Holistic Processing Supports Familiarity-Based Associative Recognition for Faces
23.4091, Saturday, 16-May, 8:30 am - 12:30 pm, Pavilion
Mitchell Meltzer1, Gowtham Ganesan1, Michelle Min1, James Bartlett1; 1The University of Texas at Dallas
Associative recognition memory is measured by assessing discrimination between copies of studied items (intact items) and re-combinations of parts of studied items (conjunction items). Such intact/conjunction (I/C) discrimination has often been thought to require recollection, a process supporting the retrieval of qualitative information about a previous encounter. However, recent findings indicate that recognition of associations encoded in a holistic or unitized fashion can rely on familiarity, a mnemonic signal varying only in strength. To test this possibility, we applied a modified process dissociation procedure to an associative recognition paradigm for four classes of stimuli; upright faces, which are thought to be processed in a unitized or holistic fashion, inverted and misaligned faces, both thought to be processed in a part-based fashion, and faces which were both inverted and misaligned. Participants studied a list of faces, and were presented with intact, conjunction, and new faces at test. I/C discrimination was then compared for “exclusion” participants, for whom such discrimination was encouraged, and “inclusion” participants, for whom it was discouraged. According to process dissociation logic, familiarity-based associative recognition is inferred if exclusion and inclusion participants exhibit similar levels of I/C discrimination. By contrast, greater I/C discrimination in the exclusion condition as compared to the inclusion condition with no corresponding group difference in intact/new discrimination indicates recollection-based associative recognition. Consistent with previous findings, a negligible I/C discrimination advantage for exclusion versus inclusion participants was found with upright faces. Howevever, a robust I/C discrimination advantage for exclusion participants over inclusion participants was found with inverted-misaligned faces, consistent with the idea that such stimuli were processed in a part-based fashion. Broadly speaking, our results suggest that holistic processing of faces supports the use of familiarity in associative recognition, and that such processing is susceptible to a combination of face inversion and face misalignment.