Lessons from Psychophysical Studies of Somatic Sensation in Autism

Background: Atypical responses to sensory input are hypothesized to cascade into higher-order deficits in individuals with autism spectrum disorder (ASD). However, there is a gap between the precise control of the stimulus-perception relationship afforded by psychophysics and questionnaire measures of sensory reactivity in daily life, with the latter generally mapping better onto higher-order deficits such as social-communication difficulties. Adjustments to the design and execution of standard psychophysical approaches may improve the ability to bridge this gap. Examples include designs that consider motor and attention differences, as well as those that incorporate evidence accumulation and consider data points surrounding the point of binary perceptual decisions.

Objectives: To describe two recent studies that offer insights into how psychophysics may be adapted to serve as a more effective measurement tool in this population and thus better evaluate the cascading effects theory.

Methods:

Study 1: In a sample of neurotypical adults (n=43, mean age 30.1 +/- 6.9 years), measures of detection threshold and dynamic range (the width of the decision window surrounding the threshold) were obtained for a 100 msec, 35 Hz vibrotactile stimulus using the method of constant stimuli (20 levels of stimulus amplitude between 0-20 um). The relations for detection threshold and dynamic range with higher-order autism traits were evaluated according to self-reported sensory sensitivity using multiple linear regressions that tested moderated effects.

Study 2: Thermal detection thresholds were collected using the method of limits (continuous temperature increase or decrease from a 32°C baseline at a rate of 1°C/sec) in a large sample of individuals with ASD (n=84; 32 adults [mean age 28.8 years] and 51 children [mean age 11.2 years]) and typical developmental histories (n=60; 24 TD adults [mean age 29.0 years], and 36 TD children (mean age 10.03 years)). Effects of age, sex, group, trial-to-trial variability (CQV), performance IQ, and counterbalance order were included in separate regression models predicting warm and cool thresholds.

Results:

Study 1: Though neither threshold nor dynamic range was independently related to self-reported sensory or higher-order autism traits, self-reported sensory sensitivity significantly moderated the relations between dynamic range and higher-order autism traits.

Study 2: For warmth detection, trial-to-trial variability (CQV) (p<.0001) was the strongest predictor of threshold, with smaller but significant effects of performance IQ (p<.02) and sex (p<.03). For cool detection, performance IQ (p<.004) was the strongest predictor, with smaller effects of CQV (p<.02), age and sex (p’s <.04). No significant effects of group or counterbalance order were found in either model.

Conclusions:

Study 1 illustrates a) the need to consider novel psychophysical metrics, including those that incorporate information surrounding the point of a binary perceptual decision (e.g., dynamic range), in future studies of sensory function, and (b) highlight the importance of testing factors that may moderate associations of interest. Study 2 suggests that variability in reaction time may explain findings previously interpreted as between-group differences in sensory thresholds, and that method of limits or other approaches that depend on reaction time may need to be adapted or replaced for assessing sensory differences in ASD.