18474
Extending the Characterization of Semantics in ASD to Non-Verbal Domains

Thursday, May 14, 2015: 11:30 AM-1:30 PM
Imperial Ballroom (Grand America Hotel)
K. Steinman1,2, M. K. Askren3, T. Ward4, J. E. Elgin5, J. Wenegrat6, R. K. Earl2, R. Bernier2 and T. J. Grabowski3, (1)Seattle Children’s Research Institute, Seattle, WA, (2)University of Washington, Seattle, WA, (3)Radiology, University of Washington, Seattle, WA, (4)Clinical Psychology, Seattle Pacific University, Seattle, WA, (5)Seattle Children's Autism Center, Bothell, WA, (6)University of Washington Autism Center, Seattle, WA
Background: Recent research on semantic memory – using semantic dementia, the prototypical neurologic semantic disorder, as a model – has demonstrated an amodal global semantic knowledge system that underlies both verbal (i.e., language-based) and non-verbal semantic challenges seen in this disorder. In autism spectrum disorders (ASD), verbal semantic impairments are well-recognized and common. However, the characterization of semantics in ASD has been almost exclusively limited to verbal semantics; non-verbal semantic knowledge has yet to be examined in depth.

Objectives:

(1) To characterize verbal and non-verbal semantic knowledge of 7–15-year-olds with ASD vs matched typically-developing controls (TYP).

(2) To examine the relationship between verbal and non-verbal semantic deficits in 7–15-year-olds with ASD.

Methods: Participants with ASD, functional language, and PIQ≥85 as well as age- and PIQ-matched TYP participants underwent IQ testing (WASI-II) and a computer-adapted semantic battery to assess semantic knowledge across various modalities. Response accuracy (percent incorrect) and mean response latency (for correct responses) were collected on two sets comprised of corresponding verbal and non-verbal tasks that use the same semantic stimuli: (1) Naming (of a pictured object; Nam, verbal), Comprehension (matching word to picture; Comp, verbal), and Semantic Association (matching related pictures; SemA, nonverbal); and (2) Word-Picture (WP, verbal) and Sound-Picture (SP, non-verbal) Matching tasks. Accuracy and (item-wise) response latency differential scores were defined for corresponding nonverbal versus verbal tasks. Two-sample t-tests were used to assess between-group differences in accuracy and response latencies. Differential scores were compared within and between groups with one- and two-sample t-tests, respectively.

Results: ASD participants (n=19) displayed lower accuracy (higher percentage of incorrect responses) than TYP participants (n=20) on all verbal tasks (Nam p=0.02; Comp p=0.04; WP p= 0.005) and one non-verbal task (SemA p=0.02), but performed no differently from TYP participants on the other non-verbal task (SP p=0.3). Latencies for correct responses did not differ between ASD and TYP on any verbal or non-verbal tasks (p≥0.3). Comparing accuracy on non-verbal versus corresponding verbal tasks, ASD participants exhibited the same pattern as TYP of greater accuracy on SemA than Nam, similar accuracy on SemA and Comp, and less accuracy on SP than WP. Both groups exhibited longer latencies on SemA than Comp (both p<0.001), but only TYP exhibited significantly longer latencies on SP than WP(p<0.001 vs. ASD p=0.06). No significant differences between groups were found for differential accuracy or response latencies on non-verbal versus verbal tasks.

Conclusions: As anticipated, ASD participants exhibited lower accuracy on verbal semantic tasks than TYP participants. ASD participants also demonstrated impairments on a non-verbal picture matching task, but not a non-verbal auditory-picture matching task. This profile suggests an intact underlying global semantic knowledge system in ASD, with differential impairments in various verbal and non-verbal semantic tasks resulting from modality-specific abnormalities (e.g., language) in the semantic network. Future multivariable regression analysis will investigate semantic task performance differences adjusting for age, PIQ, and standardized verbal semantic tasks to yield a better understanding of the relationship between verbal and non-verbal semantic task impairments and deficits in different non-verbal semantic domains.