The Role of Intersensory Redundancy in the Typical Development of Social Orienting across Infancy: a New Hypothesis for Autism

Background: Children with autism show impairments in social orienting and attention. Understanding these impairments requires understanding the typical development of social orienting across infancy, the period during which it develops. However, no research has systematically assessed changes in attention to social versus nonsocial events across infancy. According to the Intersensory Redundancy Hypothesis (IRH, Bahrick & Lickliter, 2002), infants show heightened attention to multimodal events that provide intersensory redundancy (synchrony, rhythm, tempo common to audible and visible stimulation). Relative to nonsocial events, social events provide an extraordinary amount of intersensory redundancy (across face, voice, and gesture). We hypothesize that if sensitivity to intersensory redundancy underlies the development of social orienting, then infants should show differences in basic measures of attention across age as a function of redundancy. A slight disturbance of intersensory processing could then promote social orienting impairments in autism.

Objectives: We assessed the typical developmental trajectory of attention to social and nonsocial events that provide intersensory redundancy (audiovisual) versus no redundancy (unimodal visual) across the ages of 2-8 months. We predicted greater attention (more processing time and fewer disengagements) to events with more redundancy: audiovisual social and nonsocial events > unimodal visual social and nonsocial events, and audiovisual social events > than all other event types, given that social events typically amplify redundancy.

Methods: Data from 705 infants at 2, 3, 4-5, or 6-8 months of age (N = 140, 157, 239, and 169, respectively) were analyzed. Infants were habituated to dynamic displays of bimodal audiovisual versus unimodal visual (silent) social events (women speaking in infant-directed speech) or bimodal audiovisual versus unimodal visual nonsocial events (toy hammer tapping a rhythm). Mean looks away per minute (disengagement) and mean length of time to habituation (processing time) were evaluated.

Results: Age (2, 3, 4-5, 6-8 months) x event type (social, nonsocial) x condition (unimodal, bimodal) between subjects ANOVAs indicated main effects of age with decreasing processing time and increasing disengagements across age, ps < .001. Consistent with our predictions, main effects of condition and event type indicated longer processing times and less disengagement for social than nonsocial events (ps < .005) and for bimodal redundant than unimodal nonredundant stimulation (ps < .001). These main effects were each qualified by interactions with age (ps < .01) where differences between conditions were most apparent for older infants. Slope analyses indicated that attention to bimodal, redundant social events was maintained across age (p > .10), whereas attention to all other event types decreased across age (p < .01).

Conclusions: Consistent with predictions of the IRH, these findings demonstrate that attention (processing time and disengagement) to bimodal, redundant social events is maintained across age, from 2-8 months, whereas attention to unimodal and nonsocial events declines across this period. These are the first findings indicating that social orienting develops gradually across infancy, emerging by 3 months, and is a function of intersensory redundancy. Together with findings of impaired intermodal functioning in autism, these findings suggest that intermodal processing disturbance may underlie social orienting impairments in autism.