Gennady Erlikhman, Ph.D. - US grants

? DESCRIPTION (provided by applicant): The neural correlates of object perception have been traditionally identified in ventral cortical regions that form a visual processing hierarchy, with deeper regions containing cells with larger receptive fields that are thought to underlie abstract (viewpoint-, position-, and scale-invariant) object representations. These forms of representational invariance are important for the categorization and identification of static objects. However, most objects in our natural environment are not fully visible due to occlusion and are not static as they or as we move. Rather than representational invariance, the visual system requires precise representations about object fragment position, velocity, and orientation in order to integrate many fragments across space and time into perceptual wholes or spatiotemporal objects. We have conducted two preliminary studies that suggest that dorsal regions (V3A, V3B, and the intraparietal sulcus) may be involved in the representation of dynamic forms. This opens the question of whether shape information in dorsal and ventral areas is similar or redundant or whether there is communication and a relaying of information between them that ultimately leads to both invariant representations in ventral areas and specific representations in dorsal areas. To identify the signaling pathway (Aim 1), subjects will perform a shape identification task with objects that are gradually revealed over time (spatiotemporal objects) or with objects whose shape information is available all at once while EEG data is recorded. In order to relate the signaling time-course to cortical areas, structural and functional MRI data will be collected in order to enhance source-localization accuracy and to define ROIs in source-localized space. Using MVPA, we will decode shape identity in each ROI as a function of time. If spatiotemporal object representations are first constructed in dorsal areas and then passed down to ventral ones, shape decoding will be possible earlier after stimulus onset in dorsal areas than in ventral. Static object identity may only be decodable in ventral regions. To confirm the causal role of dorsal areas in spatiotemporal object representation (Aim 2), we will conduct two experiments in which we apply transcranial direct current stimulation (tDCS) to posterior parietal cortex (dorsal visual areas) to disrupt dynamic form integration. If dorsal areas are necessary for integration of shape information over time, tDCS should decrease the time over which such forms can be integrated, but should have no effect on the perception of static shapes.