Automated Measurement of Head Movement in Children with and without ASD

Thursday, May 12, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
K. B. Martin1, Z. Hammal2, J. Cohn3, J. Cassell4 and D. S. Messinger1, (1)Psychology, University of Miami, Coral Gables, FL, (2)The Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, (3)Psychology, University of Pittsburgh, Pittsburgh, PA, (4)School of Computer Science, Carnegie Mellon University, Pittsburgh, PA

Motor and social deficits are associated in children with Autism Spectrum Disorders (ASD). Successful social interactions require typical coordination and motor movement initiations (Piek & Dyck, 2004). Motor abnormalities, such as head motion atypicalities, in development may contribute to the perceptual and social impairments that characterize individuals with ASD. To date, deficits in motor movement in children with ASD have been characterized descriptively by human observers. In lieu of laborious human coding, automated measurement provides objective, continuous measurement to quantify head position and head movement.


To objectively quantify differences in head movement in children with and without ASD.


Participants were 48-to-68- month old children with a diagnosis of ASD (n=20) or no evidence of ASD (n=20). A diagnosis of ASD or no ASD was confirmed for all children with the Autism Diagnostic Observation Schedule (Lord et al., 2000), which yielded severity scores for social affect  and restricted, repetitive behaviors (Hus, Gotham, & Lord, 2014). Children were video recorded while watching a 16 minute video of six stimuli blocks, 3 social and 3 nonsocial stimuli. Three degrees of rigid head movement—pitch (nodding), yaw (head turns), and roll (lateral head inclinations) were tracked from the video recordings using an automatic person-independent tracker (Zface; Jeni, Cohn, & Kanade, 2015) (Figure 1). To measure the dynamics of head movement, head angles were converted into angular displacement and angular velocity. The root mean square was then used to measure the magnitude of variation of the angular displacement and the angular velocity for yaw, pitch, and roll for each stimulus block and for each infant separately.


Repeated measures ANOVA revealed differences in yaw angular displacement and roll angular velocity by stimulus block, ps<.05. There were no interaction effects of stimulus block and ASD status in angular displacement or velocity of yaw, pitch, and roll. Children with ASD exhibited larger angular displacement of yaw and roll than children without ASD, F(1, 36)=5.29, p=.027, partial ηp2=.13 and F(1, 36)=4.10, p=.05, ηp2=.10, respectively, indicating greater variability in head position. They also exhibited greater angular velocity of yaw, F(1, 36)=4.09, p=.050, ηp2=.10, and roll, F(1, 36)=7.69, p<.01, ηp2=.18, than children without ASD (Figure 2), indicating greater variability in head motion. There were no ASD status effects in angular displacement or velocity of pitch, ps>.10. Over both groups, children with greater angular displacement in yaw showed higher levels of repetitive behaviors (r=.37, p<.01) and social affect deficits (r=.30, p=.036). 


To our knowledge, this is the first study to use objective measurements to quantify head movement differences in children with and without ASD. The findings suggest that children with ASD nod (pitch) and turn their heads (yaw) with greater variability than children without ASD. Moreover, greater variability in head turns (yaw displacement) was associated with higher ASD symptomatology in both the social affect and repetitive behavior domains. These head movement differences may create sensory experiences for children with ASD that involve greater accommodation to changes in head position and movement than those experienced by children without ASD.