Abstract
Individuals with Autism Spectrum Disorder (ASD) show impaired recognition of faces and facial expressions. Face recognition deficits are believed to result from a lack of (or dysfunction in) configural/holistic processing mechanism and a shift toward feature-based recognition strategies. The current study addressed this deficit using a two-part experiment that independently measures face perception and face recognition strategy (Schwartz & Chang, 2008). In Part 1 of the experiment, we measured face discrimination thresholds for seven internal face dimensions (i.e. eye shape, nose-mouth distance, etc) using a same-different task. Face stimuli were caucasian and were displayed in grayscale, and an adaptive thresholding procedure (QUEST) was used. In Part 2 of the experiment, we measured face recognition strategy using a 3-alternative forced error (3-AFE) design, a modified 3-AFC, delayed match-to-sample task in which there is no correct answer. Distractors were calibrated using thresholds measured in Part 1 and each distractor varied in only one dimension from the target so participant response indicated the dimension with the least utility. Participants were grouped according to degree of autistic symptomatology using the AQ-50. Participants scoring 2 standard deviations above and below norms were placed into High AQ and Low AQ groups, respectively. High AQ individuals showed lower discrimination threshold for eye shape, nose shape, eye-eye distance, eye-nose distance, and nose-mouth distance (five of the seven dimensions manipulated). These results suggest that high ASD individuals are significantly more sensitive to differences in those dimensions. When recognizing a target face from a set of distractors, High AQ individuals used eye-nose distance and nose-mouth distance more than Low AQ individuals, but used eye shape and eyebrow shape less than Low AQ controls. These results suggest high ASD individuals will outperform neurotypical observers when recognizing faces in which nose- and mouth-related distances are diagnostic of identity or facial expression.
Meeting abstract presented at VSS 2013