In occupational domains such as sports, healthcare, driving, and military, both individuals and small groups are expected to perform challenging tasks under adverse conditions that induce transient cognitive states such as stress, workload, and uncertainty. Wearable and standoff 6DOF sensing technologies are advancing rapidly, including increasingly miniaturized yet robust inertial measurement units (IMUs) and portable marker-less infrared optical motion tracking. These sensing technologies may offer opportunities to track overt physical behavior and classify cognitive states relevant to human performance in diverse human-machine domains. We describe progress in research attempting to distinguish cognitive states by tracking movement behavior in both individuals and small groups, examining potential applications in sports, healthcare, driving, and the military. In the context of military training and operations, there are no generally accepted methods for classifying transient mental states such as uncertainty from movement-related data, despite its importance for shaping decision-making and behavior. To fill this gap, an example data set is presented including optical motion capture of rifle trajectories during a dynamic marksmanship task that elicits variable uncertainty; using machine learning, we demonstrate that features of weapon trajectories capturing the complexity of motion are valuable for classifying low versus high uncertainty states. We argue that leveraging metrics of human movement behavior reveals opportunities to complement relatively costly and less portable neurophysiological sensing technologies and enables domain-specific human-machine interfaces to support a wide range of cognitive functions.
Keywords: cognitive state estimation; inertial measurement units; machine learning; movement dynamics; optical motion capture; uncertainty; workload.