Differential Sensitivity to Synthetic Face Stimuli Across Viewpoint In Autism

Background:  Face perception is the most commonly used visual metric of social abilities in autism. However, when found to be atypical, the nature of its origin is often contentious. One hypothesis proposes dysfunction of face-sensitive brain areas as a potential origin, reflecting a more general altered functioning of the social brain network (e.g., Schultz, 2005). Alternatively, a less “social” proposal suggests that autistics’ characteristic locally-oriented visual analysis ultimately affects (i) performance on most face tasks where global or configural analysis is optimal, and plausibly, (ii) the typical, experience-dependent development of integrative neural architecture involved in the perception of complex visual objects, including faces (e.g., Vlamings et al, 2010).

Objectives:  To evaluate these two hypotheses by assessing face identity discrimination between synthetic faces presented with and without viewpoint changes, with the former condition necessitating a more global, integrative analysis to be efficiently completed.

Methods:  Ten autistic and nonautistic adults matched for global IQ and age were asked to perform a face identity discrimination task similar to that of Habak et al (2008). Stimuli were synthetic faces (Wilson et al, 2002), consisting of simplified (hair and skin texture removed) and ecologically-validated stimuli extracted from traditional face photographs in both frontal (front) and 20° side (side) viewpoints. Face photographs were digitized from 37 points (including head shape, features, etc), which were then defined relative to mean head radius, to provide a continuous measure of facial geometry. Face identity discrimination thresholds (defined by the minimum % change in face geometry at 75% correct performance) were obtained using a two-alternative, temporal forced choice match-to-sample paradigm consisting of a target face  (1000ms), followed by a 200 ms mask, then by 2 choice faces presented side-by-side. Participants were asked to identify which choice face matched the target for three viewpoint conditions and one inverted face condition. Thresholds were measured when the target face and choice faces were all facing forward (a) front-front view, all facing 20° to the side (b) side-side view, when the target was facing forward and the choice faces to the side (c) front-side view: the viewpoint change condition, and when all faced forward but were upside-down (d) inverted.

Results:  Face identity discrimination thresholds were highest (worst) for the front-side view condition for both groups, suggesting that identity discrimination was most difficult across views. Between-group differences in performance were not evidenced for either (a) front-front view, (b) side-side view conditions, or (d) inverted front-front view. However, mean identity discrimination thresholds for the autism group was higher for the front-side view (the viewpoint change condition) compared to that of nonautistics.

Conclusions:  The selective decrease in autistic performance for the viewpoint change condition suggests that face identity discrimination in autism is more difficult when (i) access to local cues are minimized, and/or (ii) an increased dependence on integrative analysis is introduced to the face task used. This finding suggests that atypical face perception in autism may originate from “non-social” perceptual origins related to integrative information processes.