Potential movement biomarkers for autism in children and adolescents

Background -- While social communication deficits are the hallmark of autism spectrum disorders (ASD), motor deficits are known to be common in this population as well. Recently, members of our research team showed that kinematic markers collected by playing a tablet game may be a promising biomarker for identification of ASD as compared to a typically developing population (TD) in children ages 3-6 years old (Anzulewicz et al, 2016). To our knowledge, no one has replicated this finding in an older population. Purpose -- To replicate and extend previous findings of kinematic differences in children with ASD to an older population of children (9-14 years old). Methods -- Four TD children and 5 children with ASD (aged 9-12) played an iPad drawing game (Anzulewicz et al, 2016) that measured gesture kinematics and gesture force using inertial sensors and touch screen touch displacements. 212 features were calculated from the inertial sensor and touch screen data (ibid). A Kolmogorov-Smirnov (K-S) test was run to identify motor features distinct between ASD and TD children. Results -- K-S test identified seven significantly different features (JerkMagnitudeMax, JerkMin_y, JerkRange_y, AttitudeRange_y, RotationRMS_x, RotationStdDev_x, JerkZeroCrossing_x) between ASD and TD groups that represented differences in acceleration of finger movements and the displacement of the iPad during movements. Conclusions -- Results demonstrated inertial movement sensor parameter differences are key identifiers between 8-12 year old ASD and TD children, common to children 3-6 years old. Contact forces and the distribution of forces during coloring may serve as important identifiers of ASD irrespective of age during childhood, while other parameters may be age-dependent. Research Support NIH R01 (1R01HD079432-01A1