Final Term Paper ixk51

Selective Interfascicular Neural Recording

Ilya Kolb

Grant proposal

Specific Aims

Since peripheral nerves carry many different signals from the brain to limbs and organs, closed-loop devices such as advanced prosthetics and functional electrical stimulation (FES) systems can use neural signal as feedback for control. Selective recording, or the ability to record from a specific fascicle without picking up signal from other fascicles, is therefore vital for this purpose. There is literature available about selective extra-neural cuff electrodes as well as intra-fascicular microelectrode arrays, but no published record exists of using interfascicular electrodes (ones that reside in the epineurium) for the purpose of recording. The need is evident for an electrode that does not invade the fascicle (in order to prevent neuronal damage) and is still able to selectively record neural signals to be useful as a high-fidelity control signal. Our hypothesis is that interfascicular electrodes are selective and safe for the purpose of neural recording. Testing the hypothesis will be accomplished by addressing the specific aims outlined below. First, a mathematical model will be constructed to determine the theoretical effectiveness of inter-fascicular electrodes (Aim 1). Second, selectivity results from acute experiments will be compared to the mathematically-obtained selectivity measures to validate the model and prove the short-term safety and efficacy of the electrodes (Aim 2). Finally, a set of chronic experiments will be performed to evaluate long-term performance of the electrode and ultimately pave the way for human trials (Aim 3). These experiments will serve as a comprehensive quantification of the design space available for the purpose of interfascicular neural recording.

Aim 1: To validate the claim that inter-fascicular electrodes can selectively record signals from separate fascicles using a finite element model a. Determine the relationship between the number of inter-fascicular contacts and the selectivity of the electrodes. b. Find the maximum distance that a contact can be away from a target fascicle and still selectively record from it. c. Evaluate the effect of adding more fascicles and inhomogeneities (blood vessels, connective tissue) on the selectivity of the electrodes.

Aim 2: To quantify the recording selectivity of interfascicular electrodes in animal models a. Develop system for interfascicular electrode implantation in vivo b. Find the minimum number of inter-fascicular contacts required for the selectivity to approach a maximum. c. Record spontaneous neural activity to evaluate the signal-to-noise ratio of inter-fascicular electrodes. d. Review the histology of nerves to confirm electrode placement within the epineurium and rule out acute neuronal damage.

Aim 3: To determine the viability of using interfascicular electrodes in chronic animal studies a. Evaluate the selectivity of the inter-fascicular electrode as a function of time within a 10-week observation period. b. Determine the signal-to-noise ratio of the electrodes within a 10-week observation period to evaluate deterioration in signal quality.

Research strategy: Significance

Figure 1 (Source: Modified from Sergi et al. 2006)

Importance of problem: A selective, high fidelity neural signal can serve as an indicator of neuronal activity, which can be used in a multitude of control and therapy applications

A stable chronic interface between the nervous system and man-made electronics allows us to “listen” to neuronal activity. It is impossible to fully appreciate all the potential applications of peripheral nerve recording since some, if not most of them, have yet to be invented; however, there are several particular applications which demonstrate a clear need for advances in this technology.

The prosthetic limb industry has been around for centuries in one form or another, yet the idea of using “smart” prosthetics, ones that mimic actual human motion, is relatively new. In order to ensure that a prosthetic device is responsive to the user’s movement wishes, there must be a command signal to drive the motor(s) in the artificial joint. This signal may be obtained from extra-corporeal sources such as sensor/switch systems, as well as from the body’s natural cortical, muscle-based, or peripheral nerve activities. Using external sensors and switches, while plausible, poses problems such as donning, doffing, calibration, and lack of space. Of the aforementioned biologically inspired options, detecting activity from the nerve that previously innervated the lost limb is the most intuitive interface method. It taps into the control signal of the muscles themselves, such that neural information that previously went to the patient’s missing limb is now fully accessible. An ideal (yet currently unavailable) neural interface would pick out signals from individual neurons, which would enable prosthetic devices to fully mimic human motion using an artificial limb.

Peripheral neural recording has been shown to be potentially useful in several applications involving controlling FES devices (Popovic et al. 1993), more specifically electrical stimulation to treat obstructive sleep apnea (OSA) (Sahin, Durand, and Haxhiu 1999), foot drop (M. K. Haugland and Sinkjaer 1995), hand grasp (M. Haugland et al. 1997) and others (see (Hoffer et al. 1996) for a review of applications).

Barrier to overcome: There is currently no optimized minimally invasive electrode system for the purpose of selectively recording ENG (electroneurograms).

In FES and prosthetic systems with a large number of degrees of freedom (DOF), controlling each DOF is crucial for functionality. In such a case, it becomes necessary to control each DOF with its own command signal, which could be accomplished by implanting multiple electrodes, which is overly invasive due to the need for many incisions, overabundance of wires, and overall complexity of the system. One of the greatest challenges of modern neural recording systems is developing a single electrode that could discriminate between the firing patterns of different parts of the nerve, obviating the need for multiple implantations.

The ability of an electrode to discriminate between signals from various sources is defined as its selectivity, depicted on the x-axis in Figure 1. Unsurprisingly, selectivity increases proportionally with invasivity since electrodes that are physically closer to neurons are better able to pick up their activity but are also more invasive. Invasivity in this context is defined as the measure by which neural tissue is negatively affected due to implantation. Man-made devices that closely interact with neurons are therefore more invasive than those that are physically further away.

The challenge has been to optimize the interface in such a way that selectivity is maximized while invasivity (and therefore damage) is minimized (labeled “Ideal electrode” in Figure 1). Currently, no electrode exists that claims to cause no damage to neurons and at the same time records with superior selectivity, though several electrodes can be placed at different points along the selectivity-invasivity line (Figure 1) (Navarro et al. 2005)(Yoo and Durand 2005)(S. Micera et al. 2011)(Yoshida, Hennings, and Kammer 2006)).

Importance in the field: The novel electrode design may challenge the paradigm that selectivity and invasivity are directly related.

This project will address the challenge of creating a selective yet relatively non-invasive interface (closer to the ideal electrode on the Selectivity-Invasivity plot in Figure 1) by proposing an interfascicular electrode, a novel interface which has been sparingly used for stimulation (Tyler and Durand 1997)(Nielsen, Sevcencu, and Struijk 2011) and has not been documented for recording applications. The goal of the current project is to show that this design is minimally invasive and at the same time maximally selective (represented by “Interfascicular?” in Figure 1). This study will quantify the short-term and long-term performance of interfascicular electrodes by assessing three clinically relevant metrics: selectivity, signal-to-noise ratio (SNR), and neuronal damage. In addition, it will elucidate a reliable method of wire insertion into the epineurium, which would be beneficial for purely surgical purposes as well. If successful, this electrode could serve as a viable recording interface between the peripheral nervous system and man-made devices. The development of this technology could ultimately improve the control mechanism for patients using FES systems or prosthetic devices, which would ultimately lead to a lower level of functional impairment and an improved standard of living. As more devices are made that can utilize neural signals, using interfascicular electrodes may become a standard practice for those applications as well.

Approach

The proposed project will serve as a method of quantifying the recording capabilities of interfascicular electrodes, a novel method of interfacing with the nervous system. While the procedures outlined below are presented in increasing order of complexity (mathematical model to acute experiments to chronic experiments), the aims can be performed in parallel, such that a temporary setback in achieving one specific aim will not hinder the progress of another.

Figure 2 FEM model of a nerve containing 5 fascicles. Interfascicular electrodes are in the epineurium, and have an insulated surface (black) and a contact surface (white). (Source: Modified from (Perez-Orive and Durand 2000))

Aim 1: To validate the claim that inter-fascicular electrodes can selectively record signals from separate fascicles using a finite element model

Several models of peripheral nerves have been created for the purpose of elucidating a mathematical relationship between a specific type of electrode and its ability to selectively record from (Perez-Orive and Durand 2000; Yoo and Durand 2005) or stimulate (Veltink, van Alste, and Boom 1988; Tyler, Peterson, et al. 2011) different fascicles. These models are primarily based on the groundbreaking work of Hodgkin and Huxley (Hodgkin and Huxley 1952) and the core conductor model (or cable theory) which has been improved and adapted through the years to include myelin (McNeal 1976).

Methodology

Creating mathematical model of nerve and electrodes

A finite-element model (FEM) of various shapes of nerves will be constructed uses Ansoft (ANSYS Inc.; Canonsburg, PA). The electrical properties of the neurons will be modeled using a double-cable model, as used in (McIntyre, Richardson, and Grill 2002) to account for the properties of myelin. The neurons will be bathed in low-impedance solution, corresponding to that of the endoneurium. A high-impedance perineurium will separate fascicular space from the endoneurium and the fascicles will be suspended in a relatively high-impedance solution corresponding to the epineurium. The specific impedance values of neural tissue will be obtained from literature, primarily (Geddes and Baker 1967; Choi, Cavanaugh, and Durand 2001). Finally, an insulating cuff electrode (modeled as a silicone sheath) will be circumscribed around the nerve and interfascicular contacts will be placed into the epineurium (Figure 2). Neural activity will be recorded using a tripolar electrode configuration, with the outer contacts of the tripole shorted together on the outside of the nerve and the inner contact being recorded using the interfascicular electrode.

Evaluating selectivity of various electrode configurations

A selectivity index, as defined in (Perez-Orive and Durand 2000) will be computed for various electrode locations and nerve shapes. For instance, selectivity will be evaluated as a function of distance from the fascicle of interest and as a function of the number of interfascicular contacts in the nerve. Nerve geometry will be taken into account as well, since a reshaping cuff electrode will be used in the study; circular and flat nerve cross-sections will be tested, and the selectivity in both geometries will be compared.

Adding inhomogeneities, modeling in a non-ideal environment

The final step in the modeling study will involve removing the homogeneity constraints in the model and addressing a more physiologically-relevant case, where fascicles are spread out unevenly across the epineurium and other longitudinal structures (such as blood vessels) are also present. The simulations will be run several times with randomized non-homogeneity conditions to ensure that the electrode does not rely on a perfect symmetry to achieve selectivity.

Expected outcomes, potential problems, risks

Interfascicular electrodes have been shown to be highly selective in stimulation experiments in finite-element models (Tyler, Peterson, et al. 2011). Since such electrodes can be implanted just a few microns away from active fascicles, there is reason to believe that they should exhibit excellent recording selectivity as well. For instance, in Figure 2, electrodes with the contact surface facing a fascicle are likely to record a large signal from the nearby fascicle and a significantly weaker signal from all other neighboring fascicles.

Several problems have been anticipated in the modeling phase and are addressed here. First, it is possible that simulating electrical activity of millions of axons will prove to be too time-consuming, and ultimately unnecessary to show the needed data. In this case, the model will be simplified to a computationally efficient, yet still physiologically faithful model of neural activity (Peterson, Izad, and Tyler 2011). Likewise, the selectivity results of the non-homogenous model may turn out to be more based on the specific geometry of the nerve than on the configuration of electrodes. This is somewhat expected since the electrodes cannot be selective in all geometries. If this is the case, we will simulate real nerves by using actual histology data of real nerve cross-sections. This should still validate the use of interfascicular electrodes without the variability associated with randomly picking nerve geometries.

Aim 2: To quantify the recording selectivity of interfascicular electrodes in animal models

Figure 3: (left) interfascicular wires will be implanted into the epineurium, between fascicles (unpublished schematic from lab of Dustin Tyler). (Right) Example of neuronal degeneration close to intrafascicular electrode implant site (*). Arrow indicates thick tissue that surrounds implant and may worsen signal quality. Surrounding neurons are in an injured state, as evident by thinning of myelin sheath. (Source: (Right) Lefurge et al. 1991)

A logical step in evaluating the performance of interfascicular electrodes is testing them in an actual animal model. This can be done in parallel to Aim 1; when both are complete, it would be beneficial to evaluate the fidelity of the mathematical model by comparing its selectivity metrics to those obtained when using the animal model.

Methodology

Experimental procedures

Teflon-insulated Platinum-Iridium wires of diameters ranging from 50 to 100 µm will serve as interfascicular electrodes. One hundred microns of one side (180º) of each wire will be de-insulated at the tip. Additionally, two tripolar FINE electrodes (one for stimulation and one for recording, with a contact separation of 1 cm) will be manufactured and the recording FINE will be modified to house the interfascicular wires. Multiple wires of this diameter may be housed in the cuff.

Acute experiments will be performed on anesthetized rabbits. The sciatic nerve will be exposed bilaterally and the stimulation FINE will be implanted approximately 1 cm distal to the bifurcation of the sciatic nerve onto the tibial and the peroneal branches. The recording electrode with up to 10 interfascicular wires will be implanted just proximal to the bifurcation in order to record antidromic compound action potentials (see Implantation procedure). All experimental procedures will be approved by the Institutional Animal Care and Use Committee (IACUC).

The common peroneal and tibial branches of the sciatic nerve will be activated using the stimulating electrode, and compound action potentials (CAPs) will be captured using the modified recording FINE, filtered, amplified and digitized. Selectivity will be quantified by the electrode’s ability to record and recognize neural signals emanating from different branches. A selectivity index (SI), as defined in (Yoo and Durand 2005) will also be computed. Spontaneous neural activity will be recorded by activating the afferent sensory fibers in the leg of the animal. This will be done by pinching the toe (to stimulate nociceptive fibers) and moving the leg. At the end of the experiment, the nerve will be excised, fixed, preserved, sectioned and stained to reveal the location of the contact in the nerve. The actual electrode contacts will be left inside the nerve until the tissue is fixed in formalin; after that, they will be removed as to not interfere with the sectioning process.

Implantation procedure

An accurate method of electrode insertion into the epineurium will need to be developed. It is reported that the force required to penetrate the perineurium (a tough cellular layer enclosing fascicles) is significantly greater than that required to penetrate the epineurium (unpublished observations). Consequently, a minimum insertion force to penetrate the epineurium but not the perineurium has been determined. Based on this, a method of electrode insertion can be created that utilizes a stepper motor to drive the electrodes into the nerve with a force that is greater than that required to pierce the epineurium but less than that required to penetrate the perineurium. Ultimately, the wires should be implanted in the epineurium (Figure 3, left) and the depth of the implantation will be controlled once again using the stepper motor.

Evaluate selectivity, SNR of interfascicular electrode designs

Using the selectivity metrics, it will be possible to determine the optimal number of interfascicular contacts necessary to achieve maximum selectivity (selectivity index is expected to increase asymptotically, implying that after a certain point, adding more interfascicular contacts will not significantly improve selectivity). Additionally, the SNR of the electrodes will be evaluated and averaged over several experiments and separated into motion SNR (achieved when the leg of the animal is being moved, which activates the neurons synapsing on spindle fibers) and nociceptive SNR (the response to the pinching and pricking of the leg). Both metrics are physiologically relevant since they evaluate the ability of the interfascicular electrode to record actual (not electrically evoked) activity, which is useful for high-fidelity neuronal control systems.

Analysis of histology

The images of cross-sections of neurons around the implantation site will be stored and analyzed. The purpose of the analysis will be two-fold: first, it will be used to ensure that the interfascicular electrodes are in fact in the epineurium, which is a key metric in this study. Second, high-resolution histology slides will be able to show signs of acute neuronal damage such as demyelination, neuronal death, and blood vessel rupture (see Figure 3, right). The presence of any of these acute signs of damage will be beneficial in beginning the chronic studies; non-recoverable acute neuronal damage cannot be tolerated in long-term experiments.

Expected outcomes, potential problems, risks

Figure 4: Preferential recording is achieved using interfascicular wire. Prior to rotation (blue), the electrode records 100% of both tibial and c. peroneal signals. After rotation (red), the tibial branch is still recruited to 95% (circle) while the signal from the c. peroneal branch reaches only ~ 50%, indicating a recording orientation favorable to the tibial branch. Schematic on the right indicates possible positions of the interfascicular contacts. (Source: Kolb, unpublished observations)

First, we expect the histology to reveal that all interfascicular wires implanted using the developed system were in fact in interfascicular space which would validate the implantation approach, and we expect to see little to no signs of acute neuronal degeneration since the electrodes are not in direct contact with axons or the fluids surrounding them. One potential concern is that if the electrodes in the epineurium are too large, they might rupture blood vessels, causing localized edema in the nerve and an increased inflammatory response. This is a serious potential concern that can affect the ability of the electrode to perform in chronic experiments. It can be addressed by making the electrodes much smaller by using smaller-diameter wire. Wires are used for the specific purpose of making the electrode less sharp and stiff, which would minimize the risk of penetrating blood vessels.

Second, we expect the interfascicular electrodes to record selectively in vivo, as defined by their ability to record different signals for different branches of the sciatic nerve. There is preliminary data suggesting that interfascicular electrodes demonstrate the capacity for “preferential” recording, indicating a relationship between contact position within the nerve and recorded activity from different fascicles (Figure 4).

Some possible complications will be addressed. First, inserting the electrodes into the epineurium may be difficult, especially with low Young’s modulus wires that bend readily under pressure. In this case, purely for the purposes of the acute experiments, silicon-based microelectrode shanks such as the Michigan probe (www.neuronexus.com) may be inserted into the epineurium using a micromanipulator. Their location in the nerve can be assessed using a microscope. This would not be a final design for an interfascicular electrode since one of the reasons for using wires as opposed to stiff shanks is the decreased chance of neuronal damage.

Ultimately, to be useful in chronic studies, the electrode must be biocompatible. In the case that this cannot be achieved using the interfascicular wires, novel mechanically dynamic materials for electrodes have been developed and tested intracortically (Harris, Capadona, et al. 2011; Harris, Hess, et al. 2011) can be modified to function similarity in the PNS as well.

Another potential risk is the potential inability of the interfascicular electrode to reject outside interference such as electromyographic (EMG) signals from muscle activity and stimulus artifact from the inherent electric field present when stimulating (see (Triantis and Demosthenous 2006)). This problem can be avoided by using predictive template matching to remove the stimulus artifact. The EMG interference can likewise be minimized by increasing the order of the filter in the recording circuit since the frequency content of ENG signals is somewhat different than that of EMG signals (Popovic et al. 1993).

Aim 3: To determine the viability of using interfascicular electrodes in chronic animal studies

The feasibility of any device (including nerve electrodes) implanted into humans must go through a rigorous stage of chronic animal testing to determine its safety and efficacy. The third Aim will serve as a chronic animal study that will evaluate the feasibility of using this electrode in humans. Here, the animal’s normal gait will be used to assess the electrode’s ability to record very low-amplitude neural signals and selectively pick out the muscle that is currently active. The signals obtained with the nerve electrode will be compared to those obtained with EMG electrodes to correlate the two signals and determine whether the electrode identified the active muscle correctly.

Methodology

Electrode fabrication

Flat Interface Nerve Electrodes (FINEs) will be manufactured using the methods previously described in (Yoo and Durand 2005). In the modified interfascicular FINEs, the middle row of contacts will be replaced with interfasciclular electrodes that will penetrate the nerve at different depths. There will be several wires in the epineurium – the specific number will be determined by the results of Aim 1 and 2, which should quantify the optimized number of contacts necessary to attain maximum selectivity. Additional parameters of the interfascicular electrodes (such as composition, diameter, contact size, etc) will also be defined based on the results of the acute experiments.

Chronic experiment procedures

Figure 5: Ensemble recording from a dorsal root ganglion (DRG) of an awake, walking cat. The timing of spikes during normal gait (and in the case of the present study, the associated EMG data) helps identify which population of neurons is active. EMG signals will be paired with neuronal recordings to evaluate the electrode’s ability to record functionally different ENG signals. (Source: (Aoyagi et al. 2003)

Seven chronic experiments will be performed on canines under sterile surgical procedures. The canines will be anesthetized, and bilateral incisions will be made on both legs, just proximal to the knee joint to expose the bifurcation of the sciatic nerve into the common peroneal, tibial, and sural branches. In five animals, the interfascicular FINE will be implanted just proximal to the bifurcation. The two others will serve as controls; FINEs without electrodes will be implanted at the same location to evaluate the histology post-mortem. In non-control canines, EMG electrodes will be implanted in the muscles innervated by the sciatic nerve, the tibialis anterior, posterior tibialis, and gastrocnemius. The wound will be closed and bandaged, and all animals will be given time to recover from the surgery. All animal protocols will be approved by IACUC.

Evaluate chronic performance of interfascicular electrodes (selectivity, SNR)

Animals will be observed over a period of 10 weeks. Once every week, an experiment to evaluate the recording selectivity of the interfascicular electrodes will be performed. Each non-control dog will be equipped with a harness to house recording equipment and will be observed while walking on a treadmill for 10 minutes. During this time, the ENG from the interfascicular electrode and EMG from the relevant leg muscles will be recorded and stored for subsequent analysis.

The ENG recorded using each interfascicular contact will be analyzed using Principle Component Analysis (PCA) techniques to identify various spike types and compiled into a recording map (example in Figure 5) where all ENG and EMG signals are on the shown on the same time scale. Correlation coefficients will be found between each ENG and EMG signal to determine how selectively interfascicular contacts can record muscle-relevant information. Selectivity will be assessed based on how many individual muscle ENGs are reliably recorded, based on their correlation with each EMG from different muscles. A selective recording will be one that is correlated significantly with one and only one EMG signal. The SNR of each contact will be calculated using correlation to EMG: the “signal” will be present when the corresponding EMG is active” while “noise” will be defined as all non-active time regions. The correlation coefficients and the SNR will be measured within the 10-week time frame to determine the effect of long-term implantation on signal quality and ability to selectively record muscle activity.

Evaluate neuronal damage

At the end of the 10-week experiment protocol, control and non-control animals will be sacrificed. The FINE will be explanted from the nerve; for non-control animals, the interfascicular contacts will be left in the nerve until sectioning (to not interfere with the microtome blade). The nerves will be fixed, preserved, sectioned and stained. In non-control animals, the staining will reveal the location of the contacts and in both groups, neuronal damage will be assessed. Damage will be quantified by examining the nerve cross-sections closest to the implantation site for demyelination and neuronal necrosis.

Expected outcomes, potential problems, risks

It is expected that the interfascicular electrode will be able to selectively record all neuronal activity associated with normal gait. In other words, we expect each EMG recording to correlate with the recording made with at least one interfascicular contact. Additionally we expect the SNR of this electrode to be comparable to that of other chronic nerve electrodes used to record physiological activity (Sahin, Durand, and Haxhiu 1999).

One of the hazards associated with chronic implantation is the effect of electrode encapsulation and local inflammation on recording. Neuronal damage from control and non-control animals will be assessed and compared. It is expected that interfascicular electrodes to not cause significantly more damage (as defined earlier) than the cuff alone; however, since chronic studies of interfascicular electrodes have not been reported to date, the response of the nerve to electrodes in the epineurium is unknown. This hazard will be minimized by precise and minimally invasive surgery techniques that should minimize local damage and prevent a significant inflammatory response. Additionally, a mechanically adaptive polymer inspired by the sea cucumber dermis was shown to decrease local neuronal damage in intracortical recordings (Harris, Capadona, et al. 2011). This material can potentially be used as a substrate for dynamically compliant and minimally invasive interfascicular probes.

Additionally, it is possible that one or more electrode contacts fail in the duration of the experiment due to errors during surgery or inflammatory response that pushes the electrode out of the epineurium. If too many contacts are damaged (as defined by the optimal number of contacts in the epineurium in Aim 2), the experiment on that specific animal may be terminated prematurely; this will warrant acquiring another animal.

Timeline

The proposed timeline for the project is 4 years. A beneficial aspect of this proposal is that at the first two specific aims can be addressed simultaneously. We plan to begin work on the first two aims in the first year. The first Aim is expected to run approximately a year and the second, 3 years (in order to fully develop a non-damaging interfascicular electrode and evaluate its efficacy in vivo). The last aim will be completed within the span of 1 year, allotting possible time for replacing one or more of the animals in the study due to potential problems with implantation or electrode longevity.

Innovation

This research presents a novel method of interfacing with the peripheral nervous system. While studies have shown the effectiveness of using interfascicular electrodes for nerve stimulation, at the time of writing, no study indicating usage of interfascicular electrodes for neural recording has been published.

The novelty of our approach to achieving a stable and selective neural interface for recording stems from implanting electrodes in the epineurium - outside the fascicle but inside the nerve. The selection of the electrode to be used is also novel: we purposely use biocompatible electrodes with a low Young’s modulus to ensure a dynamic and chemical compatibility with nervous tissue, which should minimize local autoimmune response. This is in contrast with modern intrafascicular electrodes such as the LIFE and the Utah array where the former requires a needle to pierce the perinueurium and the latter is inserted into the nerve at high speeds for the same purpose (N. Lago et al. 2007; Aoyagi et al. 2003). Avoiding penetrating the perineurium may thus be advantageous by obviating the need for an aggressive insertion method. It may also be beneficial because intracortical electrode studies show that neuronal injury is typically concentrated around the shank of the electrode, implying that neuronal necrosis is prevalent in the neurons adjacent to a foreign body (Harris, Capadona, et al. 2011; Harris, Hess, et al. 2011; McConnell, Schneider, and Bellamkonda 2007). There is reason to believe that the same phenomenon is observed in the PNS; therefore avoiding the endoneurium may be a way to avoid acute neuronal death. In addition to recording selectivity, this metric will ultimately be relevant in evaluating the stability and efficacy of the recording interface.

The development of a selective and minimally invasive recording interface will have important clinical impacts on the areas of rehabilitation and prostheses. For rehabilitation, this reliable, selective interface will enable a control system of an FES device to stimulate several muscles independently, and in a fashion corresponding to the desires of the patient. This would be an enormous advancement towards complete functional restoration of patients suffering from post-stroke paralysis, spinal cord injuries, or neurodegenerative diseases. For the prosthetic industry, the proposed interface could enable a motorized prosthetic system to have multiple degrees of freedom, where the motor usage is directly proportional to the user’s own neuronal signals. Users of such prosthetic systems will be able to regain much of the functions associated with the lost limb, even in such complex systems as the human hand. Ultimately, the usability of the interfascicular electrode technology is not confined to the two areas of clinical practice outlined above. It can be used whenever it is necessary to “listen” to the nervous system, which in the future may become advantageous for a human-machine interface, diagnostic, or other purposes.

Bibliography

Aoyagi, Yoichiro, Richard B Stein, Almut Branner, Keir G Pearson, and Richard A Normann. 2003. “Capabilities of a penetrating microelectrode array for recording single units in dorsal root ganglia of the cat.” Journal of Neuroscience Methods 128 (1-2) (September 30): 9-20.

Choi, A. Q, J. K Cavanaugh, and D. M Durand. 2001. “Selectivity of multiple-contact nerve cuff electrodes: a simulation analysis.” IEEE Transactions on Biomedical Engineering 48 (2) (February): 165-172. doi:10.1109/10.909637.

Geddes, L. A., and L. E. Baker. 1967. “The specific resistance of biological material—A compendium of data for the biomedical engineer and physiologist.” Medical & Biological Engineering 5 (May): 271-293. doi:10.1007/BF02474537.

Harris, J P, J R Capadona, R H Miller, B C Healy, K Shanmuganathan, S J Rowan, C Weder, and D J Tyler. 2011. “Mechanically adaptive intracortical implants improve the proximity of neuronal cell bodies.” Journal of Neural Engineering 8 (6) (December): 066011. doi:10.1088/1741-2560/8/6/066011.

Harris, J P, A E Hess, S J Rowan, C Weder, C A Zorman, D J Tyler, and J R Capadona. 2011. “In vivo deployment of mechanically adaptive nanocomposites for intracortical microelectrodes.” Journal of Neural Engineering 8 (4) (August): 046010. doi:10.1088/1741-2560/8/4/046010.

Haugland, M. K, and T. Sinkjaer. 1995. “Cutaneous whole nerve recordings used for correction of footdrop inhemiplegic man.” IEEE Transactions on Rehabilitation Engineering 3 (4) (December): 307-317. doi:10.1109/86.481970.

Haugland, Morten, Andreas Lickel, Ron Riso, Margareth M Adamczyk, Michael Keith, Inger Lauge Jensen, Jens Haase, and Thomas Sinkjær. 1997. “Restoration of Lateral Hand Grasp Using Natural Sensors.” Artificial Organs 21 (3) (March 1): 250-253. doi:10.1111/j.1525-1594.1997.tb04661.x.

Hodgkin, A. L., and A. F. Huxley. 1952. “A quantitative description of membrane current and its application to conduction and excitation in nerve.” The Journal of Physiology 117 (4) (August 28): 500-544.

Hoffer, J A, R B Stein, M K Haugland, T Sinkjaer, W K Durfee, A B Schwartz, G E Loeb, and C Kantor. 1996. “Neural signals for command control and feedback in functional neuromuscular stimulation: a review.” Journal of Rehabilitation Research and Development 33 (2) (April): 145-157.

Lago, N., K. Yoshida, K. P Koch, and X. Navarro. 2007. “Assessment of Biocompatibility of Chronically Implanted Polyimide and Platinum Intrafascicular Electrodes.” IEEE Transactions on Biomedical Engineering 54 (2) (February): 281-290. doi:10.1109/TBME.2006.886617.

Lefurge, T, E Goodall, K Horch, L Stensaas, and A Schoenberg. 1991. “Chronically implanted intrafascicular recording electrodes.” Annals of Biomedical Engineering 19 (2): 197-207.

McConnell, G. C, T. M Schneider, and R. V Bellamkonda. 2007. Acute spatiotemporal changes in neuronal density surrounding microelectrode arrays implanted in rat motor cortex. In 3rd International IEEE/EMBS Conference on Neural Engineering, 2007. CNE ’07, 137-140. IEEE, May 2. doi:10.1109/CNE.2007.369631.

McIntyre, Cameron C., Andrew G. Richardson, and Warren M. Grill. 2002. “Modeling the Excitability of Mammalian Nerve Fibers: Influence of Afterpotentials on the Recovery Cycle.” Journal of Neurophysiology 87 (2) (February 1): 995 -1006.

McNeal, Donald R. 1976. “Analysis of a Model for Excitation of Myelinated Nerve.” IEEE Transactions on Biomedical Engineering BME-23 (4) (July): 329-337. doi:10.1109/TBME.1976.324593.

Micera, S., P. M. Rossini, J. Rigosa, L. Citi, J. Carpaneto, S. Raspopovic, M. Tombini, et al. 2011. “Decoding of grasping information from neural signals recorded using peripheral intrafascicular interfaces.” J Neuroeng Rehabil 8: 53. doi:1743-0003-8-53 [pii] 10.1186/1743-0003-8-53.

Navarro, Xavier, Thilo B Krueger, Natalia Lago, Silvestro Micera, Thomas Stieglitz, and Paolo Dario. 2005. “A critical review of interfaces with the peripheral nervous system for the control of neuroprostheses and hybrid bionic systems.” Journal of the Peripheral Nervous System: JPNS 10 (3) (September): 229-258. doi:10.1111/j.1085-9489.2005.10303.x.

Nielsen, T, C Sevcencu, and J Struijk. 2011. “Fascicle-Selectivity of an Intra-Neural Stimulation Electrode in the Rabbit Sciatic Nerve.” IEEE Transactions on Bio-Medical Engineering (September 26). doi:10.1109/TBME.2011.2169671. http://www.ncbi.nlm.nih.gov/pubmed/21954195.

Perez-Orive, J., and D. M. Durand. 2000. “Modeling study of peripheral nerve recording selectivity.” IEEE Trans Rehabil Eng 8 (3): 320-9.

Peterson, E J, O Izad, and D J Tyler. 2011. “Predicting myelinated axon activation using spatial characteristics of the extracellular field.” Journal of Neural Engineering 8 (August 1): 046030. doi:10.1088/1741-2560/8/4/046030.

Popovic, D. B., R. B. Stein, K. L. Jovanovic, R. Dai, A. Kostov, and W. W. Armstrong. 1993. “Sensory nerve recording for closed-loop control to restore motor functions.” IEEE Trans Biomed Eng 40 (10): 1024-31. doi:10.1109/10.247801.

Sahin, M., D. M. Durand, and M. A. Haxhiu. 1999. “Chronic recordings of hypoglossal nerve activity in a dog model of upper airway obstruction.” J Appl Physiol 87 (6): 2197-206.

Sergi, P. N, M. C Carrozza, P. Dario, and S. Micera. 2006. “Biomechanical Characterization of Needle Piercing Into Peripheral Nervous Tissue.” IEEE Transactions on Biomedical Engineering 53 (11) (November): 2373-2386. doi:10.1109/TBME.2006.879463.

Triantis, I. F, and A. Demosthenous. 2006. “The effect of interference source proximity on cuff imbalance.” IEEE Transactions on Biomedical Engineering 53 (2) (February): 354-357. doi:10.1109/TBME.2005.862569.

Tyler, D. J, E. J Peterson, N. Brill, and K. White. 2011. Increased selectivity of clinical peripheral nerve interfaces. In 2011 5th International IEEE/EMBS Conference on Neural Engineering (NER), 257-260. IEEE, May 27. doi:10.1109/NER.2011.5910536.

Tyler, D. J., and D. M. Durand. 1997. “A slowly penetrating interfascicular nerve electrode for selective activation of peripheral nerves.” IEEE Trans Rehabil Eng 5 (1): 51-61.

Veltink, P. H., J. A. van Alste, and H. B. Boom. 1988. “Simulation of intrafascicular and extraneural nerve stimulation.” IEEE Trans Biomed Eng 35 (1): 69-75. doi:10.1109/10.1338.

Yoo, P. B., and D. M. Durand. 2005. “Selective recording of the canine hypoglossal nerve using a multicontact flat interface nerve electrode.” IEEE Trans Biomed Eng 52 (8): 1461-9. doi:10.1109/TBME.2005.851482.

Yoshida, K., K. Hennings, and S. Kammer. 2006. Acute Performance of the Thin-Film Longitudinal Intra-Fascicular Electrode. In Biomedical Robotics and Biomechatronics, 2006. BioRob 2006. The First IEEE/RAS-EMBS International Conference on, 296-300.