Adults with ASD Signal Conversational Turn Taking with Their Hands

Background:

Co-speech hand gestures serve many communicative functions, often denoted by gesture “types.” For example, gestures can be deictic (pointing), or representational (depicting physical properties of referents, e.g., shape/movement). The evidence for whether or not verbally fluent people with ASD use gestures to fulfill similar functions as those without ASD is mixed. However, to date, no ASD study has examined interactive gestures, which serve primarily pragmatic and discourse functions, such as indicating uncertainty or regulating conversational turn-taking.

Objectives:

Co-speech gestures in ASD have primarily been studied via elicited narratives, which offer limited opportunities for back-and-forth interaction, and thus, reduced performance of interactive gestures. Gestures serve different functions based on social context, thus, the objective of the current study was to examine communicative functions of co-speech gestures during back-and-forth conversation.

Methods:

Adults with ASD (n=21) and age-, gender, and IQ-matched typically developing controls (TDC; n=21) completed a five-trial collaborative referential communication task designed to elicit spontaneous back-and-forth conversation in a controlled setting. We examined whether gestures broadly fulfill different communicative functions in ASD vs. TDC by coding gesture types; gestures were coded as interactive, representational, deictic, beat (i.e., moving hands in time to speech), numeric (e.g., holding up two fingers to indicate two of something), and other (primarily unclassifiable gestures). As described below, a surprising finding emerged, which prompted us to further explore the varied functions of interactive gestures used by this sample. Interactive gestures were thus further coded according to the pragmatic/discourse function served, as follows: regulating turn-taking, expressing uncertainty, marking language as reflecting shared information, indicating agreement, and indicating disagreement.

Results:

Participants with ASD gestured at a marginally lower rate overall (p=.09, d=0.54). Across groups, participants primarily used representational gestures, with beat and interactive gestures the next most common (main effect of type, p<.001, η²=.79, see Figure). Groups differed in terms of their preference for different types (group X type interaction, p=.002, η²=.45), a large effect that was primarily driven by an increased rate of interactive gestures (p=.03, d=0.78), and other/unclassifiable gestures (p=.01, d=1.00) in ASD. This finding ran counter to our original hypothesis that adults with ASD would use fewer interactive gestures, so we conducted an exploratory analysis to investigate the interactive functions adults with ASD may be signaling in gesture. Groups were comparable across all subcategories of interactive gestures, with the exception of turn-taking, which was signaled via gesture over three times as often in the ASD group (p=.01, d=0.98, see Table).

Conclusions:

Adults with ASD surprisingly used more gestures than controls to serve interactive functions in their discourse. An exploratory analysis of these functions suggested that adults with ASD may use gesture to regulate conversational turn taking to a greater extent than controls. Conversational turn taking in general has been understudied in ASD, and the results of the current study provide compelling evidence that we should broaden our units of analysis to include nonverbal communication (including gesture, but also including, e.g., eye gaze, body movements, and prosody) to fully understand conversational strengths and weaknesses in ASD.