Engineering Direct Electrical Stimulation of Human Sensorimotor Cortex

Abstract

Damage to the nervous system due to stroke, spinal cord injury, and limb loss leads to significant sensory and motor deficits. Despite the variety of reasons behind cortical injury, targeted direct electrical stimulation (DES) may be able to help restore both motor and sensory function in future neuroprosthetic applications. Our research addresses fundamental barriers through a principled engineering approach for translation of DES to humans. We primarily focused on the physics of stimulation, the signal processing of these concurrent stimulation and recording signals, how sensory stimulation compares to natural touch, and how to induce plasticity to modify cortical connectivity through DES. First, we modeled how electrical stimulation propagates through the tissue with analytic models, and found that analytic models with flat and spherical geometries fit the measured voltages on the surface well. We also discovered the importance of the assumptions about the overall geometry of the brain in these flat and spherical models and how these affect the subsequent interpretations of the resistivity of the brain tissue. These results have implications for the appropriate ways to make resistivity measurements in-vivo on the human brain, and establishes benchmarks for comparisons with more computationally intensive finite-element models. We subsequently developed novel unsupervised algorithms to extract the neural response to DES with concurrent stimulation validated on multiple human datasets, allowing for analyses of neural data contaminated by stimulation artifacts. This work will aid in the interpretation of data from experiments involving stimulation in humans, and we have made public all of the code and data for future algorithmic development and comparisons. We discovered that humans respond more slowly to DES of sensory parts of the brain relative to natural peripheral touch, which may require future neuroprosthetic devices to account for this delay to allow for smooth and intuitive control. Concurrent DES of sensory parts of the brain and natural peripheral touch are perceived independently, which suggests that local cortical circuitry is not completely inhibited by DES and will allow for overlapping inputs in future neuroprosthetic devices. Furthermore, modified DES waveforms during sensory stimulation can modify percepts and reaction times, suggesting that better neuroprosthetic performance could be achieved by moving beyond the constant amplitude waveforms. We tested two stimulation protocols to induce plasticity in human cortex, with potential applications in helping the brain heal after stroke. The first involved beta-oscillation driven stimulation, where we found short term (less than two seconds in duration) enhancement of cortically evoked potentials. This was driven by a dominant effect between the number of conditioning stimuli delivered and the size of the evoked potential following stimulation. Delivering stimuli during the depolarizing surface potential phase appeared to have an potentiating interaction effect during trains where the greatest number of conditioning stimuli were delivered. The second was a paired pulse stimulation paradigm carried out intraoperatively in subjects undergoing deep brain stimulator placement, where the optimal lag to induce plasticity as assessed through evoked potentials was longer than that predicted by the principles of spike timing dependent plasticity. We similar found that paired site stimulation was more effective than stimulation within the same site, suggesting that both the timing and location of stimulation are critical for the induction of plasticity. These two stimulation protocols lay the groundwork for future clinical trials in humans using these protocols to try and induce plasticity in damaged regions of cortex. Our research in these areas of engineered stimulation of sensorimotor cortex in humans with implanted electrodes moves DES closer towards clinical rehabilitation use for sensory neuroprosthetics and neuromodulation to induce cortical plasticity.