Personalizable computational models of electrical spinal cord stimulation to restore function after neurological disorders

Spinal Cord Injury (SCI) disrupts the communication between the brain and spinal circuits below the lesion, leading to a plethora of neurological impairments, including the loss of motor function. At present, the only medical practices to enhance recovery after paralysis are activity-based therapies. In severe cases, SCI-patients fail to produce active movements voluntarily and are therefore unable to fully benefit from these therapies. Epidural Electrical Stimulation (EES) applied over the lumbosacral spinal cord has demonstrated to enable locomotion in preclinical animal models and humans with SCI. EES may therefore play a pivotal role in the recovery of locomotion after SCI.However, neither has EES reached the same level of efficacy in humans as in preclinical animal models, nor is there consensus on the definition of clinically meaningful stimulation protocols. In the absence of clear understanding on how EES promotes the formation of motor patterns, stimulation protocols have been largely based on empirical observations. Instead, hypothesis-driven definition of stimulation parameters has demonstrated superior efficacy in other neuromodulation therapies, such as deep brain stimulation.In this thesis, I outline a computational approach to inform EES strategies for the recovery of neurological function after SCI. For this purpose, I leveraged a combination of finite element methods, neurophysical and neuromusculoskeletal models to refine our understanding on how EES promotes locomotion after SCI, inform the design of neuromodulation technologies and develop personalized EES treatments in humans with SCI.In the first part of this thesis, I provide evidence that EES recruits proprioceptive and cutaneous afferents in the dorsal roots. I suggest that the recruitment of proprioceptive feedback circuits promotes the formation of motor patterns. In contrary, the continuous recruitment of cutaneous afferents may disrupt the production of locomotion. In the next chapter, I study interspecies difference of EES. I argue that these differences necessitate the adaption of stimulation protocols to enable translation of EES as a therapy from preclinical models to humans. I propose the application of spatiotemporal protocols with low-amplitude, high-frequency bursts to improve the efficacy of EES in humans. In the following chapter, I describe a technological framework that enables the implementation of such stimulation paradigms in people with SCI. I outline how personalized computational modeling can be used for treatment planning. I then summarize how I advanced this rudimentary computational tool into a computational framework for the rapid elaboration of highly realistic personalizable computational models. I explain how I leveraged this framework to inform the arrangements of electrodes on a paddle lead to enable locomotion in large and diversified patient-cohorts, perform preoperative treatment planning and validate the efficacy of both the lead and treatment planning protocol in individuals with severe, chronic SCI. Finally, I summarize how I translated my computational framework for other use-cases including the control of hemodynamics and the evaluation of an optoelectronic system to enable locomotion in preclinical models.

Brain-Controlled Neuroprosthetic Therapies And Mechanisms Of Recovery After Spinal Cord Injury

Galyna Pidpruzhnykova

Spinal cord injury (SCI) is the second leading cause of paralysis, which induces abrupt impairments with devastating consequences for the quality of life of patients. Despite significant progress in SCI treatment through neurorehabilitation training and spinal cord stimulation (SCS) techniques, there is still no cure for SCI and the mechanisms of recovery are poorly understood. My thesis represents a series of studies which investigate the neural mechanisms underlying locomotor recovery after SCI and leverage these findings for the development of brain-controlled neuroprosthetic therapies to enhance the potential of neurorehabilitation therapies. The first study focused on investigating neural mechanisms underlying locomotor recovery after clinically relevant contusion SCI. During this study, we assessed changes in over a hundred kinematic parameters using gait pattern extraction and principal component analysis. Our results show that the animals learned to step even in the absence of electrochemical enabling factors. Remarkably, we observed a carryover effect of locomotor recovery to the previously unpracticed swimming task. Afterwards, we dissected the neural mechanisms underlying the observed behavioral effect using viral tracing techniques and a double-virus construct allowing targeted pathway inactivation. As a result, we were able to reversibly abolish stepping capacity of rehabilitated rats, suggesting the critical importance of the reticulospinal tract for the recovery from contusion SCI. These findings gave us the rationale for the second study in which we developed a novel rehabilitation paradigm of brain-controlled stimulation of the mesencephalic locomotor region (MLR). The MLR sends excitatory descending input to reticulospinal neurons in the brainstem, which relay the descending locomotor command through the tissue bridge spared after contusion SCI. Thus, we hypothesized that MLR stimulation may lead to improved locomotion in rats after SCI when combined with neurorehabilitation training and electrochemical SCS. We found that brain-controlled MLR stimulation was able to improve kinematic output in rehabilitated animals and to significantly reduce side effects of non brain-controlled MLR stimulation. Thus, this novel neuroprosthetic paradigm is capable of enhancing existing rehabilitation techniques. However, the long-term effectiveness of this intervention during SCI rehabilitation needs to be further investigated to evaluate its full therapeutic benefit. Finally, after evaluating the possible beneficial and adverse effects of MLR stimulation in a translational context, we proposed another type of brain-controlled stimulation that directly bypasses the lesion site through an electronic bridge between the brain and the spinal cord. For this, we created a neuroprosthetic system allowing for direct cortical control of spinal cord neuromodulation during gait rehabilitation. This refinement of SCS delivery significantly improved both immediate locomotor output and the long-term recovery of rats after SCI. In conclusion, this thesis uncovers mechanisms of recovery from SCI and demonstrates the therapeutic potential of brain-controlled neuroprosthetic therapies. Furthermore, it raises important questions and identifies challenges in translating these therapies from the bench to the bedside. Hence, this work constitutes a valuable input for both basic science and the development of clinically relevant interventions.

EPFL2017

Brain-spine interface technologies to alleviate gait deficits after Parkinson's disease

Flavio Raschella

Gait and balance deficits are motor disabilities that typically develop late in the progression of Parkinson's disease (PD). These deficits are thought to be the result of dysregulated neural activity in the cortico-basal ganglia-thalamic-cortical motor loop, caused by the depletion of dopaminergic and cholinergic circuits. PD patients exhibit gait abnormalities characterized by postural instability and reduced locomotor control, which diminish their functional independence and strongly affect their quality of life. This is further aggravated by the onset of falling events. These deficits represent a therapeutic challenge as they typically respond poorly or not at all to standard therapies, such as dopaminergic medication and deep brain stimulation (DBS). Whereas the efficacy of DBS in treating cardinal symptoms of PD, such as tremor, rigidity and bradykinesia is well established, effects on gait and balance deficits are still controversial. With disease progression the majority of PD patients develop a resistance to DBS treatment, which results in a significant worsening of gait and postural stability, while cardinal symptoms remain partially stable. In the last decade, Epidural Electrical Stimulation (EES) of the spinal cord has been proposed as a possible solution for improving locomotor performance in people with PD. Initial results on rodent and marmoset models of PD were promising. In these studies, EES was delivered continuously at the thoracic level of the spinal cord, without taking into consideration the locomotor state. However, translation to human patients led to conflicting results. In parallel, studies in spinal cord injury (SCI), in both animal models and patients, have drawn attention to the use of spatially and temporally specific stimulation protocols that independently engage flexion and extension muscles synergies. This strategy has allowed unprecedented progress, enabling full weight-bearing and restoration of locomotion after only a few days of application. In order to promote voluntary control of locomotion after SCI, a first brain-spine interface (BSI) was engineered to link the intended motor state to the EES protocols, re-establishing the altered communication between the brain and the spinal cord. In this thesis, I show that gait and balance deficits developed in the non-human primate (NHP) 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model of PD are similar to those observed in PD patients. Next, I describe the design of spatiotemporal EES protocols using the activation of the spinal motor neurons. I subsequently show that, although MPTP treatment drastically reduces the number of dopaminergic neurons in the substantia nigra, it has no detectable impact on the cortical dynamics underlying locomotion. Furthermore, I demonstrate that a BSI immediately alleviates gait and balance deficits in an NHP MPTP model of PD. Finally, I investigate the interaction between DBS and BSI in a MPTP-treated NHP. Although DBS increases the alertness and locomotor activity of the NHP, it has no effect on kinematic parameters during gait. The BSI, instead, alone or in combination with DBS, improves locomotor performance such that it approaches healthy locomotion. Therefore, I demonstrate that the BSI and DBS can be simultaneously applied, opening up the possibility of a combined therapeutic approach in human patients. As such, the subsequent step should be to test the efficacy of the BSI in treating PD patients.

EPFL2019

Brain-Controlled Neuroprosthetic Therapies And Mechanisms Of Recovery After Spinal Cord Injury

Galyna Pidpruzhnykova

EPFL2017

Brain-spine interface technologies to alleviate gait deficits after Parkinson's disease

Flavio Raschella

EPFL2019