Statistical Motor Biomarkers Characterizing age-dependence in Neurodevelopmental Disorders

We studied the statistical properties of motor kinematics of individuals with Neurodevelopmental Disorders (NDD). Humans show random arm kinematic trajectories that appear smooth to the naked eye. By using high definition motion capture sensors attached to moving arms we found millisecond time fluctuations leading to a statistical biomarker [1]. We broadened our previous studies of subjects with Autism Spectrum Disorder (ASD), to subjects with Attention Deficit Hyperactivity Disorder (ADHD), Comorbid ASD and ADHD. The subjects moved their arm towards touch screen targets appearing randomly in the monitor. We measured the statistical properties of the angular velocity, acceleration, and jerk of upper extremity movements. We introduce a new Fano-like Factor (FF) biomarker of the distribution of fluctuations in the kinematic variables. Implementing a Support Vector Machine Learning tool allowed us a separation in the cognitive conditions of the NDD subjects studied. Subjects with ASD+ADHD show an age dependent change as well as a small sample of ADHD subjects. This is consistent with clinical studies that record improvement in clinical ADHD severity and impairment with age.
[1] Wu, et al. Scientific Reports (Nature), 8(1):614, 2018.