Temporal Considerations for Stimulating Spiral Ganglion Neurons with Cochlear Implants

Abstract

A wealth of knowledge about different types of neural responses to electrical stimulation has been developed over the past 100 years. However, the exact forms of neural response properties can vary across different types of neurons. In this review, we survey four stimulus-response phenomena that in recent years are thought to be relevant for cochlear implant stimulation of spiral ganglion neurons (SGNs): refractoriness, facilitation, accommodation, and spike rate adaptation. Of these four, refractoriness is the most widely known, and many perceptual and physiological studies interpret their data in terms of refractoriness without incorporating facilitation, accommodation, or spike rate adaptation. In reality, several or all of these behaviors are likely involved in shaping neural responses, particularly at higher stimulation rates. A better understanding of the individual and combined effects of these phenomena could assist in developing improved cochlear implant stimulation strategies. We review the published physiological data for electrical stimulation of SGNs that explores these four different phenomena, as well as some of the recent studies that might reveal the biophysical bases of these stimulus-response phenomena.

Figures

FIG. 1

Illustration of effective pulse rates for electrical stimulation of spiral ganglion neurons (SGNs) by a cochlear implant. A The positioning of an electrode array inserted into the cochlea (drawn as the gray mesh wireframe) relative to the SGNs that form the auditory nerve. It is desirable for stimulating electrical currents from different electrodes or electrode pairs (highlighted blue, green, yellow, and red) to maximally stimulate distinct subpopulations of SGNs (also highlighted correspondingly with blue, green, yellow, and red), such that the tonotopic arrangement of SGNs is utilized in transmitting information about different sound frequencies. However, in practice, there is substantial current spread along the length of the cochlea, such that a single SGN is subjected to a weighted sum of the currents delivered by the nearby electrodes. For example, plotted in B are current pulse trains delivered by electrodes 1–8 for a short speech segment encoded at a rate of 900 pulses/s on each electrode. C An electrode separation of 1.4 mm and a monopolar stimulation attenuation of 0.5 dB/mm (Merzenich and White 1977) translate to the current spread profile (shaded red) that smears the contribution of all 8 electrodes to an example SGN situated between electrodes 4 and 5. This compound stimulation of an SGN results in an effective pulse rate that is much higher than the single-electrode rate of 900 pulses/s. Each biphasic pulse has a duration of 25 μs/phase and a gap of 8 μs between positive and negative phases. Image in A courtesy of Cochlear Americas, © 2015, adapted from Gray’s Anatomy textbook.

FIG. 2

Stimulus-response phenomena and their associated mechanisms. The left column (Phenomenon) shows sample SGN membrane potentials (blue) in response to monophasic current pulses (red) representing the different phenomena. These were generated with a Hodgkin–Huxley-type SGN membrane model (Negm and Bruce 2014). The horizontal black dot-dashed line indicates the resting threshold current for the SGN model. Possible responsible mechanisms for each are listed in the middle column (Mechanisms), with the source listed in the right column (Reference). Note that each panel (A–D) represents one trial outcome, and in general, many trials are required to characterize each behavior due to the stochastic nature of the membrane potential and thus the resulting spiking. A Refractoriness appears as reduced excitability to the second pulse given a spike in response to the first pulse, whereas at longer interpulse intervals, a second spike is more probable. B Facilitation acts as membrane integration of two subthreshold pulses at small interpulse intervals to enable an action potential in response to the second pulse, whereas in the case of C accommodation, the states of some ion channels are responsible for reducing excitability after a subthreshold masker pulse such that an action potential may not be generated in response to a following pulse above the resting threshold current. D In response to ongoing spiking due to pulse train stimulation, spike rate adaptation refers to the diminished spiking activity over longer timescales than refractoriness. HCN hyperpolarization-activated cyclic nucleotide-gated, KLT low-threshold potassium, IAF integrate-and-fire, RMP resting membrane potential.

FIG. 3

Published data from cat SGN recordings illustrating the four stimulus-response phenomena: A refractoriness, B facilitation, C accommodation, and D spike rate adaptation, and E a summary of the timescale ranges of their operation. Data in panels A–C were collected with monophasic pulses, while the data in panel D were in response to biphasic pulses. Data in panels A–C were obtained with masker-probe pairs of pulses at a range of intervals, and the responses were characterized by the ratio of the threshold current for the second (probe) pulse to the single-pulse threshold (SPT; also referred to as the unmasked or resting threshold). A To determine the absolute and relative refractory behavior, only cases when the masker pulse elicits a spike are considered. Elevated probe pulse thresholds due to refractoriness are shown for multiple SGNs (solid dots) and their average (open circles) is fitted by the function (black line) with the equation given. B Reduced probe pulse thresholds due to facilitation are observed in the range of 100 to 300 μs. Data are shown for multiple SGNs and their average is fitted by an exponential function. C Both facilitation and accommodation were observed by using a masker-probe stimulation protocol including longer interpulse intervals. The masker pulse is set to levels of 2 to 0.5 dB below the single-pulse threshold, while the level of the delayed probe is varied. Accommodation is seen at probe threshold values greater than 0 dB, whereas facilitation occurs below 0 dB. D Responses to a masker train (left panels, stimulating at a rate of 5000 pulses/s) and following responses to a probe train (right panels, stimulating at 100 pulses/s) displayed using normalized post-stimulus time histograms. Probe responses (shown in the right panels) are displayed as black bars if preconditioned with a masker train; otherwise, they are displayed as gray bars. In the top-left panel, the masker train is delivered with a constant pulse current level, substantially above the SPT, whereas in the bottom-left panel, the masker train is delivered with a subthreshold current level. In all conditions, probe trains are set to a constant current level, close to threshold. Reduced excitability to the start of the probe train is observed for both the suprathreshold masker (top-right panel) and the subthreshold masker (bottom-right panel) compared to the cases with no masker train. The bottom-left panel case is indicative of accommodation while the top-left panel case may include the combined effects of spike rate adaptation and accommodation. E The time ranges that refractoriness, facilitation, accommodation, and spike rate adaption operate at are shown as black bars. The black-to-white gradients indicate the variability in the time ranges. The white bar outlined by a dashed black line represents the time range of accumulated accommodation in response to pulse train stimulation shown in panel D. Panel A reprinted with kind permission of Springer Science & Business Media: Fig. 7 from Miller et al. (2001), © 2001. Panel B reprinted with kind permission of Elsevier: Fig. 5 from Cartee et al. (2000), © 2000. Panel C is used with permission from Fig. 3-2 of Dynes (1996). © Massachusetts Institute of Technology. Panel D adapted with kind permission of Springer Science & Business Media: Fig. 1 from (Miller et al. 2011), © 2011.

FIG. 4

Hyperpolarization-activated cyclic nucleotide-gated (HCN) channel subunit expression in rat cochlea identified on type I SGN. Labeled in green, both HCN1 and HCN4 subunits are localized to the nodes of Ranvier neighboring the cell body and the first peripheral node of Ranvier, or the inner spiral plexus (ISP). Shown in red, vesicular glutamate transporter 3 (VGLUT3) was used to identify inner hair cells (IHCs). Reprinted with kind permission of the American Physiological Society: Fig. 4 from Yi et al. (2010), © 2010.

FIG. 5

Illustration of passive and active contributions to the facilitation (temporal summation) phenomenon, generated with a Hodgkin–Huxley-type SGN membrane model (Negm and Bruce 2014). A Relative membrane potential and B percentage of open Na channels responding to a C monophasic masker-probe stimulation paradigm with 75 μs/phase pulse durations. Responses were averaged over 100 simulation trials for the relative membrane potential and percentage of open Na channels. Panels A and B show instances when the model SGN spiked (green curve) in response to the probe pulse, when it did not spike (magenta curve) in response to the second pulse, and where the fraction of open ion channels were fixed at their resting values (passive response; black, dot-dashed curve). By comparing the cases of spiking versus no spiking, it is apparent that when the SGN spiked, it was caused by an increased number of Na channels flicking open in response to the first pulse, such that the membrane potential decayed back towards rest more slowly than it did for cases where the SGN did not spike. The increased depolarization of the membrane at the time when the second pulse is delivered contributes to greater facilitation than would be produced by the passive response.