Inducing Change In Visual Scanning of Natural Scenes In Infants with ASD by Manipulating Physical Contingencies

Background: In recent research, we found that two-year-olds with autism did not orient towards point-light displays of biological motion but attended instead to non-social, physical contingencies—contingencies that were disregarded by control children.  In follow-up studies, we also found that toddlers with ASD varied their fixation on the faces of others as a function of the physical contingencies embedded therein. Finally, we also found that detection of audiovisual synchrony was not greatly influenced by social context in infants with ASD, although social context did alter preferential viewing patterns in typically-developing infants.

Objectives: The goal of the current project is to induce changes in visual scanning of naturalistic social scenes in infants with ASD by experimentally manipulating audiovisual synchrony embedded in those scenes, and to then quantify those behavioral changes in visual scanning as predictors of dimensional outcome.

Methods: Infants with ASD and typically-developing infants (TD), ages 12-24 months, watched a series of clips of caregivers interacting with an infant against static backgrounds showing settings familiar to the infant. In addition to the caregiver, we introduced a time-varying, moving object into the room that was dynamically synchronized with the caregiver’s speech. Specifically, we co-varied the object’s rotational motions, rocking motions, and luminance with the amplitude envelope of each utterance, and dynamically varied the degree of synchrony between caregiver and object throughout the movie. Eye-tracking technology was used to track infants’ looking patterns, using the relative fixation time on caregiver and object as our dependent measure.

Results: Preliminary results show that when the non-social object is not fully synchronous, infants with ASD look more at the synchronous face.  However, as the synchrony of the non-social object is increased, infants with ASD look more towards the object, diminishing their attention to the face. TD infants, on the other hand, preferentially attend to the socially relevant stimuli and spend more time looking at the face rather than the object, even when the motion of the object is fully synchronous to the audio of the utterance.  In addition, preliminary results show that those infants with ASD who look more towards faces in the scenes (even when the object is fully synchronous) exhibit less impaired social behavior as measured by the ADOS 1 Social Affect Score.

Conclusions: The present study suggests that while TD infants preferentially attend to socially relevant stimuli, even in the presence of competing non-social, physical contingencies, such contingencies can be highly distracting to infants with ASD, particularly in naturalistic settings.  These findings indicate an early interruption of typical social experience and suggest one mechanism by which infants with ASD may not attend to important social cues in their environment.  By inducing change in visual scanning to measure individual sensitivity to non-social contingencies, we may also learn how to develop effective early interventions for infants with ASD by either (1) minimizing the impact of distracting environmental cues, or (2) using cues that may be innately attention-getting to infants with ASD to foster socially relevant learning.