Prediction of Individual Cochlear Implant Recipient Speech Perception With the Output Signal to Noise Ratio Metric.

OBJECTIVES: A cochlear implant (CI) implements a variety of sound processing algorithms that seek to improve speech intelligibility. Typically, only a small number of parameter combinations are evaluated with recipients but the optimal configuration may differ for individuals. The present study evaluates a novel methodology which uses the output signal to noise ratio (OSNR) to predict complete psychometric functions that relate speech recognition to signal to noise ratio for individual CI recipients. DESIGN: Speech scores from sentence-in-noise tests in a "reference" condition were mapped to OSNR and a psychometric function was fitted. The reference variability was defined as the root mean square error between the reference scores and the fitted curve. To predict individual scores in a different condition, OSNRs in that condition were calculated and the corresponding scores were read from the reference psychometric function. In a retrospective experiment, scores were predicted for each condition and subject in three existing data sets of sentence scores. The prediction error was defined as the root mean square error between observed and predicted scores. In data set 1, sentences were mixed with 20 talker babble or speech weighted noise and presented at 65 dB sound pressure level (SPL). An adaptive test procedure was used. Sound processing was advanced combinatorial encoding (ACE, Cochlear Limited) and ACE with ideal binary mask processing, with five different threshold settings. In data set 2, sentences were mixed with speech weighted noise, street-side city noise or cocktail party noise and presented at 65 dB SPL. An adaptive test procedure was used. Sound processing was ACE and ACE with two different noise reduction schemes. In data set 3, sentences were mixed with four-talker babble at two input SNRs and presented at levels of 55-89 dB SPL. Sound processing utilised three different automatic gain control configurations. RESULTS: For data set 1, the median of individual prediction errors across all subjects, noise types and conditions, was 12% points, slightly better than the reference variability. The OSNR prediction method was inaccurate for the specific condition with a gain threshold of +10 dB. For data set 2, the median of individual prediction errors was 17% points and the reference variability was 11% points. For data set 3, the median prediction error was 9% points and the reference variability was 7% points. A Monte Carlo simulation found that the OSNR prediction method, which used reference scores and OSNR to predict individual scores in other conditions, was significantly more accurate (p < 0.01) than simply using reference scores as predictors. CONCLUSIONS: The results supported the hypothesis that the OSNR prediction method could accurately predict individual recipient scores for a range of algorithms and noise types, for all but one condition. The medians of the individual prediction errors for each data set were accurate within 6% points of the reference variability and compared favourably with prediction methodologies in other recent studies. Overall, the novel OSNR-based prediction method shows promise as a tool to assist researchers and clinicians in the development or fitting of CI sound processors.

An Evaluation of Output Signal to Noise Ratio as a Predictor of Cochlear Implant Speech Intelligibility.

OBJECTIVES: Cochlear implant (CI) sound processing strategies are usually evaluated in clinical studies involving experienced implant recipients. Metrics which estimate the capacity to perceive speech for a given set of audio and processing conditions provide an alternative means to assess the effectiveness of processing strategies. The aim of this research was to assess the ability of the output signal to noise ratio (OSNR) to accurately predict speech perception. It was hypothesized that compared with the other metrics evaluated in this study (1) OSNR would have equivalent or better accuracy and (2) OSNR would be the most accurate in the presence of variable levels of speech presentation. DESIGN: For the first time, the accuracy of OSNR as a metric which predicts speech intelligibility was compared, in a retrospective study, with that of the input signal to noise ratio (ISNR) and the short-term objective intelligibility (STOI) metric. Because STOI measured audio quality at the input to a CI sound processor, a vocoder was applied to the sound processor output and STOI was also calculated for the reconstructed audio signal (vocoder short-term objective intelligibility [VSTOI] metric). The figures of merit calculated for each metric were Pearson correlation of the metric and a psychometric function fitted to sentence scores at each predictor value (Pearson sigmoidal correlation [PSIG]), epsilon insensitive root mean square error (RMSE*) of the psychometric function and the sentence scores, and the statistical deviance of the fitted curve to the sentence scores (D). Sentence scores were taken from three existing data sets of Australian Sentence Tests in Noise results. The AuSTIN tests were conducted with experienced users of the Nucleus CI system. The score for each sentence was the proportion of morphemes the participant correctly repeated. In data set 1, all sentences were presented at 65 dB sound pressure level (SPL) in the presence of four-talker Babble noise. Each block of sentences used an adaptive procedure, with the speech presented at a fixed level and the ISNR varied. In data set 2, sentences were presented at 65 dB SPL in the presence of stationary speech weighted noise, street-side city noise, and cocktail party noise. An adaptive ISNR procedure was used. In data set 3, sentences were presented at levels ranging from 55 to 89 dB SPL with two automatic gain control configurations and two fixed ISNRs. RESULTS: For data set 1, the ISNR and OSNR were equally most accurate. STOI was significantly different for deviance (p = 0.045) and RMSE* (p < 0.001). VSTOI was significantly different for RMSE* (p < 0.001). For data set 2, ISNR and OSNR had an equivalent accuracy which was significantly better than that of STOI for PSIG (p = 0.029) and VSTOI for deviance (p = 0.001), RMSE*, and PSIG (both p < 0.001). For data set 3, OSNR was the most accurate metric and was significantly more accurate than VSTOI for deviance, RMSE*, and PSIG (all p < 0.001). ISNR and STOI were unable to predict the sentence scores for this data set. CONCLUSIONS: The study results supported the hypotheses. OSNR was found to have an accuracy equivalent to or better than ISNR, STOI, and VSTOI for tests conducted at a fixed presentation level and variable ISNR. OSNR was a more accurate metric than VSTOI for tests with fixed ISNRs and variable presentation levels. Overall, OSNR was the most accurate metric across the three data sets. OSNR holds promise as a prediction metric which could potentially improve the effectiveness of sound processor research and CI fitting.

Long-term anesthetic protocol in rats: feasibility in electrophysiology studies in visual prosthesis.

Electrical stimulation of excitable cells provides therapeutic benefits for a variety of medical conditions, including restoration of partial vision to those blinded via some types of retinal degeneration. To improve visual percepts elicited by the current technology, researchers are conducting acute electrophysiology experiments, mainly in cats. However, the rat can provide a model of a range of retinal diseases and possesses a sufficiently large eye to be used in this field. This article presents a long-term anesthetic protocol to enable electrophysiology experiments to further the development of visual prostheses. Six Long-Evans rats (aged between 14 and 16 weeks) were included in this study. Surgical anesthesia was maintained for more than 15 h by combining constant intravenous infusion of ketamine (24.0-34.5 mg/kg/h), xylazine (0.9-1.2 mg/kg/h), and inhaled isoflurane in oxygen (<0.5%). Overall heart rate, respiratory rate, and body temperature remained between 187-233 beats/min, 45-58 breaths/min, and 36-38 °C, respectively. Neural responses to 200-ms light pulses were recorded from the superior colliculus using a 32-channel neural probe at the beginning and before termination of the experiment. Robust responses were recorded from distinct functional types of retinal pathways. In addition, a platinum electrode was implanted in the retrobulbar space. The retina was electrically stimulated, and the activation threshold was determined to be 5.24 ± 0.24 µC/cm2 . This protocol may be used not only in the field of visual prosthesis research, but in other research areas requiring longer term acute experiments.

Understanding the retina: a review of computational models of the retina from the single cell to the network level.

The vertebrate retina is a clearly organized signal-processing system. It contains more than 60 different types of neurons, arranged in three distinct neural layers. Each cell type is believed to serve unique role(s) in encoding visual information. While we now have a relatively good understanding of the constituent cell types in the retina and some general ideas of their connectivity, with few exceptions, how the retinal circuitry performs computation remains poorly understood. Computational modeling has been commonly used to study the retina from the single cell to the network level. In this article, we begin by reviewing retinal modeling strategies and existing models. We then discuss in detail the significance and limitations of these models, and finally, we provide suggestions for the future development of retinal neural modeling.

Cross-species Blink Characterisation Tool for the Analysis of Emerging Interventions for Overcoming Facial Paralysis.

The loss of the ability to blink is considered the most severe consequence of facial nerve paralysis. Surgical techniques and implantable technologies continue to be developed to reanimate the eye; however, few analyse the full movement of blink when evaluating success. Here, we describe a method of taking high-quality, and high-speed video recordings of the eye, to non-invasively extract meaningful data about the dynamic movement of blinking. This can then be used to assess the effectiveness of a new technology in mimicking the natural movement. The tool was validated on humans (N=2, authors) before testing on an ovine recording (N=1), to confirm the cross-species utility of the tool, for use during preclinical development of technologies. It was found to be accurate and comprehensive, able to give insights on blinking in both human and ovine cases.

Transformer-based light-evoked retinal spiking activity prediction.

Retinal visual prosthetic devices aim to restore vision via electrical stimulation delivered on the retina. While a number of devices have been commercially available, the stimulation strategies applied have not met the expectations of end-users. These stimulation strategies involve the neurons being activated based on their spatial properties, regardless of their functions, which may lead to lower visual acuity. The ability to predict light-evoked neural activities thus becomes crucial for the development of a retinal prosthetic device with better visual acuity. In addition to temporal nonlinearity, the spatial relationship between the 2-dimensional light stimulus and the spiking activity of neuron populations is the main barrier to accurate predictions. Recent advances in deep learning offer a possible alternative for neural activity prediction tasks. With proven performance on nonlinear sequential data in fields such as natural language processing and computer vision, the emerging transformer model may be adapted to predict neural activities. In this study, we built and evaluated a deep learning model based on the transformer to explore its predictive capacity in light-evoked retinal spikes. Our preliminary results show that the model is possible to achieve good performance in this task. The high versatility of deep learning models may allow us to make retinal activity predictions in more complex physiological environments and potentially enhance the visual acuity of retinal prosthetic devices in the future by enabling us to anticipate the desired neural responses to electrical stimuli.

Cross-modal activation of the primary visual cortex by auditory stimulation in RCS rats: considerations in visual prosthesis.

An important brain re-wiring, the so-called cross-modal plasticity, occurs during progression of retinal degenerative diseases to compensate for lack of visual input. The visual cortex does not go 'unused', instead it is devoted to processing other sensory modalities. In this study we recorded, in the visual cortex, visual- and auditory-evoked potentials in an anesthetized murine model of retinal degeneration. The latency to the first peak of the recorded local field potentials was used to assess the speed of the response. Visual responses occurred significantly faster in the control group. Conversely, auditory responses appeared significantly faster in animals with retinal degeneration. This suggests the compensatory neural rewiring is optimizing the performance of other sensory modalities, hearing in this case. This phenomenon may play an important role in visual neuro-rehabilitation. Whether or not it can promote or deter the interpretation of artificially encoded neural signals from a visual prosthesis remains to be studied.

Selective neuromodulation of retinal ganglion cells via a hybrid optic-nerve and retinal neuroprosthesis for visual restoration.

A visual neuroprosthesis delivers electrical stimulation to the surviving neural cells of the visual pathway to produce prosthetic vision. While the retina is often chosen as the stimulation site, current retinal prostheses are hindered by the lack of functional selectivity that impairs the resolution. A possible strategy to improve the resolution is to combine the retinal stimulation and the stimulation of the optic nerve bundle, which contains myelinated fibres of retinal ganglion cells (RGCs) axons that vary in diameter. In this study, we used a computational model of retinal ganglion cells (RGCs) with myelinated axons to predict whether the frequency of electrical stimulation delivered to the optic nerve can be modulated to preferentially inhibit a subset of optic nerve fibres classified by diameter. The model combined a finite element model of bipolar penetrating electrodes delivering sinusoidal stimulation in the range of 25-10000 Hz to the optic nerve, and a double-cable model, to represent an optic nerve fibre. We found that the diameter of the axon fibre and ion kinetic properties of the RGC affect the neuron's frequency response, demonstrating the potential of an optic nerve stimulation to produce selective inhibition based on the axon fibre size. Clinical Relevance-This establishes the importance of considering the size of the nerve cell axons, as well as the functional type of the RGC, in stimulating the optic nerve. This can be exploited to facilitate functionally selective neuromodulation when used in conjunction with retinal stimulation.

Implications of Neural Plasticity in Retinal Prosthesis.

Retinal degenerative diseases such as retinitis pigmentosa cause a progressive loss of photoreceptors that eventually prevents the affected person from perceiving visual sensations. The absence of a visual input produces a neural rewiring cascade that propagates along the visual system. This remodeling occurs first within the retina. Then, subsequent neuroplastic changes take place at higher visual centers in the brain, produced by either the abnormal neural encoding of the visual inputs delivered by the diseased retina or as the result of an adaptation to visual deprivation. While retinal implants can activate the surviving retinal neurons by delivering electric current, the unselective activation patterns of the different neural populations that exist in the retinal layers differ substantially from those in physiologic vision. Therefore, artificially induced neural patterns are being delivered to a brain that has already undergone important neural reconnections. Whether or not the modulation of this neural rewiring can improve the performance for retinal prostheses remains a critical question whose answer may be the enabler of improved functional artificial vision and more personalized neurorehabilitation strategies.

Facial nerve paralysis: A review on the evolution of implantable prosthesis in restoring dynamic eye closure.

Facial nerve paralysis (FNP) is a debilitating condition that leaves those affected with disfigurement and loss of function. The most important function of the facial nerve is protecting the eye through eye closure and blinking. A series of reanimation techniques have been reported to restore dynamic function in FNP, but the lack of a universally accepted method that is reliable and reproducible with immediate effect has led to the introduction of several implantable devices. Most of these devices have been applied to assist blinking; however, the delicate anatomy and unique mechanics of eye closure are difficult to replicate. Lid loading is the most frequently used implant today, which is a passive device that can aid in volitional eye closure but has a limited effect on blinking. Dynamic action can be achieved with active prostheses but achieving successful long-term function remains elusive. Device action must also be coupled with a real-time feedback mechanism in order to capture the natural variation in facial muscle movements. This review discusses all prostheses used for restoring eye closure and blinking to date and explores their relative merits.

Overcoming Facial Paralysis with an Implantable Actuator for Restoration of Blink.

The loss of the ability to blink the eyelid is considered the most severe effect of facial nerve paralysis. The delicate homeostasis of the eye is disrupted, and without frequent intervention, the cornea can become damaged, ultimately resulting in blindness. The psychosocial impact is also significant, with individuals withdrawing from society to hide what they perceive to be a disfigurement. Surgical and engineering interventions have been devised to reanimate blink, however, a solution has yet to be designed which addresses both functional and aesthetic concerns. Here we describe an implantable electromagnetic actuator to restore the capacity to blink. Triggered synchronously with the contralateral eye, and externally modifiable to tailor treatment post-operatively to the individual, this implant restores complete blinking and a natural appearance. Cadaver studies (N=12) have been used to validate the device design, including the form factor and force required to elicit a blink, while a passive in vivo study (N=1) has verified the surgical protocol and recovery.

Implantation and long-term assessment of the stability and biocompatibility of a novel 98 channel suprachoroidal visual prosthesis in sheep.

Severe visual impairment can result from retinal degenerative diseases such as retinitis pigmentosa, which lead to photoreceptor cell death. These pathologies result in extensive neural and glial remodelling, with survival of excitable retinal neurons that can be electrically stimulated to elicit visual percepts and restore a form of useful vision. The Phoenix99 Bionic Eye is a fully implantable visual prosthesis, designed to stimulate the retina from the suprachoroidal space. In the current study, nine passive devices were implanted in an ovine model from two days to three months. The impact of the intervention and implant stability were assessed using indirect ophthalmoscopy, infrared imaging, and optical coherence tomography to establish the safety profile of the surgery and the device. The biocompatibility of the device was evaluated using histopathological analysis of the tissue surrounding the electrode array, with a focus on the health of the retinal cells required to convey signals to the brain. Appropriate stability of the electrode array was demonstrated, and histological analysis shows that the fibrotic and inflammatory response to the array was mild. Promising evidence of the safety and potential of the Phoenix99 Bionic Eye to restore a sense of vision to the severely visually impaired was obtained.

Safety and biocompatibility of a bionic eye: Imaging, intraocular pressure, and histology data.

The data presented here are related and supplementary data to the research article "Implantation and long-term assessment of the stability and biocompatibility of a novel 98 channel suprachoroidal visual prosthesis in sheep" [1]. In Eggenberger et al., nine sheep of the Suffolk (N=2) and Dorper (N=7) breeds were implanted in the left eye with an electrically inactive, suprachoroidal retinal stimulator (Bionic Eye) for durations of up to 100 days. The surgical safety, implant stability and device biocompatibility were assessed. Intraocular pressure measurements, indirect and infrared ophthalmoscopy and optical coherence tomography were performed at fixed time points to evaluate the clinical effects of the surgery and device implantation. Post-mortem eye tissue collection and histology was performed to measure the effects of the intervention at the cellular level. The data, including a comprehensive collection of fundus, infrared, optical coherence tomography and histology images can be used as a reference for comparison with other research, for example, active retinal stimulators. Furthermore, these data can be used to evaluate the suitability of the sheep model, in particular Dorper sheep, for future research.

Recreation of eyelid mechanics using the sling concept✰.

BACKGROUND: Paralytic lagophthalmos causes major functional, aesthetic and psychological problems in patients with facial paralysis. The Bionic Lid Implant for Natural Closure (BLINC) project aims to restore eyelid function using an implanted electromagnetic actuator combined with an eyelid sling. The authors performed a preliminary study using cadaveric heads to investigate the optimal application of an eyelid sling in various configurations around the orbit. METHODS: The sling was tested in a cadaveric sheep head using 2 medial anchor points and 4 lateral ostectomy points. An impulse was generated using gravitational force to test each combination of medial and lateral sling insertion sites using weights between 10 and 50 g. Each generated blink was recorded and analysed. The final result was validated in a human cadaveric model. RESULTS: The maximum amount of eye closure and closure speed displayed in sheep were 83.7 ±â€¯9.4% of total closure and 70.6 ±â€¯6.9 mm/s at a maximum force of 490 mN, respectively. The 2 inferior lateral attachments performed better at displacing the eyelid than the superior attachments. The position with the highest degree of eye-closure (improvement of 21.6%, p < 0.001) and speed (improvement of 30.4 mm/s, p < 0.001) was the combination of a posterior medial attachment and an inferior-posterior lateral attachment, which resulted in a near physiological closure in human cadaver. CONCLUSION: Closure improved with an inferior lateral position due to increased force acting in the direction of closure. Posterior positioning increases force acting radially, towards the centre of eyelid movement. The latter directs the closure force to effectively move the eyelid around the curved globe.

Direct activation and temporal response properties of rabbit retinal ganglion cells following subretinal stimulation.

In the last decade several groups have been developing vision prostheses to restore visual perception to the profoundly blind. Despite some promising results from human trials, further understanding of the neural mechanisms involved is crucial for improving the efficacy of these devices. One of the techniques involves placing stimulating electrodes in the subretinal space between the photoreceptor layer and the pigment epithelium to evoke neural responses in the degenerative retina. This study used cell-attached and whole cell current-clamp recordings to investigate the responses of rabbit retinal ganglion cells (RGCs) following subretinal stimulation with 25-mum-diameter electrodes. We found that direct RGC responses with short latency (

An Investigation of Audibility Effects on Cochlear Implant Speech Perception Prediction.

Output Signal to Noise Ratio (OSNR) is the Signal to Noise Ratio (SNR) at the output of a cochlear implant (CI) sound processor. Whereas other prediction metrics typically predict mean speech-in-noise test scores for a group of subjects, an OSNR-based model has been shown to accurately predict scores for individual CI recipients. The OSNR model was unable to predict scores for aggressive Ideal Binary Mask (IdBM) sound processing. This algorithm calculated Input Signal to Noise Ratio (ISNR), in each CI channel, and applied a gain function to suppress noise when a gain threshold was exceeded.The current study investigated the effect of IdBM processing on the separate speech and noise signals to determine whether audibility was affecting intelligibility. A novel metric, "OSNR and Power" (OSNRP), which combined the effect of the reduction in output speech power with OSNR, was proposed.It was found that the IdBM reduced the output speech level, likely causing audibility issues, at poor ISNRs. OSNRP accurately predicted individual speech-in-noise test scores for aggressive IdBM.The novel OSNRP metric has potential as a tool for calculating optimum configurations for sound processor parameter settings for individual CI recipients. We propose using a prescribed set of reference test conditions, the results of which can be utilized to predict outcomes when using alternative sound processing parameters and techniques, and to tailor them to the individual needs of individual CI recipients.

Mediating Retinal Ganglion Cell Spike Rates Using High-Frequency Electrical Stimulation.

Recent retinal studies have directed more attention to sophisticated stimulation strategies based on high-frequency (>1.0 kHz) electrical stimulation (HFS). In these studies, each retinal ganglion cell (RGC) type demonstrated a characteristic stimulus-strength-dependent response to HFS, offering the intriguing possibility of focally targeting retinal neurons to provide useful visual information by retinal prosthetics. Ionic mechanisms are known to affect the responses of electrogenic cells during electrical stimulation. However, how these mechanisms affect RGC responses is not well understood at present, particularly when applying HFS. Here, we investigate this issue via an in silico model of the RGC. We calibrate and validate the model using an in vitro retinal preparation. An RGC model based on accurate biophysics and realistic representation of cell morphology, was used to investigate how RGCs respond to HFS. The model was able to closely replicate the stimulus-strength-dependent suppression of RGC action potentials observed experimentally. Our results suggest that spike inhibition during HFS is due to local membrane hyperpolarization caused by outward membrane currents near the stimulus electrode. In addition, the extent of HFS-induced inhibition can be largely altered by the intrinsic properties of the inward sodium current. Finally, stimulus-strength-dependent suppression can be modulated by a wide range of stimulation frequencies, under generalized electrode placement conditions. In vitro experiments verified the computational modeling data. This modeling and experimental approach can be extended to further our understanding on the effects of novel stimulus strategies by simulating RGC stimulus-response profiles over a wider range of stimulation frequencies and electrode locations than have previously been explored.

An Investigation of the Effect of AGC Gain on the Output Signal to Noise Ratio in Cochlear Implant Sound Processing.

Measurement of speech intelligibility of cochlear implant (CI) recipients is typically carried out with a speech-innoise test procedure. Metrics which predict speech intelligibility can pre-screen new sound processing strategies prior to comprehensive testing with human subjects.The Output Signal to Noise Ratio (OSNR) metric calculates the Signal to Noise Ratio (SNR) which is present at the CI sound processor output. Watkins et al. (2018) found OSNR was an accurate predictor of speech intelligibility that could predict intelligibility in scenarios where other predictors could not.The current study investigated the effect of the sound processor automatic gain control (AGC) on OSNR and a simplified metric, Separate gain SNR (SSNR), which calculated the SNR at the CI output, assuming no interaction between the signal and noise in the sound processor. Prediction accuracy of OSNR was compared to that of Input SNR and SSNR.It was found that AGC-induced distortion and SNR degradation in speech gaps worsened OSNR. For scenarios with significant non-linear, time-varying processing, OSNR was the most accurate prediction metric. SSNR was found to be an inaccurate predictor.

Optic nerve and retinal electrostimulation in rats: direct activation of the retinal ganglion cells.

Visual prosthesis is competing with biological approaches to restore vision to the blind. Understanding and developing the ability to replicate the neural code of the retina are key factors that can bring bionic vision significant advantage. Here, electrically evoked potentials were recorded in anesthetized rats from the dorsal surface of the superior colliculus. Electrical stimuli of different amplitudes were delivered at the retina and the optic nerve. An evoked potential appeared in both cases within the first 5 ms post-stimulus suggesting that this component of the response was initiated by direct activation of the retinal ganglion cells. However, in the case of retinal neurostimulation, a second evoked potential occurred $9.0 \pm 3.4$ ms after the stimulus delivery. Because this component was not present in the case of optic nerve electrostimulation, it is expected to be originated by the activation of other cells in the retinal network.

Computational Models And Tools For Developing Sophisticated Stimulation Strategies For Retinal Neuroprostheses.

Improvements to the efficacy of retinal neuroprostheses can be achieved by developing more sophisticated neural stimulation strategies to enable selective or preferential activation of specific retinal ganglion cells (RGCs). Computational models are particularly well suited for these investigations. The electric field can be accurately described by mathematical formalisms, and the population-based neural responses to the electrical stimulation can be investigated at resolutions well beyond those achievable by current state-of-the-art biological techniques. In this study, we used a biophysically-and morphologically-detailed RGC model to explore the ability of high frequency electrical stimulation (HFS) to preferentially activate ON and OFF RGC subtypes, the two major information pathways of the retina. The performance of a wide range of electrical stimulation amplitudes (0 - $100~\mu \mathbf {A}$) and frequencies (1 - 10 kHz) on functionally-distinct RGC responses were evaluated. We found that ON RGCs could be preferentially activated at relatively higher stimulation amplitudes $( > 50 {\mu } \mathrm {A})$ and frequencies $( >2$ kHz) while OFF RGCs were activated by lower stimulation amplitudes (10 to $50 {\mu } \mathrm {A})$ across all tested frequencies. These stimuli also show great promise in eliciting RGC responses that parallel RGC encoding: one RGC type exhibited an increase in spiking activity during electrical stimulation whilst another exhibited decreased spiking activity, given the same stimulation parameters.