Theodoros Bermperidi ^1* Richa Rai ¹ Elizabeth B. Torres ^2*

• ¹ Rutgers, The State University of New Jersey, New Brunswick, New Jersey, United States
• ² Psychology, Computer Science, Cognitive Science, Rutgers, The State University of New Jersey, New Brunswick, New Jersey, United States

Dyadic social interactions evoke complex dynamics between two agents that while exchanging unequal levels of body autonomy and motor control, may find a fine balance to synergize, take turns and gradually build social rapport. To study the evolution of such complex interactions, we currently rely exclusively on subjective pencil and paper means. Here we complement this approach with objective biometrics of socio-motor behaviors conducive of socio-motor agency. Using a common clinical test as the backdrop of our study to probe social interactions between a child and a clinician, we demonstrate new ways to streamline the detection of social readiness potential in both typically developing and autistic children, by uncovering a handful of tasks that enable quantification of levels of motor autonomy and levels of motor control. Using these biometrics of autonomy and control, we further highlight differences between males and females and uncover a new data type amenable to generalize our results to any social setting. The new methods convert continuous dyadic bodily biorhythmic activity into spike trains and demonstrate that in the context of dyadic behavioral analyses, they are well characterized by a continuous gamma process that can classify individual levels of our thus defined socio-motor agency during a dyadic exchange. Lastly, we apply signal detection processing tools in a machine learning approach to show the validity of the streamlined version of the digitized ADOS test. We offer a new framework that combines stochastic analyses, nonlinear dynamics, and information theory, to streamline and to facilitate scaling the screening and tracking of social interactions with applications to autism.