Electrical Stimulation Induced Current Distribution in Peripheral Nerves Varies Significantly with the Extent of Nerve Damage: A Computational Study Utilizing Convolutional Neural Network and Realistic Nerve Models.

Electrical stimulation of the peripheral nervous system is a promising therapeutic option for several conditions; however, its effects on tissue and the safety of the stimulation remain poorly understood. In order to devise stimulation protocols that enhance therapeutic efficacy without the risk of causing tissue damage, we constructed computational models of peripheral nerve and stimulation cuffs based on extremely high-resolution cross-sectional images of the nerves using the most recent advances in computing power and machine learning techniques. We developed nerve models using nonstimulated (healthy) and over-stimulated (damaged) rat sciatic nerves to explore how nerve damage affects the induced current density distribution. Using our in-house computational, quasi-static, platform, and the Admittance Method (AM), we estimated the induced current distribution within the nerves and compared it for healthy and damaged nerves. We also estimated the extent of localized cell damage in both healthy and damaged nerve samples. When the nerve is damaged, as demonstrated principally by the decreased nerve fiber packing, the current penetrates deeper into the over-stimulated nerve than in the healthy sample. As safety limits for electrical stimulation of peripheral nerves still refer to the Shannon criterion to distinguish between safe and unsafe stimulation, the capability this work demonstrated is an important step toward the development of safety criteria that are specific to peripheral nerve and make use of the latest advances in computational bioelectromagnetics and machine learning, such as Python-based AM and CNN-based nerve image segmentation.

Color and cellular selectivity of retinal ganglion cell subtypes through frequency modulation of electrical stimulation.

Epiretinal prostheses aim at electrically stimulating the inner most surviving retinal cells-retinal ganglion cells (RGCs)-to restore partial sight to the blind. Recent tests in patients with epiretinal implants have revealed that electrical stimulation of the retina results in the percept of color of the elicited phosphenes, which depends on the frequency of stimulation. This paper presents computational results that are predictive of this finding and further support our understanding of the mechanisms of color encoding in electrical stimulation of retina, which could prove pivotal for the design of advanced retinal prosthetics that elicit both percept and color. This provides, for the first time, a directly applicable "amplitude-frequency" stimulation strategy to "encode color" in future retinal prosthetics through a predictive computational tool to selectively target small bistratified cells, which have been shown to contribute to "blue-yellow" color opponency in the retinal circuitry. The presented results are validated with experimental data reported in the literature and correlated with findings in blind patients with a retinal prosthetic implant collected by our group.

In Vivo Magnetic Stimulation of Rat Sciatic Nerve With Centimeter- and Millimeter-Scale Solenoid Coils.

Previous reports of magnetic stimulation of the peripheral nervous system (PNS) used various coil geometries, all with outer diameters larger than 35 mm, and stimulation energies in the 50 J range to evoke neural excitation. Recent reports of central nervous system (CNS) activation used sub-mm-scale solenoid coils with mJ energy levels. The goal of this study was to translate the lower energy levels from the CNS to the PNS via using smaller coils placed in closer proximity to the neural tissue. Such a performance improvement would advance the state of the art of magnetic stimulation and provide a path towards new neuroprosthetic devices. Primarily, we investigated the range of coil outer diameters from 25 mm down to 5 mm to better understand the dependence of coil diameter on energy required for PNS activation. Nine cm- and mm-scale copper solenoid coils, with various resistances, inductances, inner and outer diameters, and heights were compared by quantizing neuromuscular responses to magnetic stimulation via capacitive discharge excitation of rat sciatic nerves in vivo. Additionally, the effects of stimulus duration and coil position were investigated. As opposed to prior work, this study compares a subset of stimulation parameters in an intact nerve preparation, and shows that magnetic stimulation with coils that abut the nerve is a reliable, effective method of neuromuscular stimulation. Although we observed different energies required for neuromuscular activation depending on the coil and excitation parameters used, for the experimental configuration, devices, and stimulus waveform shapes presented in this manuscript, no systematic dependence of PNS activation on coil diameter was found, even for the mm-scale coils investigated herein. However, there was a clear relationship between discharge circuit capacitance and energy required to evoke a neuromuscular response. Coils approximately 12 mm in outer diameter and larger consistently evoked responses, whereas coils 5 mm in outer diameter did not. Furthermore, we observed meaningful neuromuscular excitation when stimulating with energies as low as 20 J. Although this is an improvement over prior work, it is still orders of magnitude greater than the energy required for conventional electrical stimulation, suggesting that these devices are presently not suitable for use in an application requiring continued pulsed stimulation. Nevertheless, these devices are suitable for basic research and as clinical tools that infrequently stimulate, such as in diagnostic applications.

A µm-Scale Computational Model of Magnetic Neural Stimulation in Multifascicular Peripheral Nerves.

There has been recurring interest in using magnetic neural stimulation for implantable localized stimulation. However, the large stimulation voltages and energies necessary to evoke neuronal activity have tempered this interest. To investigate the potential of magnetic stimulation as a viable methodology and to provide the ability to investigate novel coil designs that can result in lower stimulation threshold voltages and energies, there is a need for a model that accurately predicts the magnetic field-tissue interaction that results in neuronal stimulation. In this study, we provide a computational framework to accurately estimate the stimulation threshold and have validated the model with in vivo magnetic stimulation experiments. To make such predictions, we developed a micrometer-resolution anatomically driven computational model of rat sciatic nerve and quantified the effect of tissue heterogeneity (i.e., fascicular organization, axon distribution, and density) and axonal membrane capacitance on the resulting threshold. Using the multiresolution impedance method, we computed the spatial-temporal distribution of the induced electric field in the nerve and applied this field to a Frankenhaeuser-Huxley axon model in NEURON to simulate the nonlinear mechanisms of the membrane channels. The computational model developed predicts the stimulation thresholds for four magnetic coil designs with different geometrical parameters within the 95% confidence interval (experiments count = 4) of measured in vivo stimulation thresholds for the rat sciatic nerve.

Magnetic stimulation of mammalian peripheral nerves in vivo: an alternative to functional electrical stimulation.

Functional electrical stimulation is the current gold standard for stimulating neuronal interfaces for functional neuromuscular and cortical applications, but it is not without its drawbacks. One such fault is the need to have direct electrical contact with the nerve tissue, and any side effects this causes. Functional magnetic stimulation, which works though electromagnetic induction, does not require electrical contact and may be a viable alternative to functional electrical stimulation. We are investigating the capabilities of magnetic stimulation with centimeter scale (< 2.5 cm) coils in feline and rodent sciatic nerves in vivo. We have shown that magnetic stimulation can consistently produce the same levels of neuromuscular activation as electrical stimulation. Additionally, the position of the coil relative to the nerve influences neuromuscular activation, suggesting the possibility of selective muscle activation.

Modeling and percept of transcorneal electrical stimulation in humans.

Retinal activation via transcorneal electrical stimulation (TcES) in normal humans was investigated by comparing subject perception, model predictions, and brain activation patterns. The preferential location of retinal stimulation was predicted from 3-D admittance modeling. Visual cortex activation was measured using positron emission tomography (PET) and (18)F-fluorodeoxyglucose (FDG). Two different corneal electrodes were investigated: DTL-Plus and ERG-Jet. Modeling results predicted preferential stimulation of the peripheral, inferior, nasal retina during right eye TcES using DTL-Plus, but more extensive activation of peripheral, nasal hemiretina using ERG-Jet. The results from human FDG PET study using both corneal electrodes showed areas of visual cortex activation that consistently corresponded with the reported phosphene percept and modeling predictions. ERG-Jet was able to generate brighter phosphene percept than DTL-Plus and elicited retinotopically mapped primary visual cortex activation. This study demonstrates that admittance modeling and PET imaging consistently predict the perceived location of electrically elicited phosphenes produced during TcES.

Two-dimensional SPICE-linked multiresolution impedance method for low-frequency electromagnetic interactions.

A multiresolution impedance method for the solution of low-frequency electromagnetic interaction problems typically encountered in bioelectromagnetics is presented. While the impedance method in its original form is based on the discretization of the scattering objects into equal-sized cells, our formulation decreases the number of unknowns by using an automatic mesh generation method that does not yield equal-sized cells in the modeling space. Results indicate that our multiresolution mesh generation scheme can provide a 50%-80% reduction in cell count, providing new opportunities for the solution of low-frequency bioelectromagnetic problems that require a high level of detail only in specific regions of the modeling space. Furthermore, linking the mesh generator to a circuit simulator such as SPICE permits the addition of arbitrarily complex passive and active circuit elements to the generated impedance network, opening the door to significant advances in the modeling of bioelectromagnetic phenomena.