Children and adults with autism spectrum disorder (ASD) and little or no spoken language, labeled ‘minimally verbal’ (MV), often fail conventional assessments of cognitive abilities and are typically judged to be low functioning or cognitively impaired. However, it is unclear whether these failures reflect a true cognitive impairment or, alternatively, might reflect a severe deficit in the ability to perform the necessary motor expressions used to evaluate cognitive skills in certain types of testing. Indeed, case descriptions of MV-ASD individuals and similarities with catatonia suggest that these individuals may have a core action-control deficit for voluntary or intentional behavior, which would result in a significant gap between the observed and actual cognitive abilities. Such a deficit could explain the unreliable pointing behaviors observed even in the simplest assessments and the typical floor performance seen in cognitive assessments.

In this pilot study, Yoram Bonneh proposes to investigate the hypothesis that MV-ASD individuals suffer from core action-control deficits. To do so, he and his collaborator Niva Shefer-Kaufmann will characterize cognition in MV-ASD individuals by analyzing small, involuntary eye movements, with the idea that these involuntary movements may bypass any voluntary action-control deficit and could reveal unknown cognitive skills, such as reading. Bonneh and his team will use advanced eye-tracking methods based on high temporal precision to assess basic reading and receptive language by analyzing the involuntary eye movements of individuals hearing and watching brief presentations of spoken words, text and pictures.

The study will focus on two specific aims: (1) the investigation of hidden reading skills and knowledge in MV-ASD adolescents and young adults who are considered cognitively impaired and (2) the investigation of a possible gap between pointing behavior and eye-gaze performance in young (2–5 years old) children with ASD who are either MV or exhibit delayed speech.

Preliminary results with numerous neurotypical adults, six typically developing young children (2–3 years old) and, importantly, two MV-ASD adolescents initially diagnosed as illiterate, validate the utility of these proposed methods. The ability to identify minimally verbal individuals with ASD who have a significant gap between their observed and actual cognition would have a dramatic impact on the way these individuals are treated and on the interventions applied to them.