Computational Analysis of the Neurophysiological Effects of Spinal Cord Stimulation

Abstract

Chronic pain is a major public health burden and cause of global human suffering. Spinal cord stimulation (SCS) is a common neurostimulation treatment option for patients with refractory chronic pain. Unfortunately, despite decades of clinical experience, only about 60% of patients successfully respond to the therapy (typically defined as a reduction of pain by at least 50%), and many patients report loss of efficacy over time. This inadequate success rate is largely because very little is known about the neurophysiological effects of SCS and its analgesic mechanisms of action. This dissertation describes a computational modeling approach to investigate the effects of SCS on the nervous system. The first study analyzes the most fundamental question: which neurons are directly driven to fire action potentials by various available SCS waveforms including conventional, burst, and 10-kHz SCS? We used a finite element model of the lower thoracic spinal cord to assess the spatial electric potential generated by clinical SCS systems, which was then paired with biophysical multi-compartment models of the relevant neural populations. This study provides insights into the neural recruitment order during clinically relevant SCS, as well as the effects of various factors such as axonal collateralization on activation thresholds. One important finding is that all waveforms produced the same neural order of activation, albeit at different amplitudes. Study two expands this analysis to evaluate recently proposed mechanisms for novel SCS modalities beyond simple activation of large diameter myelinated axons. Specifically, we evaluated potential alterations in spike timing in afferent fibers, altered dorsal horn neuron excitability due to membrane polarization, and selective activation of unmyelinated C fibers. We also evaluated the effects of including stochastic ion channel properties on the dorsal column fiber response to stimulation. Overall, our results refute several proposed mechanisms of action and highlight the value of including stochastic ion channels in models of extracellular stimulation to produce realistic variability in firing responses. Finally, the third study comprehensively evaluated the effects of stimulation frequency on the neural response to SCS, including effects on dorsal column fiber activation thresholds, synaptic transmission in the brainstem, action potential fidelity in primary afferent collateral arbors, and the output of the dorsal horn pain processing network. We found that high frequency stimulation of at least 100 Hz reduced activation thresholds (mirroring clinical paresthesia perception thresholds), produced asynchronous firing in the brainstem, and promoted branch point failure within the branching collateral arbors of the dorsal horn. Overall, this dissertation contributes to our understanding of the neural response to clinically relevant SCS. The results presented in this study clarify which neurons are most likely to be activated during SCS as well as higher order properties such as firing properties during suprathreshold stimulation and synaptic processing. Ultimately, these insights will help guide developing future SCS systems to optimize pain relief while minimizing power consumption.