Comparison of Methods for Assessing Receptive Language Knowledge in Minimally Verbal Children and Young Adults with Autism Spectrum Disorders

Background: Assessing receptive language skills in individuals with autism spectrum disorder (ASD) who fail to acquire functional speech (20-30%) is challenging due to pragmatic, motivational and behavioral problems.  Research has often not included these individuals and their actual linguistic knowledge remains poorly characterized.  Interest in devising alternative methods for probing comprehension abilities in minimally-verbal (MV) individuals with ASD is growing (Tager-Flusberg & Kasari, 2012), underscoring the need for analyses of the feasibility of these methods and comparisons of their efficacy.  We are conducting a multimethod study of different novel assessment measures of receptive language in MV individuals with ASD (ages 7-21) compared to verbal, high functioning (HFA) children with ASD, using eye-movements, touch screen responses, and EEG recordings alongside standardized testing and parental reports of language comprehension. 

Objectives: We aim to evaluate the feasibility of novel experimental methods to measure receptive language in school-age and older MV children with ASD, to compare their efficacy to providing valid data, and to explore individualized approaches to assessment.  

Methods: A looking-while-listening paradigm was administered to monitor language-mediated eye-movements as implicit measures of lexical knowledge using a TobiiT-60 eye- tracker. Image-pairs were displayed for 5 seconds and the target word (noun, verb, adjective - matching one image) was played 2.5 seconds after stimulus onset. Trials were distributed over 3 developmentally ordered blocks. Participants completed a similar paradigm on a touch screen device, as well as a standardized test of receptive vocabulary (PPVT-4). Parents reviewed a list of all of the words used as stimuli and indicated which were “understood only” or “understood and produced” by the child. Five MV (Mean age=16.1 years) and 4 HFA participants (Mean age=13.1 years) completed all assessments over 3-4 visits.  

Results: Analyses of eye-tracking data suggest that measures of language-mediated visual attention deployment (total duration of fixations and mean fixation count on the target image after auditory stimulus onset) provide evidence of implicit language comprehension in the MV group. The MV participants looked longer and more often at the matching versus the foil in pair (M=.23 sec. versus .16 sec., t(3) = 2.9, p =.05 and M= 1.1 versus .63, t(3) =8.2, p=.004, respectively), consistent with the pattern found in typically developing individuals, and replicated in our HFA group. This pattern was found for 4 MV participants, and matched the words parents indicated as understood by their child. Performance on the touch screen device was the most reliable indicator of comprehension for two MV participants (98% and 75% accuracy respectively, when compared to the percentage of trials with increased attention deployment to targets after word-onset, 68% and 32%, respectively, in the looking-while-listening task).  

Conclusions: The significance of alternative methods for receptive language assessment in MV persons using more sensitive measures of comprehension (e.g., attention deployment or motor response) versus relying on compliance with standardized testing will be discussed in light of the findings.  We will also discuss the need for individualized testing protocols including possible strategies for designing flexible assessments that provide valid, reliable data for this population.