Magnitude of Perceptual Peaks in Autism Is Partially Dependent on the Matching Variable: The Example of Pitch Discrimination

Background:  When comparing cognitive performance between autistic and comparison populations, groups are matched for intellectual functioning, assuming that the effect of the matching variable on performance is the same for both groups. The most frequent matching variable in autism studies is Wechsler Full Scale IQ (FSIQ; Mottron, 2004). However, it has been shown that this measure underestimates autistics’ cognitive level when compared to another major test of general intelligence, Raven’s Progressive Matrices (RPM; Dawson et al., 2007), which minimizes the need for typical speech comprehension and production. It is therefore possible that autistics’ perceptual peaks are, at least partially, an artifact of the matching variable.

Objectives:  To investigate this possibility, the current study tested the influence of RPM, as a more representative measure of autistic intelligence, on the magnitude of perceptual peaks in autism.

Methods:  34 autistic and 34 typically developing adolescents and adults, aged between 14 and 35 years, were tested on Wechsler FSIQ, RPM, and pitch discrimination, using an adaptive psychophysical task.

Results:  Regression analyses revealed a Group X IQ interaction (p=.01): Wechsler FSIQ predicted pitch discrimination performance in control participants (p=.004 R²=.176), but not in autistic participants (p>.20). RPM predicted discrimination performance similarly in both groups (p=.001; no Group X RPM interaction p=.23). Group comparisons entering either Wechsler FSIQ or RPM as a control variable consistently revealed significantly better performance in autistics compared to controls. However, the effect size was smaller when using RPM, rather than Wechsler FSIQ, thereby reducing the magnitude of the peak of ability.

Conclusions:  RPM but not Wechsler FSIQ predicted perceptual ability in autistics, suggesting that RPM is a more consistent measure of autistic intelligence than Wechsler scales. Matching on RPM diminished but did not eliminate an auditory perceptual peak, indicating that enhanced perceptual functioning cannot be explained by an incorrect matching strategy. This finding, in addition to  the very low level of perceptual architecture where superiorities are evident, and the very young age at which these superiorities can be demonstrated (Kaldy et al., 2011),  argues in favour of a primary role of early perceptual alteration in the cascade of effects resulting in the autistic cognitive and behavioural phenotype.