ihybrid
Figure3
Reconstructed faces (a) and results (b) from the posttest validation analysis. Using the difference maps from the iHybrid analyses (a), we isolated the regions associated with local-identification responses (colored in red) and the regions associated with global-identification responses (colored in blue). In the example shown here, the iHybrid stimulus has local information from Brad Pitt and global information from William H. Macy. One hybrid was created by applying a spatial-frequency (SF) filter to select the local information for Pitt and completing the image with the complementary information from Macy. Another hybrid was created by using an SF filter to select the global information for Macy and filling in the image with the complementary information from Pitt. Thus, two hybrids were created for each iHybrid stimulus by applying local and global SF filters separately to the iHybrid The graph (b) shows the proportion of identifications that corresponded with the information selected by the SF filter (predicted identity) and the proportion of face identifications that corresponded with the person from whom the complementary information was derived (other identity). Error bars show standard errors of the mean.
Figure3
Reconstructed faces (a) and results (b) from the posttest validation analysis. Using the difference maps from the iHybrid analyses (a), we isolated the regions associated with local-identification responses (colored in red) and the regions associated with global-identification responses (colored in blue). In the example shown here, the iHybrid stimulus has local information from Brad Pitt and global information from William H. Macy. One hybrid was created by applying a spatial-frequency (SF) filter to select the local information for Pitt and completing the image with the complementary information from Macy. Another hybrid was created by using an SF filter to select the global information for Macy and filling in the image with the complementary information from Pitt. Thus, two hybrids were created for each iHybrid stimulus by applying local and global SF filters separately to the iHybrid The graph (b) shows the proportion of identifications that corresponded with the information selected by the SF filter (predicted identity) and the proportion of face identifications that corresponded with the person from whom the complementary information was derived (other identity). Error bars show standard errors of the mean.