Larger Receptive Field Size as a Mechanism Underlying Atypical Motion Perception in Autism Spectrum Disorder

Abstract

Atypical visual motion perception has been widely observed in individuals with autism spectrum disorder (ASD). The pattern of results, however, has been inconsistent. Emerging mechanistic hypotheses seek to explain these variable patterns of atypical motion sensitivity, each uniquely predicting specific patterns of performance across varying stimulus conditions. Here, we investigated the integrity of two such fundamental mechanisms-response gain control and receptive field size. Twenty children and adolescents with ASD and 20 typically developing (TD) age- and IQ-matched controls performed a motion discrimination task. To adequately model group differences in both mechanisms of interest, we tested a range of 23 stimulus conditions varying in size and contrast. Results revealed a motion perception impairment in ASD that was specific to the smallest sized stimuli (1°), irrespective of stimulus contrast. Model analyses provided evidence for larger receptive field size in ASD as the mechanism that explains this size-specific reduction of motion sensitivity.

Keywords: autism spectrum disorder; motion perception; receptive field size; response gain control.

Figure 1

(A) Predictions derived from impairments in gain control and receptive field size. Results are shown for a full range of contrasts, assuming a small stimulus size. Atypicalities in neural responses (left) result in reciprocal changes of perceptual thresholds (right). Solid line shows typical neural responses and perceptual thresholds over a range of stimulus contrasts. Dotted line illustrates effects of impaired response gain control, which increases neural responses and decreases perceptual thresholds, particularly at higher contrasts. Dashed line illustrates effects of atypically large receptive field size, which reduces neural responses across all contrast levels. (B) Perceptual consequences of changes in gain control and receptive field size. Left panel shows the typical pattern of results over a full range of stimulus sizes and contrasts, in which data suggest spatial summation at low contrast and spatial suppression at high contrast. The remaining panels depict selective changes in the excitatory receptive field size (center) and response gain parameters (right), and the resulting predicted thresholds. These predictions are derived from the main model used in this study. (C) We selected three stimulus conditions—mixed-contrast/small size, mixed-size/high-contrast, and mixed-size/low-contrast—to best test predictions derived from our two main hypotheses. All conditions included eight stimulus levels (note: top and bottom panels are depicted here on different scales to better illustrate the stimuli). (D) Predicted thresholds change resulting from enlarged receptive field size (left) and increased response gain (right; specifically, by reduction in suppressive gain). Red and blue colors indicate increased and decreased thresholds, respectively, compared to typical results (as shown in the left panel of B). In sum, changes in gain control and receptive field size lead to markedly different patterns in motion sensitivity.

Figure 2

Results from main experiments. (A-C) Perceptual thresholds across three experimental conditions for individuals with ASD and TD. Model fits to the psychophysical data from the main model are represented by the solid (ASD) and dashed (TD) lines. For the mixed-contrast/small-size condition (A), individuals with ASD showed higher thresholds (impaired motion sensitivity) across all contrast levels compared to those with TD. For the mixed-size/high-contrast condition (B) and the mixed-size/low-contrast condition (C), no group differences were observed. (D) Estimated excitatory receptive field size over a range of stimulus contrasts. For both ASD (red) and TD (blue), estimated excitatory receptive field size decreased with increasing contrast. Solid lines represent the model estimate. Shaded regions indicate bootstrapped 68% confidence intervals. The excitatory receptive field size was significantly greater in ASD compared to TD.

Figure 3

Control experiment and supplemental analyses. (A) Exact replication of Foss-Feig et al. (2013), testing three stimulus sizes at high-contrast using the QUEST procedure. Foss-Feig et al. (2013) are plotted for comparison (diamonds and dashed lines). Our results show no group difference in motion sensitivity, consistent with our results shown in Figure 3B. When compared with Foss-Feig et al. (2013), the ASD group from Foss-Feig et al. (2013) shows enhanced performance relative to the other three groups. See Results for more details. (B) Comparison of results using different adaptive presentation methods (FAST versus QUEST). Data largely fall near the unity line, indicating no difference between the two procedures (all p's > .34). (C) Effects of stimulus context. As in panel B, data points on the unity line indicate no effects of stimulus context. Data show that for identical (high contrast, 1° size) stimuli, thresholds were largely unaffected by stimulus context, with ASD performing slightly worse in the mixed-contrast/small-size condition. Both groups show the predicted pattern of worse performance in the mixed-contrast/small-size condition where stimuli were presented at 2% contrast compared to the mixed-size/low-contrast condition where stimuli were presented at 2.3% contrast.