Distinct neural systems for the encoding and recognition of topography andfaces

In a series of three positron emission tomography experiments the functiona l neuroanatomy of four different types of visual stimuli was investigated w ithin the same experimental context. The stimuli were unknown buildings, la ndscapes, human faces, and animal faces. The purpose of the present study w as to compare the stimulus types, both within the same category and across category, by examining if, at encoding (with several seconds exposure to ea ch stimulus) or recognition lover time scales of minutes compared to the se conds of usual perception/one-back studies), common or different neural cir cuits were activated for all types/categories of stimuli. Within category a nd although visually very different, the encoding of both buildings and lan dscapes activated a similar set of brain regions, including bilateral parah ippocampal gyrus. This was in contrast to the encoding of both human and an imal faces, both of which resulted in activation of the fusiform gyrus bila terally. Despite the perceptual inputs being identical to those during enco ding, the recognition of both buildings and landscapes activated only unila teral right parahippocampal gyrus, while recognition of both human and anim al faces activated unilateral right fusiform gyrus. In addition, right supe rior frontal gyrus and right inferior and medial parietal areas were more a ctive during recognition compared with encoding for all stimulus types. Ove rall the data identify differential patterns of activation for encoding com pared with retrieval of visual stimuli. Furthermore, medial temporal struct ures specifically are involved in the explicit learning and long-term recog nition of topographically relevant stimuli, be they buildings or landscapes , while lateral temporal structures support non-topographical learning and recognition, in this case either human or animal faces. (C) 2001 Academic P ress.