Concept Identification and Formation in High-Functioning Adolescents with ASD

Background:  Abstraction refers to the ability to find common features across different stimuli. Concept identification involves recognizing patterns of features created by an external agent. Concept formation is more difficult, requiring independent creation of schemata to organize information. Impairments in concept identification and formation are theorized to underlie a variety of practical difficulties of individuals with ASD (e.g., failure to generalize learning in one context to a similar, but new context). However, past research has yielded mixed results, with some finding significant impairment and others finding intact concept identification and formation. Contradictory findings may be due to differences in assessment methodology. For example, studies utilizing strictly perceptual tasks likely minimized impairment in ASD, while studies using strictly verbal tasks likely exaggerated impairment. 

Objectives:  We sought to use a single task that would allow the evaluation of (1) concept formation relative to concept identification, and (2) verbal concept formation and identification relative to perceptual concept formation and identification. Also, we sought to explore (3) the relationship between concept identification and formation and autism symptomatology. We hypothesized that (1) we would replicate previous findings of intact concept identification but impaired concept formation in ASD (Minshew et al., 2002); (2) impaired concept formation would be more notable on verbal tasks; and (3) worse impairments would be associated with more severe autism symptoms.

Methods:  The sample consisted of 27 high-functioning (FIQ > 80) adolescents with ASD and 27 age (M 14.8 years)- and FIQ (M 102.8)-matched typically-developing controls. Concept formation and identification abilities were assessed using the Delis-Kaplan Executive Function System (D-KEFS) Card Sorting task.  One-way ANOVAs explored group differences on task performance variables.  Partial correlations explored the relationship between task performance and autism symptomatology measured using the Social Responsiveness Scale (SRS-2). Given the significant correlation between FIQ and every measure of task performance, FIQ was entered as a covariate in all analyses.

Results:  Controlling for FIQ, (1) there was no group difference on tasks requiring concept identification (Sort Recognition; assessor sorts cards); however, there was a small significant effect of group on concept formation tasks (Free Sort; participant sorts carts) (F(1,51) =4.218, p = .045, η² = .02), with the ASD group exhibiting worse performance; (2) there was no group difference on verbal or perceptual tasks of concept identification and formation; and (3) SRS-2 Total scores (higher scores indicate more severe autism symptoms) were associated with concept formation (r(54) = -.31, p < .05), but not concept identification (r(54) = -.20, p = .15).

Conclusions:  Our findings add to a growing body of research showing a dissociation between concept identification and concept formation in ASD. This dissociation, which existed after covarying FIQ, suggests that concept formation deficits may underlie impairments across levels of cognitive functioning in ASD. While previous research demonstrated that a combined measure of concept identification and formation was correlated with autism symptom severity (McLean et al., 2014), our finding that autism symptoms are significantly associated with only concept formation further strengthens the theoretical significance of concept formation deficits in ASD.