Abstract
Despite decades of investigation of prosopagnosia, the nature of the perceptual deficit underlying this condition has never been rigorously assessed. Nor even has it been established if there is a perceptual deficit, as opposed to a deficit in memory or a failure of view invariance. Individuals classified as prosopagnosic do have difficulty in discriminating some non-face visual entities. Performance on a minimal simultaneous match-to-sample test (Fig. 1) in which the participant must select the face that is an exact match to the sample from a foil: a) firmly establishes that there is a perceptual component to prosopagnosia, b) correlates with the magnitude of face configural effects, and c) accounts for much of the predictable variance of other validated measures of face recognition abilities, such as the CFMT, Faceblind.org, or the PI20. These findings suggest that the scores yielded by the various measures, although not ostensibly measuring perceptual processing, are largely—if not solely—reflecting perceptual effects. Individuals who show poor performance on the match-to-sample test also show deficits in their discrimination of fine metric differences between non-face stimuli, but only when these stimuli are complex, such as teeth, where the location of the differences are uncertain. The underlying deficit in prosopagnosia can arise from a dearth of large, overlapping receptive fields in posterior face-selective areas, e.g., FFA. In typical subjects, the overlap leads to activation of r.f.s with centers far removed from the features, thus serving to magnify the perceptual impact of small metric differences. With small r.f.s, the lack of overlap in the r.f.s in face-selective areas may explain why prosopagnosics do not show configural effects (Fig. 2) and why their deficit in distinguishing fine metric differences among non-face objects is apparent only when the objects are complex (Fig. 1) but not when discriminating simple geons (Fig. 3).
Meeting abstract presented at VSS 2018