Dissertation Title: Building a digital nervous system: Bioengineering neuromuscular interfaces for biohybrid closed-loop brain-body neural prostheses

Abstract: 

Neural prostheses have demonstrated the ability to artificially control the nervous system towards neurological recovery. However, the current neuroprosthetic paradigm relies solely on synthetic electronic components for neural control, limiting precise neural targeting and hence hindering therapeutic impact. On the stimulation front, the standard technique to artificially stimulate neuromusculature, functional electrical stimulation (FES), cannot selectively activate neural structures, resulting in poor control and rapid fatigue, limiting chronic neuromodulation therapies and implantable organ actuation. On the sensing front, obtaining high-fidelity real-time neuromuscular states such as force, which are critical for closed-loop neural prostheses, remains elusive. In an alternative paradigm described in this dissertation, neuromuscular components are engineered at different scales, from the molecular to the organ scale, and interfaced with electronic components, to achieve augmented stimulation, sensing, and closed-loop control capabilities, realizing the potential of chronic closed-loop neuromodulation of brain-body circuits. Spanning genetic, regenerative, and bioelectronic approaches for artificial neural stimulation, implantable magnetic systems for wireless neuromuscular sensing, and model-based closed-loop control policies, this dissertation builds a platform for the digital control of organs to reestablish brain-body communication in neurological conditions.

In the first part of the dissertation, we present an optogenetic system that shows motor units can be recruited naturally for force production. Leveraging this mechanism, we designed a closed-loop controller based on a biophysical neuromuscular model that recapitulates optogenetically-stimulated muscle dynamics. The optogenetic system enabled continuous closed-loop control of skeletal muscle with high fidelity and without inducing fatigue. To advance the translational potential of peripheral optogenetic therapies, in the second part, we present a minimally immunogenic transduction strategy to express optogenetic molecules in the periphery for long-term neuromodulation. We show that direct neural transduction enables chronic optogenetic expression in peripheral nerves, by delaying the humoral immune response and minimizing pro-inflammatory, opsin-specific T-cell-mediated cellular immunity. In the third part, we present the design of an implantable biohybrid actuator based on a regenerative approach to engineer motor recruitment. We demonstrate that sensory neurons can establish cholinergic synapses with muscle fibers and the axonal architecture of sensory nerves normalizes activation thresholds. Leveraging this mechanistic discovery, we show fatigue-resistant control of the actuator under FES, enabling the design of biohybrid organ systems to modulate neural afferents and organ mechanics. For the fourth part, we present a minimally-invasive wireless sensing modality that estimates neuromechanical force by tracking tendon dynamics in real-time. The system combines chronically implanted magnetic beads in tendon tissue with a skin-mounted magnetometer sensing array. The system provided stable high-resolution tendon dynamics, which captured gait-phase and speed-dependent loading during walking and exhibited strong correlation with ground reaction forces. We also present a system for wireless detection of muscle activation by measuring magnetic flux changes resulting from muscle vibrations of an implanted magnetic bead. The system captures neuromechanical firing rates and is able to effectively measure muscle activation in a contactless manner. For the fifth part, we present a fully-implantable peripheral bioelectronic platform for chronic wireless sensing, stimulation, and adaptive closed-loop control of individual muscle neuromechanics. The platform resolves fundamental neuromechanical states including twitch, recruitment, and firing-rate dynamics. We demonstrate submillimeter-accuracy closed-loop control of ankle dorsiflexion and knee flexion under sustained and cyclic reference trajectories, as well as precise modulation across changing dynamic mechanical demands.

This dissertation presents contributions to: 1) an optogenetic system that solves the longstanding problem of natural control of skeletal muscle using artificial stimulation, 2) a minimally immunogenic strategy to express optogenetic molecules in the periphery for chronic neuromodulation, 3) a regenerative tissue construct that serves as a fatigue-resistant actuator for implantable control of organ systems, 4) two minimally-invasive systems for wireless neuromechanical force and activation sensing, and 5) a clinically-ready bioelectronic platform for muscle-specific neuromechanical control. Altogether, this work lays the groundwork for high-performance closed-loop neural prostheses and bioelectronic therapies for chronic neuromodulation of brain-body circuits.

Committee members: 
Prof. Hugh Herr, Ph.D.
Professor of Media Arts and Sciences
Massachusetts Institute of Technology 

Prof. Edward Boyden, Ph.D.
Y. Eva Tan Professor in Neurotechnology
Massachusetts Institute of Technology

Prof. James Guest, M.D., Ph.D.
Professor of Neurological Surgery
University of Miami Neurosurgery and the Miami Project to Cure Paralysis