A Computational Approach for the Design of Epidural Electrical Spinal Cord Stimulation Strategies to Enable Locomotion after Spinal Cord Injury

Spinal cord injury (SCI) is a major cause of paralysis with currently no effective treatment. Epidural electrical stimulation (EES) of the lumbar spinal cord has been shown to restore locomotion in animal models of SCI, but has not yet reached the same level of efficacy in humans. The mechanisms through which EES promotes locomotion, and the causes underlying these inter-species differences remain largely unknown, although essential to fully exploit the therapeutic potential of this neuromodulation strategy. Here, we addressed these questions using a deductive approach based on computer simulations and hypothesis-driven experiments, and proposed complementary strategies to enhance the current efficacy of EES-based therapies.

In the first part of this thesis, we studied the mechanisms through which EES enables locomotion in rat models of SCI. Performing simulations and behavioral experiments, we provided evidence that EES modulates proprioceptive afferents activity, without interfering with the ongoing sensory signals. We showed that this synergistic interaction allows muscle spindle feedback circuits to steer the unspecific excitation delivered by EES to functionally relevant pathways, thus allowing the formation of locomotor patterns. By leveraging this understanding, we developed a stimulation strategy that allowed adjusting lesion-specific gait deficits, hence increasing the therapeutic efficacy of EES.

In the second part of this thesis, we evaluated the influence of trunk posture on proprioceptive feedback circuits during locomotion, and thus on the effect of EES, in rat models of SCI. By combining modeling and experiments, we showed that trunk orientation regulates leg proprioceptive signals, as well as the motor patterns produced during EES-induced stepping. We exploited these results to develop a control policy that by automatically regulating trunk orientation significantly enhanced locomotor performance.

In the last part of this thesis, we investigated the causes underlying species-specific effects of EES. Hypothesis-driven simulations suggested that in humans continuous EES blocks the proprioceptive signals traveling along the recruited fibers. We corroborated this prediction by performing experiments in rats and people with SCI. In particular, we showed that EES disrupts the conscious perception of leg movements, as well as the afferent modulation of sensorimotor circuits in humans, but not in rats. We provide evidence that in humans, due to this phenomenon, continuous EES can only facilitate locomotion to a limited extent. This was insufficient to provide clinically relevant improvements in the tested participants. Finally, we proposed two sensory-compliant stimulation strategies that might overcome these limitations, and thus augment the therapeutic efficacy of EES.

In this thesis we elucidated key mechanisms through which EES promotes locomotion, we exposed critical limitations of continuous EES strategies when applied to humans