Abstract
Developmental prosopagnosics (DPs) have otherwise normal vision and generally show normal recognition of non-face visual entities, such as objects and scenes. They do suffer, in comparison to controls, on a wide range of standard face recognition tests that define DP. As simple as these tests appear at first perusal, none make it clear just what it is that a DP cannot do visually. Moreover, as these tests typically include some aspect of memory or transformation, it is unclear if there is even a purely visual deficit. Method. To assess whether a perceptual deficit underlies DP, subjects performed simple simultaneous match-to-sample tests of triangular arrays of three faces, three blobs with face texture, or three geons (Fig. 1). The upper stimulus was an exact match to one of the two lower stimuli, with the foil differing metrically from the match. The similarity of the differences between foils and matching stimuli was scaled to be equal across the different stimulus classes, according to a model of V1 simple cell similarity. Matching and foil faces differed in the height of the cheekbones and the vertical distance between eyes, nose, and mouth. These variations produced differences between the faces that could not be readily articulated by the subjects. The diagonal relation between sample and matching stimuli defeated a simple horizontal or vertical alignment strategy in which differences in local pixel clusters or contour features could be employed to detect the distractor. Results and Conclusion. DPs performed worse than controls on both the face and blob tasks but not the geon task. Moreover, accuracy (percent correct) on the face task was significantly correlated with performance on a variety of standardized tests for prosopagnosia (Table 1). A deficit in the simultaneous discrimination of small metric differences of complex, biologically appearing stimuli may thus underlie DP.
Meeting abstract presented at VSS 2017