whether figure-ground signals are cue independent within these regions. Preliminary data analysis shows that texture defined figure-ground segregation was associated with an increase in BOLD responses in V1, V2 and LO. Interestingly motion defined figure-ground segregation signals were only found in LO. These results suggest that where V5 primarily responds to the actual presence of motion, LO can keep an object active even though the actual motion has stopped. Currently we combine rTMS, EEG and the above-described paradigm to investigate the temporal and spatial dynamics of V1, LO and V5 in figure-ground segregation. The first results indicate that rTMS disturbs figure-ground related processes in V1 when LO and V5 are stimulated.