Abstract
One potential explanation for the deficit in holistic processing exhibited by congenital prosopagnosics (CPs) is that CPs have smaller receptive fields (r.f.s) in face selective regions (Witthoft et al., 2016). Large, overlapping receptive fields allow small differences in similar faces to affect the activation of cells whose r.f.s are distributed throughout the face. Xu et al. (2014) demonstrated that the face configural effect, described as the benefit in accuracy in identifying a studied individual’s eyes, nose, or mouth in the context of the individual’s face as compared to when presented in isolation, could be attributed to large r.f.s. While effective in measuring the configural effect in control subjects, their memory-based paradigm was ill-suited to measure the configural effect in individuals who struggle with learning faces. The present study used an odd-man-out paradigm in which 216 subjects had to detect which one of a diagonal array of three faces was different. Each image included a pair of eyes, a nose, and a mouth, either arranged from top to bottom or in the reverse order and either viewed in the context of a head or in isolation (Fig. 1). Subjects who failed to demonstrate normal proficiency on standard tests of face recognition (USCFPT, CFMT, PI20) (Fig. 2) showed significantly reduced benefits on their detection RTs from the presence of a head and the natural ordering of the face parts (Fig. 2) indicating their failure to benefit from a configural representation
Acknowledgement: Harold W. Dornsife Research Fund