Visual processing and decision-making in autism and dyslexia: Insights from cross-syndrome approaches

Abstract

Atypical visual processing has been reported in developmental conditions like autism and dyslexia, and some accounts propose a causal role for visual processing in the development of these conditions. However, few studies make direct comparisons between conditions, or use sufficiently sensitive methods, meaning that it is hard to say whether atypical visual processing tells us anything specific about these conditions, or whether it reflects a more general marker of atypical development. Here I review findings from two computational modelling approaches (equivalent noise and diffusion modelling) and related electroencephalography (EEG) indices which we have applied to data from autistic, dyslexic and typically developing children to reveal how the component processes involved in visual processing and decision-making are altered in autism and dyslexia. The results identify both areas of convergence and divergence in autistic and dyslexic children's visual processing and decision-making, with implications for influential theoretical accounts such as weak central coherence, increased internal noise, and dorsal-stream vulnerability. In both sets of studies, we also see considerable variability across children in all three groups. To better understand this variability, and further understand the convergence and divergence identified between conditions, future studies would benefit from studying how the component processes reviewed here relate to transdiagnostic dimensions, which will also give insights into individual differences in visual processing and decision-making more generally.

Keywords: Motion perception; cognitive modelling; decision-making; developmental conditions; diffusion model; equivalent noise.

Figure 1.

Schematic representation of motion tasks presented in studies using the equivalent noise approach (A–B) and the equivalent noise function (C). Source. Figure adapted from Manning, Hulks, et al. (2022). A. Schematic representation of a trial from the motion coherence task in which 40% of dots are signal dots moving in a coherent direction (rightward in this example, outlined in red for illustrative purposes) among randomly moving noise dots. The participant is asked to determine whether the overall motion is towards the red (left) or green (right) rocks. B. Schematic representation of a trial from the Gaussian motion task, in which the dot directions are taken from a Gaussian distribution with a standard deviation of 10° and mean offset of +45°. The participant is asked to determine whether the overall motion (i.e., mean offset) is towards the red (−45°) or green (+45°) reef. C. Example equivalent noise function relating direction discrimination thresholds to external noise (i.e., the standard deviation of dot directions presented in the Gaussian motion task (B)). Direction discrimination thresholds are relatively unaffected by low levels of external noise, as internal noise dominates. However, as external noise is increased further, the internal noise is swamped and thresholds start to increase. In our tasks with children, the equivalent noise function was constrained by data from two conditions. In the no-noise condition (blue), the standard deviation was fixed at 0° and the no-noise threshold was obtained by varying the mean offset. In the high-noise condition (red), the mean offset was fixed at ±45°, and the standard deviation was varied to find the maximum tolerable noise. Sampling and internal noise were then estimated. Reduced sampling shifts the function upwards, with reduced discrimination performance at all levels of internal noise. By contrast, increased levels of internal noise lead to higher thresholds at low levels of external noise and a rightwards shift of the elbow of the function, so that more external noise is required before thresholds start to increase.

Figure 2.

Schematic representation of the decision-making process in the diffusion model for a trial with rightward motion. Source. Figure reproduced from Manning, Hassall, Hunt, Norcia, Wagenmakers, Snowling, et al. (2022). Decision-making process represented as a noisy accumulation of evidence from a starting point, z, towards one of the two decision bounds. In Manning, Hassall, Hunt, Norcia, Wagenmakers, Snowling, et al. (2022) and Manning, Hassall, Hunt, Norcia, Wagenmakers, Evans, and Scerif (2022), participants discriminated between leftward and rightward motion as quickly and accurately as possible, so the decision bounds corresponded to left and right responses. Boundary separation, a, represents the width between the two bounds and reflects response caution. Wider decision boundaries reflect that more evidence is required before making a decision (i.e., more cautious responses). Drift rate, v, reflects the rate of evidence accumulation, which depends on both the individual’s sensitivity to a stimulus and the stimulus strength. Nondecision time, ter, is the time taken for sensory encoding processes prior to the decision-making process and response generation processes after a bound is reached.