Development of an Experimental Protocol to Study the Neural Control of Force and Impedance in Wrist Movements with Robotics and fMRI

Abstract

Robotic exoskeletons have emerged as beneficial tools in the field of rehabilitation, yet their full potential is impeded by our limited knowledge of the neural control of movements during human-robot interaction. To personalize exoskeleton protocols and improve individuals motor recovery, we must advance our understanding of how the brain commands movements in physical interaction tasks. However, interpreting the neural function associated with these movements is complex due to the simultaneous expression of at least two control policies: force and impedance control. This hinders our ability to isolate these control mechanisms and pinpoint their neural origins. In this study, we evaluate the capacity of externally applied forces to decouple the expression of force and impedance in a wrist-pointing task, a necessary step in isolating their neural substrates via neuroimaging. We first conducted simulations using a neuromuscular model to examine how both force and impedance commands are updated when participants are asked to perform reaching movements in the presence of an externally applied force. Then, we recruited seven participants to perform a wrist-pointing task with the MR-SoftWrist, an MRI-compatible wrist robot. The task included four different force conditions - no force, positive constant force, negative constant force, and divergent force, each carefully selected to decouple expression of force and impedance control. Furthermore, we evaluated the efficacy of our proposed conditions for a neuroimaging experiment through simulations of neural activity. We show that these applied forces elicit distinct and predictable torque and stiffness expression, laying the groundwork for reliably identifying their associated neural activity in a future neuroimaging study.