Dissociations in the processing of "what" and "where" information in working memory: an event-related potential analysis.

Abstract Based on recent research that suggests that the processing of spatial and object information in the primate brain involves functionally and anatomically different systems, we examined whether the encoding and retention of object and spatial information in working memory are associated with different ERP components. In a study-test procedure subjects were asked to either remember simple geometric objects presented in a 4 by 4 spatial matrix irrespective of their position (object memory task) or to remember spatial positions of the objects irrespective of their forms (spatial memory task). The EEG was recorded from 13 electrodes during the study phase and the test phase. Recognition performance (reaction time and accuracy) was not different for the two memory tasks. PCA analyses suggest that the same four ERP components are evoked in the study phase by both tasks, which could be identified as N100, P200, P300, and slow wave. ERPs started to differ as a function of memory task 225 msec after stimulus onset at the posterior recording sites: An occipital maximal P200 component, lateralized to the right posterior temporal recording site, was observed for the object memory but not for the spatial memory task. Between-tasks differences were also obtained for P300 scalp distribution. Moreover, ERPs evoked by objects that were remembered later were more positive than ERPs to objects that were not remembered, starting at 400 msec postsimulus. The PCA analysis suggest that P300 and a slow wave following P300 at the frontal recordings contribute to these differences. A similar differential effect was not found between positions remembered or not remembered later. Post hoc analyses revealed that the absence of such effects in the spatial memory task could be due to less elaborated mnemonic strategies used in the spatial task compared to the object memory task. In the face of two additional behavioral experiments showing that subjects exclusively encode object features in the object memory task and spatial stimulus features in the spatial memory task, the present data provide evidence that encoding and rehearsal of object and spatial information in working memory are subserved by functionally and anatomically different subsystems.