ABSTRACT
INTRODUCTIONElectrodes used in neural stimulation and monitoring serve a crucial role as the point of interface between an engineered electronic system and natural biological tissue. While an electrode may be as simple as a section of exposed conductive material, in practice, an array of considerations constrain the engineering of practical neural electrodes. Among these considerations are size and shape, current density, electronic impedance, surface chemistry, material biocompatibility, lifetime/longevity, and surgical practicality of installation and removal. Electrodes also differ relative to their intended use, which
defines the direction of electrical current and information flow. Some electrodes are purely passive and serve only to monitor and record endogenous voltages (signals). In such cases, current and information flows from the biological tissue through the
electrode to the electronic system. Other electrodes are designed for application of voltage (and therefore current) from the electronic system to neural tissue for the purpose of stimulation. Finally, there are electrodes designed to serve both purposes or to serve a hybrid purpose. For example, in fast-scan cyclic voltammetry (FSCV), a small transient voltage is applied for the purpose of monitoring neurotransmitter concentration via oxidation/reduction chemistry, measured by current flow. Here, the electronic system applies a voltage to a carbon-based electrode within neural tissue, for diagnostic and monitoring purposes rather than for the purpose of neural stimulation. ELECTRODE DESIGN CONSIDERATIONS
The most obvious differences between neural electrode designs are the construction material and the size and shape of the overall electrode and its individual electrical contacts. Electrodes differ in the number of distinct (electrically isolated) contacts or “channels” they contain: generally from one to dozens. The greater the number of channels or contacts, the greater the number of distinct areas of the brain or of an individual nerve (contiguous volume) they can stimulate or monitor. However, electrodes with more channels are necessarily larger and thus more difficult to implant and more disruptive to the host tissue. Electrodes with many channels are also more difficult to fabricate than single channel electrodes and require more complex wiring harnesses, connectors, and interface electronics.Each distinct electrically isolated channel of an electrode is associated with an electrical conductor which interfaces electrically with the neural tissue. The size of this conductor impacts the size of the area of neural tissue in contact with the electrode, and hence the volume of associated tissue stimulated or sampled and the spatial precision of the stimulation or monitoring. Electrode sizes can vary from the low single-digit micrometers (10-6 m) to low single-digit millimeters (10-3 m). Electrodes at the smaller end of this spectrum are capable of interacting with individual neurons but tend to be very fragile, short-lived, and yield very
weak (high impedance) signals that are suitable only for monitoring or research use. Electrodes at the larger end of this spectrum are more robust and more suited for chronic surgical implantation into patients for purposes of stimulation. The size (area) of an electrode also affects its electrical properties. In the case of stimulating electrodes, larger electrodes deliver current over a larger area, thus decreasing the “current
density” (the number of electrons flowing between the electrode and the tissue per unit area per unit time). This is important because high-current densities may damage the electrode, the neural tissue, or both. In engineering and regulatory practice, various empirical safety limits on current density are observed, such as that published by Shannon [1]. On the other hand, excessively large electrodes may stimulate too large a volume of tissue, leading to undesirable off-target effects. However, advanced stimulation electrodes can be made to apply different voltages to a patterned array of channels so as to “steer” the resulting electrical field in a defined way. This technique can be used to refine the exact volume that is subject to stimulation [2]. Monitoring/recording electrodes, on the other hand, are faced with a problem on the opposite (small-signal) end of the spectrum. The amount of electrical charge produced endogenously by the brain or other neural tissue is comparatively small. That is, as a high-impedance voltage source, it cannot deliver a large current. An excessively large monitoring electrode would act as a capacitor, thus requiring too much charge to change its potential, leading to a slow dynamic response and temporally ambiguous data (i.e., the electrode would act as a low-pass filter). Additionally, large electrodes monitor a large number of neurons with potentially heterogeneous behavior, leading to spatially ambiguous data. On the other hand, excessively small electrodes lead to higher impedance (as well as higher thermal noise) which reduce data quality [3].Electrodes are also commonly formed into multi-conductor arrays so as to monitor or introduce electric potentials over a grid of points. By recording voltages at several adjacent points in the grid, triangulation can be used to estimate the position of individual active neurons-even if the neuron in question is not in direct contact with one of the grid point contacts.
