Abstract
Object substitution masking (OSM) is a powerful tool for limiting visual awareness. Behavioral work demonstrates that OSM disrupts, but does not eliminate basic representation of object size. However, OSM does appear to wipe out face- and shape-specific brain activity, as measured by EEG and fMRI. This suggests that OSM is particularly strong late in visual processing, at or just prior to high-level face representation. We sought to test this hypothesis behaviorally, particularly because faces sometimes exert surprising perceptual effects when suppressed from awareness by other types of masking. We leveraged an effect based on the poor spatial resolution of ventral-visual neurons to evaluate how strongly OSM disrupts face processing. When two emotional faces fall within a visual hemifield and selective attention is not deployed, people tend to see the average of those two faces. This perceptual averaging does not occur when two faces appear across visual hemifields, consistent with the receptive field sizes of high-level ventral visual neurons. If OSM disrupts face representation in addition to disrupting awareness, then this averaging process should be reduced when one face from a pair is suppressed by OSM. Observers viewed one neutral and one emotional face (happy/angry) for 20-ms, arranged either vertically to the left/right of fixation (within-hemifield) or horizontally above/below fixation (across-hemifields). On some trials, a trailing four-dot mask limited visual awareness of one face from each pair. On all trials, a subsequently presented arrow prompted observers to indicate how positive or negative one of the faces appeared to be, and to indicate how many faces they perceived. Without OSM, only faces within a visual hemifield were perceptually averaged, replicating previous work. This averaging did not occur when awareness of one face was suppressed by OSM. Behavioral and physiological evidence thus agree—OSM is particularly strong in late stages of visual processing.
Meeting abstract presented at VSS 2016