Abstract
Typical observers find it harder to match sequentially presented face halves when the target regions are aligned with task irrelevant distractor halves, than when the target and distractor regions are misaligned. This "composite face effect" is thought to be a product of holistic face processing whereby information from disparate facial regions is integrated into a unified percept. Interestingly, individuals with developmental prosopagnosia (DP), a neurodevelopmental disorder associated with difficulties recognising faces, are thought to show reduced composite face interference, suggestive of aberrant holistic face processing. To determine whether reduced composite face effects previously reported in DP reflect a domain specific impairment, or a broader perceptual issue, we sought to compare composite effects seen with faces and pseudo-words, in individuals with DP and typical observers. Sixteen individuals with DP and 16 matched controls completed a simultaneous matching version of the composite paradigm. Observers were required to judge whether pairs of face or pseudo-word composites were constructed using the same upper halves. Alignment (aligned, misaligned) and composite type (faces, pseudo words) were manipulated within-subjects and interleaved within mini-blocks. In each condition, stimulus pairs were visible until participants responded. Contrary to our expectations, the two groups showed similar composite effects for both the pseudo-word and face stimuli. Observers were slower and less accurate when judging the target halves when distractors were aligned, than when misaligned. The present results indicate that individuals with DP, under some circumstances, exhibit typical behavioural markers of holistic representation. These findings challenge the popular view that the face recognition difficulties seen in DP are due to reduced feature integration. We speculate that observers with DP are able to integrate information from different facial regions, but may describe local features less efficiently.
Meeting abstract presented at VSS 2016