Early Visual Areas are Activated during Object Recognition in Emerging Images

Abstract

Human observers can reliably segment visual input and recognise objects. However, the underlying processes happen so quickly that they normally cannot be captured with fMRI. We used Emerging Images (EI), which contains a hidden object and extends the process of recognition, to investigate the involvement of early visual areas (V1, V2 and V3) and lateral occipital complex (LOC) in object recognition. The early visual areas were located with a retinotopy scan and the LOC with a localiser. The participants (N=8) then viewed an EI, followed by the hidden object’s silhouette (disambiguation), and then, the EI was repeated. BOLD responses before and after disambiguation were compared. The retinotopy parameters were used to back-project the BOLD response onto the visual field, creating spatially detailed maps of the activity change. V1 and V2 (but not V3) showed stronger response after disambiguation, while there was no difference in the LOC. The back-projections revealed no distinct pattern or changes in activity on object location, indicating that the activity in V1 and V2 is not specific for voxels corresponding to the object location. We found no difference before and after disambiguation in the LOC, which may be repetition suppression counteracting the effect of recognition.