Kaleckas Lecture & Synthesis Day

Robert Sainburg, PhD, OTR, is the Dorothy Foehr Huck and J. Lloyd Huck Distinguished Endowed Chair and Professor of Kinesiology and Neurology at The Pennsylvania State University, and Director of the Center for Movement Science and Technology. His research focuses on the neural mechanisms underlying upper limb motor control, with particular emphasis on hemispheric specialization, handedness, and predictive control of limb dynamics. A major component of his work addresses upper limb motor deficits following stroke, including impairments of both the contralesional and ipsilesional (“less affected”) arms. By integrating experimental motor control, computational modeling, and clinical studies, Dr. Sainburg’s research has directly informed novel stroke rehabilitation approaches aimed at improving arm function and daily activity performance. His work has been continuously supported by the NIH and other federal agencies and has contributed to mechanism-based interventions for individuals with neurological injury.

Abstract: Accurate control of limb movement requires the nervous system to predict the mechanical consequences of motor commands in the presence of complex and variable limb dynamics. The internal model framework, sometimes controversial, proposes that the brain acquires and utilizes structured knowledge of body and environmental mechanics to support predictive control, learning, and generalization. This presentation reviews converging evidence for internal models in the control of limb dynamics drawn from studies of deafferentation, mechanical perturbation and adaptation, interlimb transfer, lateralization, and focal brain lesions. Behavioral and electromyographic findings demonstrate that motor commands anticipate interaction torques arising from intersegmental dynamics, even when sensory feedback is absent, indicating feedforward prediction. Adaptation experiments manipulating limb inertial properties reveal persistent aftereffects and generalization across movement directions in intrinsic (joint-based) coordinates. Interlimb transfer of learned dynamics is asymmetric, occurring primarily from the dominant to the non-dominant arm, suggesting implicit representations. These findings align with hemispheric specialization, with the left hemisphere preferentially supporting predictive control and the right hemisphere contributing to impedance-based stabilization. Finally, the translational implications for stroke rehabilitation are discussed, highlighting how hemisphere-specific disruptions of internal models contribute to ipsilesional arm deficits and how targeted intervention can produce durable functional improvements.