Responding to Joint Attention Requests From Virtual and Non-Virtual Social Partners

Saturday, May 19, 2012
Sheraton Hall (Sheraton Centre Toronto)
11:00 AM
B. Lambert1, A. Gutierrez1, W. Mattson1, J. Artigas1, O. Martinez2, M. Kimijima1, J. Cassell3, J. Cohn4 and D. S. Messinger1, (1)Psychology, University of Miami, Coral Gables, FL, (2)University of Miami, Coral Gables, FL, (3)Northwestern University, Evanston, IL, United States, (4)Psychology, University of Pittsburgh, Pittsburgh, PA
Background: Responding to joint attention (RJA) refers to following a partner’s referential cues (e.g., gaze shift, head turn, pointing) by shifting gaze from the partner to the indicated object or event. Impairments in RJA have been observed in interactive and video protocols among children with autism spectrum disorders (ASDs) and are linked to delays in language development. However, little is known about how children with ASD respond to referential cues provided by different kinds of partners.

Objectives: Determine differences in RJA in response to virtual and non-virtual characters in video stimuli among children with ASD and typically developing (TD) controls.

Methods: The sample consisted of 16 TD (11 male) and 13 ASD (10 male) children from 36.8- 81.9 months of age. For the TD group, mean chronological (CA) was 55.5 months (SD=9.9) and mental age (MA) was 61.2 months (SD=7.1). For the ASD group, mean CA was 66.6 months (SD=13.4) and MA was 62.7 months (SD=15.5). Group differences in MA were not significant, p=.73.

Participants were seated in front of an LCD monitor mounted with a Tobii X50 eye tracker and presented a 14-minute video. Gaze shifts were recorded during 2 two-minute sequences featuring first a non-virtual (actual) boy and then a virtual (animated) boy using an identical audio track. In both sequences, the boy faced the child and performed a series of RJA presses (vocalizations, head turns, pointing) to direct the child’s attention to a SpongeBob cartoon on a TV to the child’s right.

Results: Children with ASD demonstrated fewer gaze shifts (M=1.01, SE=.12) from the boy to SpongeBob than TD children (M=1.53, SE=.11), F(1, 170)=9.95, p < .01, ηp2=.06. Children with ASD also demonstrated fewer gaze shifts from SpongeBob back to the boy (M=0.80, SE=.11) than TD children (M=1.33, SE=.10), F(1, 170)=12.45, p < .001, ηp2=.07. There was no effect of type of boy (actual vs. virtual) and no interaction between boy and status, ps>.25. Children in the ASD group had a lower proportion of tracked frames (M=.56, SD=.33) than children in the TD group (M=.70, SD=.30), F(1, 398)=19.68, p < .001, ηp2=.05. Analyses controlling for proportion of tracked frames (number of gaze shift per minute of tracked frames) yielded no significant main or interaction effects.

Conclusions: Children with ASD engaged in fewer instances of RJA (boy to SpongeBob) shifts and fewer instances of shifting gaze back from SpongeBob to the boy. However, children with ASD also contributed a lower proportion of tracked frames to analysis than TD children presumably because they spent less time gazing at the monitor. No deficits in gaze shifting were evident when controlling for the proportion of frames for which eye tracking data was available. There were also no differences in how ASD and TD children responded to cues from the virtual versus non-virtual social partners. Difficulties with RJA among children with ASD may be partially a function of difficulties attending to the partner who is directing their attention.

| More