Meeting ID: 928 0482 8495 Passcode: 185518
Speaker: John Krall
Advisor: Prof. B. Ajiboye
Title: Enhancing BMI+FES Neuroprothesis via Exploration of the Neural Dynamics of Multi-joint Movements in People with Tetraplegia
Abstract: The Reconnecting the Hand and Arm to the Brain (ReHAB) clinical trial integrates a Brain Machine Interface (BMI) with a Functional Electrical Stimulation (FES) system to restore volitional control of the upper limb in individuals with tetraplegia. The BMI translates signals from motor related areas of the brain into movement commands that are used to control muscle stimulation via the FES system. The BMI+FES system has been used by our study participants to successfully complete many activities of daily living including self-feeding, handshaking (greeting), object pick-and-place, and precise grasp force modulation. Our current focus is expanding the utility of the ReHAB system by increasing the complexity of movements that can be reliably performed. While we know motor cortex plays a primary role in orchestrating dexterous movements, the strategies it employs to do so are not well understood. Individual neurons have different tuning profiles across a range of single-joint movements, and large, overlapping populations of neurons are active no matter which joints are being moved. How, then, does motor cortex generate command signals for several joints simultaneously? Using intracortical microelectrode arrays, we recorded population spiking activity in the hand knob region of motor cortex while a human study participant with tetraplegia attempted various cued movements of their arm and hand. By comparing the neural state-space trajectories recorded during the course of single- and multi-joint movements, a simple cortical code is revealed. All trajectories, regardless of the joints being moved, follow a similar rotational path, though they traverse different regions of state-space (Location-Dependent Rotations decomposition, 87.1 ± 2.8% variance explained across conditions). Additionally, multi-joint trajectories closely follow the interpolated path between their constituent single-joint trajectories ( 1.1 ± 0.4 a.u. mean distance compared to 2.4 ± 1.2 a.u. mean distance to the single-joint trajectories). While simple, the strength of the relationship cannot be explained solely by the primary features of the activity of individual neurons (p < 0.05, compared to surrogate datasets that maintain covariant structure across time, neurons, and conditions but are otherwise random). The code is thus an emergent feature of the neural population that can only occur via coordinated activity between the neurons involved in controlling the upper limb. These findings shed light on the neural mechanisms underlying motor control and have significant implications for the development of advanced BMI-controlled neuroprosthetics for individuals with paralysis.