Objectives: To characterize early spatial information processing in autism by measuring contrast sensitivity functions for both luminance- and texture-defined spatial information, reflecting the connectivity of local neural networks (i.e., spatial filters) mediating static processing.
Methods: Spatial frequency tuning was assessed by measuring contrast detection thresholds to luminance- (with and without noise) and texture-defined sine-wave gratings (in gaussian envelope) of different spatial frequencies (0.5, 1, 2, 4, 8 cycles per degree [cpd]) for autistic and typically developing groups. Contrast sensitivity (1/threshold) functions (CSFs) were plotted for each group across conditions.
Results: Results demonstrate an enhanced sensitivity to luminance-defined (no-noise) gratings of high-spatial frequency (8 cpd) in the autistic group. In addition, when normalized, average peak sensitivity for the autistic group was higher (≈ 3cpd) than that of the typical group (≈2cpd).
Conclusions: Consistent with recent electrophysiological evidence (Milne et al., 2008; Mimeault et al., 2008), results of these experiments suggest that low-level static information processing is atypical in autism, and that local stimulus-driven neural networks underlying spatial filtering properties are altered, favoring the analysis of high-spatial frequency (or detailed) information. We suggest atypical lateral connectivity (i.e., enhanced lateral inhibition) within early visual areas as the most plausible type of neural alteration consistent with the behavioral data. The implication of these local network changes on larger-scale atypicalities mediating socially-related perceptual processing will be discussed.