Biomarkers of Abnormal Temporal Gesture-Speech Integration in ASD

Background:   Ineffective communication via speech and gesture (meaningful hand movements) is a primary diagnostic characteristic of Autism Spectrum Disorder (ASD). Typically developing (TD) individuals are highly sensitive to temporal asynchrony of gesture and speech.  However, individuals with ASD show decreased sensitivity to temporal asynchrony of visual and verbal stimuli. Currently, it is unclear how this decreased sensitivity to temporal asynchrony affects gesture-speech integration in ASD, and whether it is reflected in the neural signal. 

Objectives:   The objectives of this research were (1) to examine temporal gesture-speech synchrony in ASD and (2) to determine whether neural activity reflects sensitivity to temporal gesture-speech asynchrony. It was predicted that gestures and speech produced by individuals with ASD would show greater temporal asynchrony than those of TD peers. Additionally, it was hypothesized that posterior superior temporal sulcus (pSTS) and inferior frontal gyrus (IFG), two brain regions critical to language and multimodal processing, would demonstrate sensitivity to temporal gesture-speech asynchrony.

Methods:   Participants in study 1 included 18 high-functioning ASD and 23 TD individuals matched in age, gender, and verbal IQ (all ps < .05). In this study, participants viewed a brief cartoon video and retold its events to an experimenter while being tacitly video recorded. Gestures and speech were later transcribed and coded by two raters unaware of the experimental design and predictions (ICC=.79 for speech onset; .89 for gesture onset). To control for group differences in speech production, rate of gestures was normalized for speech production. Participants in study 2 included 15 TD individuals. In this study, participants viewed clips excerpted from a longer video of a speech, such that a beat (simple rhythmic) gesture always occurred at 1.5 s.  Accompanying audio clips were presented either simultaneously with or 500 ms preceding video clips.  Neural data was collected continuously using magnetoecepholagraphy (MEG), a neurophysiological method with high spatiotemporal accuracy, and was temporally synched to gesture onset for analysis.

Results:   In study 1, individuals with ASD produced more temporally asynchronous gestures than TD individuals (p=.05). In study 2, less activity was observed in pSTC and IFG in left hemisphere during concurrent speech and gesture processing than during processing of speech preceding gesture by 500 ms (p < .05, corrected; see Fig. 1). In contrast, more activity was observed in these regions in right hemisphere during concurrent gesture-speech processing than during the processing of speech preceding gesture (p < .05, corrected).

Conclusions:   The results of study 1 demonstrate greater temporal asynchrony in gesture-speech production in ASD than in TD, suggesting that insensitivity to temporal asynchrony in ASD is reflected in gesture-speech production. The results of study 2 show that activity in pSTC and IFG reflects sensitivity to temporal gesture-speech asynchrony. Furthermore, they indicate that laterality of activation in these regions reflects the directionality of this sensitivity. These results form the basis of future research that will use the methods employed in study 2 to examine the neural bases of temporal gesture-speech integration in ASD, which will reveal biomarkers of abnormal receptive communication in this disorder.