Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Apr 1.
Published in final edited form as: J Neural Eng. 2016 Feb 23;13(2):025002. doi: 10.1088/1741-2560/13/2/025002

Temporal properties of network-mediated responses to repetitive stimuli are dependent upon retinal ganglion cell type

Maesoon Im 1,2, Shelley I Fried 1,2
PMCID: PMC4931047  NIHMSID: NIHMS795420  PMID: 26905231

Abstract

Objective

To provide artificially-elicited vision that is temporally dynamic, retinal prosthetic devices will need to repeatedly stimulate retinal neurons. However, given the diversity of physiological types of retinal ganglion cells (RGCs) as well as the heterogeneity of their responses to electric stimulation, temporal properties of RGC responses have not been adequately investigated. Here, we explored the cell type dependence of network-mediated RGC responses to repetitive electric stimulation at various stimulation rates.

Approach

We examined responses of ON and OFF types of RGCs in the rabbit retinal explant to five consecutive stimuli with varying inter-stimulus intervals (10–1000 ms). Each stimulus was a 4-ms-long monophasic sinusoidal cathodal current, which was applied epiretinally via a conical electrode. Spiking activity of targeted RGCs was recorded using a cell-attached patch electrode.

Main results

ON and OFF cells had distinct responses to repetitive stimuli. Consistent with earlier studies, OFF cells always generated reduced responses to subsequent stimuli compared to responses to the first stimulus. In contrast, a new stimulus to ON cells suppressed all pending/ongoing responses from previous stimuli and initiated its own response that was remarkably similar to the response from a single stimulus in isolation. This previously unreported ‘reset’ behavior was observed exclusively and consistently in ON cells. These contrasts between ON and OFF cells created a range of stimulation rates (4–7 Hz) that maximized the ratio of the responses arising in ON vs. OFF cells.

Significance

Previous clinical testing reported that subjects perceive bright phosphenes (ON responses) and also prefer stimulation rates of 5–7 Hz. Our results suggest that responses of ON cells are weak at high rates of stimulation (> ~7 Hz) due to the reset while responses of OFF cells are strong at low rates (< ~4 Hz) due to reduced desensitization, both reducing the ratio of ON to OFF responses. In combination with previous results indicating that responses in ON cells more closely match physiological patterns (Im and Fried, 2015), our results offer a potential reason for the user preference of intermediate rates (5–7 Hz).

Keywords: Retinal prosthesis, electrical stimulation, temporal property, retinal ganglion cell, electrophysiology

INTRODUCTION

In retinal degenerative diseases such as retinitis pigmentosa and age-related macular degeneration, patients lose their sight primarily due to the degeneration of photoreceptors. However, large populations of inner retinal neurons remain morphologically and functionally preserved as those diseases mostly target outer retinal neurons (Stone et al. 1992; Santos et al. 1997; Humayun et al. 1999; Gargini et al. 2007; Mazzoni et al. 2008). Thus, electric pulses could be used to stimulate surviving neurons so that retinal ganglion cells (RGCs) could again transmit spiking signals to the brain (Zrenner, 2002). Considerable effort has been devoted to the development of microelectronic retinal prosthetics with the goal of restoring sight to the blind. For the past decade, several different groups have reported promising clinical outcomes (Humayun et al. 2003; Rizzo et al. 2003; Zrenner et al. 2011; Stingl et al. 2013; Shivdasani et al. 2014; Ho et al. 2015) such as localization/recognition of high contrast objects and reading simple words comprised of large, high-contrast letters. They also have demonstrated that electrically-elicited vision can be somewhat useful for performing tasks of daily living in those blinded by outer retinal degeneration (Zrenner et al. 2011; Ahuja et al. 2011; Humayun et al. 2012; da Cruz et al. 2013; Kotecha et al. 2014; Stingl et al. 2015; Ho et al. 2015).

To refresh elicited vision electric stimulation from the prosthetic must be periodically repeated at an appropriate rate. Thus, for improved quality of electrically-elicited vision, it is essential to understand the fundamental temporal response properties of RGCs to electric stimuli. Unfortunately however, these properties have not been well studied. For example, earlier in-vitro animal studies reported that repetitive stimuli always desensitize the responses of RGCs irrespective of cell type, and the effect of desensitization becomes more pronounced at increasing stimulation rates (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015). This is somewhat in contrast to the range of 5–7 Hz preferred by users of retinal prosthetics (Zrenner et al. 2011; Stingl et al. 2013; Chuang et al. 2014; Stingl et al. 2015). Therefore, desensitization alone does not explain the preferred range of stimulation frequencies reported from the clinical trials. Rather, a slower rate of stimulation might be preferable for stronger responses of ganglion cells.

Given the diversity of RGC types in mammalian retina (Masland 2001; Rockhill et al. 2002; Sanes and Masland, 2015) and the heterogeneity in their response to electric stimulation (Jensen and Rizzo III, 2008; Tsai et al. 2009; Cai et al. 2011; Cai et al. 2013; Lee et al. 2013; Twyford et al. 2014; Twyford and Fried, 2015; Im and Fried, 2015), temporal properties of retinal responses are likely to be more complex than previously reported. In particular, the aforementioned homogeneous desensitization is somewhat surprising for ON vs. OFF cells because retinal circuits of ON and OFF pathways are asymmetric (Ölveczky et al. 2003; Margolis and Detwiler, 2007; Liang and Freed, 2010). Further, the distinct latencies in ON vs. OFF types of RGCs in response to electric stimulation (Im and Fried, 2015) suggest that repetitive stimuli delivered at various rates could have different effects on these two pathways. If the responses to repetitive stimulation are in fact heterogeneous across different types of RGCs, some types may contribute more to psychophysical percepts than others at certain stimulation frequencies. Thus, a more systematic investigation of the cell type dependence in the responses to repetitive stimulation might help to explain the optimal stimulation rates.

In the present study, we examined the differences across ON and OFF types of RGCs in their responses to repetitive electric stimulation. Targeted RGCs were first classified into known physiological types with their light responses. Then, we recorded the responses elicited by repetitive electric stimulation in ON and OFF RGCs as a function of stimulation rate. Our results with repetitive stimulation revealed fundamental differences between responses in ON vs. OFF cells, resulting in ON/OFF response ratios that varied for different rates of stimulation. These results offer new insights into previously reported psychophysical testing with retinal implants.

METHODS

Animal preparation and retina isolation

The care and use of animals followed all federal and institutional guidelines, and all protocols were approved by the Institutional Animal Care and Use Committee of the VA Boston Healthcare System and the Subcommittee of Research Animal Care of the Massachusetts General Hospital. New Zealand White rabbits (~2.5kg) were anesthetized with an intramuscular injection of xylazine/ketamine and subsequently euthanized with an intracardial injection of pentobarbital sodium. Immediately after death, the eyes were removed. All subsequent procedures were performed under dim red illumination. After hemisection of the eye ball, the front half was removed and the vitreous was eliminated. The retina was then separated from the retinal pigment epithelium and mounted, photoreceptor side down, on a ~10 mm square piece of Millipore filter paper (0.45 μm HA membrane filter, EMD Millipore, Billerica, MA) using vacuum grease applied between the filter paper and the recording chamber (~1.0 mL volume). A ~2.1 mm circle at the center of the Millipore filter paper allowed light transmission from below to be projected onto the photoreceptor layer.

Electrophysiology

Small holes, typically smaller than 100 μm in diameter, were made in the inner limiting membrane using patch pipettes. Then, retinal ganglion cells (RGCs) with large somata (diameter > 20 μm) were targeted for recordings so as to avoid recording from displaced amacrine cells. Spiking was recorded with a cell-attached patch electrode (4–8MΩ) filled with Ames medium. Two silver chloride-coated silver wires served as the ground and were positioned at opposite edges of the recording chamber, each ~15 mm away from the targeted cell. Data were recorded and low-pass filtered at 2 kHz using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA), and digitized at 10 kHz by a data acquisition card (PCI-MIO-16E-4, National Instruments, Austin, TX). The retinal tissue was continuously perfused at 4 mL/min with Ames medium (pH 7.4) at 36°C, equilibrated with 95% O2 and 5% CO2.

Electric stimulation

Electrical stimuli were delivered via a 10-kΩ platinum-iridium electrode (MicroProbes, Gaithersburg, MD); the exposed area at the electrode tip (no Parylene-C insulation) was conical with an approximate height of 125 μm and base diameter of 30 μm, giving a surface area of ~5,900 μm2 (comparable area to a 40-μm diameter disk electrode). Stimulating electrodes were positioned 25 μm above the inner limiting membrane; the tip of the electrode was raised by micromanipulator after touching the surface of the inner limiting membrane. The tip of the stimulating electrode was laterally offset by 50 ± 10 μm from the soma towards the optic disk in the direction of the proximal axon. Two silver chloride-coated silver wires served as the return; each was positioned ~8 mm from the targeted cell and ~6 mm from the other wire. The electric stimuli were applied by a stimulus generator (STG2004, Multi-Channel Systems MCS GmbH, Reutlingen, Germany). The data acquisition and light/electric stimuli were controlled by custom software written in LabView (National Instruments) and MATLAB (Mathworks, Natick, MA); Daniel Freeman developed an earlier version of the software.

The stimulus waveform was typically a cathodic-only half-sinusoid with a duration of 4 ms, corresponding to a half-period of a single 125 Hz sine wave (Im and Fried, 2015). A comparable duration rectangular pulse showed the lowest thresholds for eliciting percepts in a previous clinical report (Zrenner et al. 2011) as well as the maximum network-mediated responses in a previous in-vitro study (Lee et al. 2013). In experiments using multiple stimuli, a series of identical stimuli (up to 10 half-sinusoids) were delivered at constant inter-stimulus intervals (100 ms), or five consecutive stimuli were delivered with inter-stimulus intervals (between the offset of a previous stimulus and the onset of a new stimulus) ranging from 10 to 1000 ms. The same set of stimuli was presented 7 times. In order to capture responses with long latencies (Freeman and Fried, 2011; Lee et al. 2013; Boinagrov et al. 2014; Im and Fried, 2015), responses were recorded for 1 s following stimulus onset. RGC activity was also recorded for 0.5 s prior to the onset of the stimulus to monitor spontaneous spiking.

Ganglion cell type classification

Targeted RGCs were classified to be either ON or OFF by their response to stationary flashed bright spots (diameter range: 100–1000 μm, presentation duration: 1 s) on a gray background, which were repeated at least three times. The flashed spots were centered on the targeted soma and projected from below onto the photoreceptor outer segments using an LCD projector (InFocus, Portland, OR). Only those cells that generated brisk responses (Caldwell and Daw 1978; Amthor et al. 1989) were targeted for further study. ON cells were further classified as brisk transient (BT) if they exhibited doublets (or triplets) in their spontaneous activity or in their weak light responses (DeVries and Baylor, 1997; Hu and Bloomfield, 2003). Brisk ON cells that were not BT were classified as brisk sustained (BS) cells. All ON cells were also tested with moving bars to ensure they were not directionally selective (Barlow et al. 1964). OFF cells were further classified into BT or BS subtype by their responses to a single electric stimulus: OFF BT cells showed low jitter (<5 ms) of the last spike as well as small interspike intervals (ISIs) (< 5ms) while OFF BS cells showed high jitter (> 10 ms) and bigger ISIs (> 5ms) (Im and Fried, 2015).

Data analysis

The timing of individual spikes was detected from raw recordings with custom software written in MATLAB. Each row in raster plots depicts elicited spikes in response to the same stimulus; a vertical line represents a single spike. Raster plots of evoked responses to repetitive stimulation (e.g. with varying numbers of stimuli or varying inter-stimulus intervals) were plotted centered at the corresponding value on the y-axis. To quantify the response arising from each stimulus when multiple stimuli were delivered, we counted the number of spikes from the stimulus onset to the onset of next stimulus, or for 0.5 s following the onset of the last stimulus. Responses were then normalized by the number of spikes elicited by a single stimulus in isolation. The response ratio between ON and OFF cells was computed as the normalized response of an ON cell divided by that of an OFF cell. The response ratio was calculated between every pair of ON (n = 6; 4 ON BT and 2 ON BS cells) and OFF cells (n = 8; 5 OFF BT and 3 OFF BS cells), and then all ratios (48 pairs in total) were averaged. The Mann-Whitney-Wilcoxon U test was used to verify statistical significance of all comparisons. P < 0.05 was considered significant. All data are presented as the mean ± one standard deviation.

RESULTS

Responses to repeating sinusoids differ from responses to a single sinusoid delivered in isolation

It had been previously reported that sinusoidal stimulation delivered at a high frequency (100 Hz) activates ganglion cells directly (Freeman et al. 2010), i.e. without simultaneously activating other presynaptic neurons within the retinal circuitry. In contrast, lower frequencies (≤ 25 Hz) tend to preferentially activate neurons further upstream in the retina which in turn leads to secondary or indirect activation of the ganglion cell (Freeman et al. 2010; Twyford and Fried, 2015). Direct activation with 100 Hz stimulation is somewhat paradoxical however given that the duration of each cathodal phase is 5 ms and pulses of comparable duration are known to strongly activate upstream neurons, i.e. activate the ganglion cell indirectly with resulting network-mediated responses that typically consist of many spikes (Lee et al. 2013; Im and Fried, 2015). In an attempt to resolve this discrepancy we measured responses of an ON BT cell to continuous sinusoidal waveforms delivered at 125 Hz (Fig. 1a) and then compared such responses to those from a single cathodal phase (half-sinusoid, duration of 4 ms) in isolation (Fig, 1b). Although multiple spikes were elicited by the first period of sinusoid (see Discussion), we found responses to the continuous sinusoid consisted of one spike per phase from the second period (Fig. 1a). This is consistent with the earlier study (Freeman et al. 2010) and highly characteristic of direct activation of the ganglion cell (Greenberg 1998; Jensen et al. 2005; Fried et al. 2006; Sekirnjak et al. 2008). However, the responses to the single cathodal phase were noticeably different and consisted of multiple bursts of spikes that could persist for durations up to ~200 ms (Fig. 1b); these prolonged multi-spike responses are highly characteristic of indirect (or network-mediated) activation (Jensen and Rizzo III, 2008; Tsai et al. 2009; Margalit et al. 2011; Eickenscheidt et al. 2012; Lee et al. 2013; Im and Fried, 2015). The presence of the large multi-spike responses to the single 125 Hz half-sinusoid indicates that such waveforms can strongly activate the network and eliminates the possibility that the discrepancy described above arises simply from the difference in the shape of the two waveforms, e.g. rectangular pulses in isolation activate the network but sinusoids of comparable duration do not. However, when the identical stimulus is incorporated into a train of sinusoids, the network is no longer activated – or its response is suppressed: the only response is the single spike that is characteristic of direct activation (Fig. 1a). Because clinical devices will utilize multiple stimuli, possibly delivered over a wide range of frequencies, it is of fundamental importance to understand why and how the bursts of spikes elicited by single half-sinusoids get suppressed when incorporated into longer trains.

Figure 1.

Figure 1

Open in a new tab

The cathodal phase influences the response more strongly than the anodal phase. (a) Raster responses to repetitive stimuli of 125 Hz sinusoids (amplitude of −100 μA). Stimulus waveform is shown with the black solid trace and elicited spikes are indicated as red vertical lines; 7 repeats of the identical stimulus are shown. (b) Raster responses to a single cathodal-only monophasic stimulus, half-period of 125 Hz. (c)–(d) Raster responses to cathodal first (c) or anodal first (d) biphasic stimulus, single period of 125 Hz. (e) Raster responses to anodal-only monophasic stimulus, half-period of 125 Hz. All stimulus waveforms are part of the identical sinusoid with the same frequency (125 Hz) and amplitude (−100 μA).

It seemed likely that the prolonged response to the single cathodal stimulus in isolation was suppressed by one of the subsequent phases from the repeating sinusoid. We therefore questioned whether the addition of an anodal phase was responsible for abolishing the multi-spike responses. However, when a single anodal phase was added immediately after or before the cathodal phase (Figs. 1c and 1d), response patterns were generally similar to that of cathodal-only monophasic stimulation (Fig. 1b). These results strongly suggest that the anodal phase is not the factor suppressing responses in repeating sinusoidal stimulation. Monophasic anodal stimulation in isolation elicited multiple spikes (Fig. 1e) with different timing than that of the anodal stimulus. But the responses to biphasic stimulation were clearly not the summed responses to both phases in isolation, e.g. Figures 1c and 1d were highly similar to the cathodal-only responses (Fig. 1b) and spikes with latencies similar to those from the monophasic anodal response were not evident. This suggests that anodal stimulus has little influence on the overall response when delivered in conjunction with a cathodal stimulus and therefore cathodal phases more strongly influence the responses to repeating sinusoids than do anodal phases. Thus, the possibility arises that subsequent cathodal phases might suppress the responses initiated by earlier stimuli.

New cathodal stimulus resets pending responses to previous stimulus

To determine whether the network-mediated response arising from a cathodal stimulus can be blocked by the delivery of a later cathodal stimulus we recorded the responses to a series of cathodal stimuli delivered at periodic intervals (Fig. 2). Onsets of stimuli are indicated by the red arrowheads at the top of each row in Figure 2. Responses of an ON BS cell to repetitive iterations of a single stimulus (shown in the group labeled ‘1’ at the top of Fig. 2a) always consisted of two periods of spiking (Im and Fried, 2015): the first consisted of one or two spikes that occurred immediately after the onset of the stimulus while the onset of the second, more prolonged burst did not occur until ~150 ms later. When a second cathodal stimulus was delivered 100 ms after the first, the prolonged burst was delayed by 100 ms and now occurred ~150 ms after the second stimulus (group labeled ‘2’) – instead of occurring ~150 ms after the first stimulus as it did when the first stimulus was delivered in isolation. A small number of spikes were observed immediately following the delivery of the second stimulus, but in just two of the seven repeats. When additional stimuli were delivered (rows labeled 3–10), the prolonged burst of spikes was always delayed to begin ~150 ms after the onset of the final stimulus. The latency of the prolonged (final) burst was calculated as a function of the number of stimuli delivered and found to be linearly related to the number of stimuli (Fig. 2b, slope: 100.6 ms per number of stimuli, intercept: 38.7 ms, adjusted R2 = 0.999); the slope confirms a consistent additional delay of ~100 ms for each additional stimulus. In addition, a short burst of spikes (typically consisting of one or two spikes) was typically found immediately following each of the stimuli.

Figure 2.

Figure 2

Open in a new tab

A new stimulus resets all pending responses to previous stimuli. (a) Raster responses to increasing numbers of cathodal stimuli; −50 μA stimuli were applied 1 to 10 times (as indicated on y-axis) with a constant inter-stimulus interval of 100 ms. The timing of stimulus delivery is indicated by a red arrowhead at the top of each raster plot. Seven trials were repeated for each number of stimulus presentations. (b) Latency of the late burst as a function of the number of stimuli. (c) Number of spikes in the last burst as a function of the number of cathodal stimuli. Shaded area indicates mean ± one standard deviation. Shaded area is not visible in panel (b) due to the highly uniform latencies across trials at each number of cathodal stimuli.

This pattern of responses suggests that each new cathodal stimulus somehow ‘resets’ the response to the previous stimulus: the new stimulus suppresses all pending responses to any previous stimuli and initiates its own response that is highly similar to the response from a single cathodal stimulus in isolation. The number of spikes within the final (prolonged) burst was found to increase for the second stimulus but then remained approximately constant (at the elevated level) for additional stimuli (Fig. 2c).

Ongoing responses are truncated by new stimuli during repetitive stimulation

The experiment of Figure 2 was designed so that each new stimulus was delivered during a period where the cell was not spiking and the new stimulus was found to reliably suppress the pending responses to previous stimuli. We questioned whether the effect might be different however if the new stimulus was delivered during a period when the cell was actively firing. To test this we varied the timing with which repetitive stimuli were delivered (Fig. 3). Five consecutive stimuli were delivered with a fixed interval that ranged from 10 to 500 ms; this range allowed us to compare the effects of delivering a new stimulus before, during and after the prolonged burst of spiking. For intervals < 100 ms, a similar pattern of responses to those in Figure 2 persisted in which each new stimulus served to ‘reset’ the response, e.g. each new stimulus elicited a brief burst of spikes but the prolonged burst did not occur until ~150 ms after the final stimulus.

Figure 3.

Figure 3

Open in a new tab

Effect of the inter-stimulus interval on the ON BS cell response to repetitive stimuli. (a) Raster plots of elicited spikes to five consecutive cathodal stimuli (−50 μA) with inter-stimulus intervals ranging from 0 to 500 ms (values given in column on left). Response to a single stimulus in isolation is shown in the top row for comparison. Timing of individual stimuli are indicated by red arrowheads at the top of each raster plot. Blue arrows show timing of stimuli with inter-stimulus interval of 275 ms (see text). (b)–(d) Response magnitudes from each stimulus were plotted as a function of stimulus number. Responses were normalized by number of spikes elicited by a single stimulus at the same current amplitude. Each data point is the average of 7 repeats. Error bars are omitted for clarity. Responses to the first and the last stimuli were highlighted by shading.

For inter-stimulus intervals that exceeded 150 ms, the new stimulus was now delivered during the prolonged burst of spikes arising from the previous stimulus. Consistent with the reset role of the new stimulus, the ongoing burst was immediately truncated and a new response initiated although it was typically difficult to distinguish the brief initial burst arising from the new stimulus from the ongoing burst associated with the previous stimulus. Nevertheless, the truncation of responses suggests that the reset occurs with a wide range of inter-stimulus intervals as long as a new stimulus is given before the completion of responses initiated by a previous stimulus. For intervals longer than 250 ms, the long burst response to the previous stimulus was completed prior to the onset of the next stimulus and the short burst was more easily detected (e.g. see the spikes indicated by the blue arrows in Fig. 3a).

Plots of the normalized responses elicited by each stimulus for three groups of inter-stimulus intervals (Figs. 3b–3d) allowed us to quickly examine the change in responsiveness as a function of inter-stimulus interval. Interestingly, the prolonged burst arising in responses to long intervals (≥ 275 ms) was consistently attenuated by later stimuli (Fig. 3d). This reduction persisted for even the slowest rates tested (intervals of 500 ms), suggesting that the retinal network had not completely returned to baseline even after an interval of 500 ms.

Reset of electric responses is consistent for both subtypes of ON cells

In a previous study, we reported that both the number of bursts as well as the timing of individual bursts generated by ON BT cells were distinct from the patterns generated by ON BS cells (Im and Fried, 2015): when a cathodal-only monophasic half-sinusoid was delivered to ON BT cells, the response consisted of three or more bursts of spikes while responses in ON BS cells consisted of two bursts (Fig. 4a, top and bottom respectively). The latencies of the later bursts were also distinct for the two cell types. These timing differences suggest that the same repetitive stimulus might have a differential effect on the responses of each type. For example, a stimulus rate of 10 Hz (100 ms intervals) would result in a new stimulus being delivered after completion of the second burst in ON BT cells but before the onset of the second burst in ON BS cells. To explore this further, we measured responses to five repetitive stimuli from both ON BT and ON BS cells (Fig. 4b). With an interval of 100 ms the first and second bursts in the ON BT cell continued to be elicited by each new stimulus while only the first burst arose in the ON BS cell. The second burst in the ON BS cell and the third burst in the ON BT cell occurred only in response to the very last stimulus, suggesting that a similar reset effect is present in both ON subtypes. To further confirm the reset function in the ON BT cell, we delivered new stimuli before the onset of the second burst (i.e. with an inter-stimulus interval of 25 ms): with this timing, the second burst was now reset as well, i.e. it did not occur during repetitive stimulation and both it and the third burst were elicited only by the last stimulus (Fig. 4c). Similar reset of electric responses by subsequent cathodal stimuli was consistently observed in both ON subtypes for all cells (n = 15/15 for ON BT; n = 17/17 for ON BS).

Figure 4.

Figure 4

Open in a new tab

Responses are reset by new stimuli in both subtypes of ON cells. (a) Raster plots of elicited spikes to a single cathodal stimulus (−100 μA) in an ON BT cell (top) and an ON BS cell (bottom). (b) Raster responses to five stimuli (inter-stimulus intervals of 100 ms) in the ON BT cell (top) and the ON BS cell (bottom). (c) Responses of the ON BT cell to five stimuli with inter-stimulus intervals of 25 ms. Stimulus timing is shown by red arrowheads at the top of each panel.

OFF cells desensitize but do not reset

In contrast to ON cells, OFF cells did not exhibit a reset in their responses but instead, consistent with previous studies (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015), the responses to subsequent stimuli were diminished (n = 14/14). Responses of an OFF BS cell to a single cathodal stimulus in isolation consisted of two bursts of spikes (group labeled ‘Single Stimulus’ at the top of Fig. 5a) (Im and Fried, 2015): the first occurred immediately after the onset of the stimulus and consisted of two spikes while the onset of the second burst occurred ~8 ms later and consisted of many spikes (best seen in the first row in the inset of Fig. 5a). In responses to the five consecutive stimuli, there was no evidence for a reset across a wide range of intervals (Fig. 5a), e.g. a complete new response did not arise from the last stimulus. For intervals ≥ 100 ms, longer than the duration of response to the single stimulus in isolation, there was a strong depression of the response to the next stimulus, consistent with earlier reports (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015). Observation of the normalized responses (Figs. 5b–5c) revealed that the responses in the second stimulus remained very low, i.e. smaller than 50 % of responses to first stimulus, even for intervals as large as 300 ms. The strength of the responses to new stimuli grew steadily for increasing intervals but did not return to baseline even for intervals as large as 1000 ms (at 1000 ms, the response to the second stimulus ~65 % of that to the first stimulus (Fig. 5c). For intervals < 100 ms, it was difficult to correlate initiated spikes to specific stimuli. For instance, we know for certain that the spikes that occurred before the second stimulus was applied were evoked by the first stimulus but assigning subsequent spikes to specific stimuli is more challenging. Interestingly, the number of elicited spikes was increased when five stimuli were delivered with intervals < 100 ms (the inset of Fig. 5a).

Figure 5.

Figure 5

Open in a new tab

OFF cell responses to repeating stimuli exhibit desensitization. (a) Raster responses to five consecutive cathodal stimuli (−50 μA) with inter-stimulus intervals ranging from 0 to 500 ms (values given in column on left). Response to a single stimulus in isolation is shown in the top row for comparison. Timing of individual stimuli are indicated by red arrowheads at the top of each raster plot. (Inset) magnified view of raster plots for the first 100 ms after the stimulus onset in responses to single stimulus and five stimuli with inter-stimulus intervals of 10–100 ms. (b)–(d) Response magnitudes from each stimulus were plotted as a function of stimulus number. Responses were normalized by number of spikes elicited in response to a single cathodal stimulus at the same amplitude. Each data point is the average of 7 repeats. Error bars are omitted for clarity. Results for inter-stimulus intervals of 10–50 ms were not plotted in (b) as it was difficult to determine whether a spike was elicited by a previous or a new stimulus. Results for inter-stimulus intervals of 750 and 1,000 ms were included in (c), but not shown in (a). Responses to the first and the last stimuli were highlighted by shading.

As previously reported, OFF BT and BS cells have different response patterns to the same cathodal stimulus (Fig. 6a): BT cells have low jitter (<5 ms) of the last spike and small ISIs (<5 ms) while BS cells have high jitter (> 10 ms) and big ISIs (> 5 ms) (Im and Fried, 2015). Also, the duration of OFF BS cell response persist longer than those of OFF BT cells (Fig. 6a). These distinct features in response to a single stimulus suggest the two subtypes might also have different responses to repetitive stimuli. When five stimuli were presented with intervals of 100 ms, both subtypes of OFF cells demonstrated reduced responses to subsequent stimuli (Fig. 6b). Interestingly however, the reduction of responses to succeeding stimuli was bigger in OFF BS cells (n = 8) than in OFF BT cells (n = 6) (Fig. 6c). The difference emerged only from the third stimulus onward (normalized responses were 47.6 ± 13.6% vs. 25.8 ± 14.0% for BT and BS subtypes, respectively; P < 0.05).

Figure 6.

Figure 6

Open in a new tab

Desensitization levels are different in the two OFF cell subtypes. (a) Raster plots of elicited spikes to a single cathodal stimulus (−100 μA) in an OFF BT cell (top) and an OFF BS cell (bottom). (b) Raster responses to five stimuli in the OFF BT cell (top) and the OFF BS cell (bottom) with inter-stimulus intervals of 100 ms. Stimulus timing is shown by red arrowheads at the top of each panel. (c) Average responses of OFF BT cells (n = 6) and BS cells (n = 8) to stimulus 2 through 5 for each of the five stimuli delivered. Responses were normalized by the number of spikes elicited by a single stimulus in each cell. The Mann-Whitney-Wilcoxon U test was applied to verify the significance of statistical comparisons; *P < 0.05. Error bars show one standard deviation.

Stimulation frequency alters the ratio between ON and OFF responses

Because the fundamental mechanism of response is different for ON and OFF cells, e.g. ON cells reset while OFF cells desensitize, it is likely that the sensitivity to different rates of stimulation will be different in each type as well. This raises the possibility that certain rates of stimulation may more strongly activate one type over the other. To explore this, we plotted the normalized response (see Methods) to an intermediate stimulus (#3 in the sequence of 5 evenly spaced stimuli) as a function of inter-stimulus interval for both ON and OFF cells (Fig. 7a). Responses to intermediate stimuli are of particular interest in retinal prosthetics as most stimuli will fall into this category while the device is active, e.g. responses to the first and last stimuli will only occur during the onset and offset of stimulation. Consistent with desensitization, responses in the OFF cells (n = 8) increased monotonically as the interval between stimuli increased (Fig. 7a, black trace). The responses in the ON cells (red trace; n = 6) also generally increased with increasing intervals but the rate of increase was more rapid for intervals < 150 ms than for intervals > 150 ms. We compared the relative strength of ON to OFF responses by plotting the pairwise ratio of ON to OFF responses (see Methods) as a function of interval (Fig. 7b) and found the peak occurred at inter-stimulus intervals of 175–225 ms, suggesting that ON vs. OFF selectivity was maximized within this range. This is because the ON cell responses were weak due to reset with short intervals (e.g. < 100 ms; Fig. 7a, red trace) while the OFF cell responses became stronger than the ON cell responses with long intervals (e.g. > 300 ms; Fig. 7a, black trace). Thus, the ON/OFF response ratios were nearly 1 or bigger than 1 (shown in dashed line in Fig. 7b) with intermediate intervals ranging from 150 to 275 ms (highlighted with a yellow band). We found similar patterns in both the responses and the ratios for the second and fourth stimuli in the train of five (data not shown). It is also worth noting that higher selectivity can be achieved for OFF cells than ON cells (OFF/ON response ratio of ~1.6) for the interval of 100 ms. However, given the higher correlations between electric and light responses in ON cells than OFF cells (Im and Fried, 2015), it may be better to selectively drive ON cells, e.g. with ON/OFF response ratios > 1, so as to better match physiological signaling patterns. Consistent with this approach, retinal prosthetic users preferentially reported bright phosphenes (Humayun et al. 1996; Humayun et al. 2003; Fujikado et al. 2007; Naycheva et al. 2012), suggesting perception may be driven by ON pathways (see Discussion).

Figure 7.

Figure 7

Open in a new tab

Response magnitudes and response ratio between ON and OFF cells vary as a function of the stimulation rate. (a) Average normalized responses in steady-state (to the 3rd stimulus among 5 stimuli; amplitude of −100 μA) were plotted as a function of inter-stimulus interval. Responses were normalized by the spike counts elicited by a single stimulus and then averaged in 6 ON (4 BT and 2 BS subtypes) and 8 OFF cells (5 BT and 3 BS subtypes). OFF cell responses were plotted only for intervals ≥ 100 ms because of the difficulty in correlating spikes to specific stimuli at intervals < 100 ms (see text). (b) Response ratios between pairs of ON and OFF cells (48 pairs in total; see Methods) were plotted as a function of inter-stimulus interval. Yellow band shows the range of inter-stimulus intervals demonstrating ratios > 1 (red dashed horizontal line). (c)–(d) Same as (a) and (b), respectively but for responses to system onset (to the 1st stimulus). (e)–(f) Same as (a) and (b) but for responses to system offset (to the last (5th) stimulus). Shaded areas in all panels indicate mean ± one standard deviation.

Finally, we examined whether responses to the first and the last stimuli in each train, corresponding to the onset and offset of stimulation (respectively), also were maximal for a specific range of stimulation frequencies. For the first stimulus (Fig. 7c), ON cell responses showed a sigmoidal shape, i.e. responses were weak for short intervals because of reset and remained approximately constant for long intervals, while OFF cell responses were generally consistent regardless of inter-stimulus interval. Accordingly, the ratio of ON vs. OFF responses to the first stimulus was sigmoidal and consistently < 1.0 (Fig. 7d). In contrast to the first stimulus, responses to the last stimulus in ON cells were large for short intervals and decreased for longer intervals. Responses in OFF cells increased only slightly with increasing inter-stimulus intervals (Fig. 7e). As a result, the response ratio for the last response showed a monotonic reduction with increasing intervals (Fig. 7f). Thus the activities of ON and OFF cells change differently as a function of inter-stimulus interval in response to the first and the last stimuli. For short inter-stimulus intervals, OFF cells are preferentially selected at stimulus onset and ON cells are preferentially selected at stimulus offset. For long inter-stimulus intervals, there is little selectivity.

DISCUSSION

Our results suggest that understanding the temporal properties of the responses to repetitive electric stimulation will be useful for optimizing the performance of retinal prosthetics. Previous studies have consistently reported that responses to repetitive stimulation become weaker for subsequent stimuli (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015), regardless of the type of ganglion cell. Contrary to earlier work however, we demonstrate here that there is a fundamental difference between the responses of ON and OFF types of RGCs (reset vs. desensitization) to repetitive stimulation. This distinction between temporal properties of those two major RGC types may offer some insight for understanding the results of psychophysical testing.

New electric stimulus resets responses of ON but not OFF cells

The responses of all OFF cells (n = 14/14) tested in our study demonstrated desensitization in response to repetitive stimuli, however the level of desensitization varied considerably (large size of shaded area in Fig. 7a). These results are consistent with previous studies (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015). However, they did not report any notable dependence on physiological types of ganglion cells (Jensen and Rizzo III, 2007; Freeman and Fried, 2011) or had not identified cell types (Lorach et al. 2015) in the level of desensitization. In contrast, we found here that OFF BT cells desensitize less in response to repetitive stimuli than OFF BS cells (Fig. 6c). We did not attempt to identify the source of this difference.

Surprisingly, the responses of all ON cells (n = 32/32) exhibited a reset behavior in which a new stimulus suppressed all pending responses to any previous stimuli and initiated its own responses that were highly similar to the response from a single stimulus in isolation. This reset feature was consistently observed in both BT and BS subtypes of ON cells (Fig. 4), suggesting that the reset is a consistent characteristic of the ON pathway in response to repeating electric stimulation. Somewhat surprisingly, this reset behavior of ON cells has not been described previously even though several earlier studies have explored responses to repetitive stimulation (Jensen and Rizzo III, 2007; Freeman and Fried, 2011; Lorach et al. 2015). It is likely that our ability to classify targeted RGCs into known types coupled with an extended post-stimulus recording time (1 second), both of which were not used in previous studies, facilitated the identification of this phenomenon.

Figure 1a suggests that responses of ON cells can be reset not only by monophasic stimulation but also by biphasic stimulation: repeating biphasic stimuli did not generate the delayed bursts elicited by single period of monophasic/biphasic stimulus (Figs. 1b–1d). Taken together with Figure 2, our results revealed that the reason why delayed spikes (i.e. second and third bursts) were not present in responses of the ON cell (Fig. 1a) to the continuous sinusoid was not the previous or subsequent anodal phase but the subsequent cathodal phase. Also, it is worth noting that number of spikes elicited during the first cathodal phase in repeating sinusoids (first 3–4 spikes in Fig. 1a) is similar to those arising from the monophasic cathodal stimulus (first 4 spikes in Fig. 1b), suggesting the first cathodal phase in Figure 1a is successfully initiating electric responses at least for its duration. However, only one spike was elicited by succeeding cathodal phases in the repeating sinusoids (Fig. 1a). Responses of OFF cells were similar in that one spike per each cathodal phase was elicited by the same repeating sinusoid (data not shown). Therefore, these properties (i.e. single spike per cathodal phase) of both ON and OFF cells could account for a previous study reporting high frequency stimulation only elicits direct spikes (Freeman et al. 2010).

Does presynaptic circuitry underlie the difference in responses of ON and OFF cells?

The unique response features of ON and OFF cells were preserved for subretinal stimulation, e.g. a new anodal stimulus reset the ON cell responses and desensitized the OFF cell responses (not shown). Also, similar results were observed for rectangular pulses (not shown). These suggest that neither the location of electrode (epi- vs. sub-retinal) nor the shape of waveform contributes to the temporal properties that were observed here. Thus, it is likely that the differences arise from intrinsic differences in the ON and OFF pathways. The most fundamental difference between ON and OFF pathways is the sign-inverting synapse that is present in the dendrites of ON bipolar cells, but lacking in OFF pathways. This means photoreceptors activated by electric stimulation will hyperpolarize downstream neurons and suppress responses in the ON system while activated photoreceptors will facilitate increased spiking activity in OFF ganglion cells. Thus, if photoreceptors were activated, ON vs. OFF responses would be expected to have some differences (Im and Fried, 2015). However, the sign-inverting synapse is not enough to completely explain the reset of ON cell responses. For instance, ongoing responses initiated by a previous stimulus were immediately truncated by a new stimulus (Fig. 3, inter-stimulus intervals of 150–250 ms), suggesting that a strong inhibitory input to ON cells arises directly from the new stimulus. Given the relatively long synaptic delay from photoreceptors to ganglion cells it is unlikely that this immediate cessation of spiking is mediated by activated photoreceptors through the sign-inverting synapse. Therefore, it is likely that one or more inhibitory amacrine cells directly synapsing onto ON ganglion cells are also strongly activated by the electric stimulus used here. Activation of inhibitory neurons is supported by an earlier study in which baseline spiking activity of RGCs was reduced by network-activating stimuli (Lee et al. 2013). We also observed a similar reduction of spontaneous firing (not shown). These results are also consistent with earlier studies in which long-lasting inhibitory currents into a retinal ganglion cell (Fried et al. 2006) or a fast hyperpolarization in inner retinal neurons (Cameron et al. 2013) were observed. While we cannot completely rule out the possibility that some portion of the inhibitory components were mediated by photoreceptor activation, the hyperpolarization of inner retinal neurons were not significantly altered in degenerate retina (rd1) (Cameron et al. 2013), suggesting that inhibitory responses are not photoreceptor-dependent.

OFF cells did not show a similar truncation of responses by a new stimulus. Instead, when a new stimulus was delivered during a period of spiking (initiated by a previous stimulus) more spikes were evoked and the overall response patterns were generally similar to the response to the single stimulus (inset of Fig. 5). This suggests that the same inhibitory pathway described above for the ON system is not activated in the OFF system. This difference in ON vs. OFF cells may result from distinct inhibitory networks in the two pathways. For example, AII amacrine cells form sign-conserving gap junctions vs. sign-inverting glycinergic synapses with ON and OFF cone bipolar cells, respectively (Famiglietti and Kolb, 1975; Raviola and Dacheux, 1987; Wässle, 2004), suggesting one possible source for the different inhibitory signals to ON and OFF cells.

Parallels with human psychophysical results

Previous clinical studies have reported that subjects were most satisfied with stimuli at a repetition rate of ~5–7 Hz (Zrenner et al. 2011; Stingl et al. 2013; Chuang et al. 2014; Stingl et al. 2015), or with inter-stimulus intervals of 200 ms for localizing multiple phosphenes (Wilke et al. 2011). Intriguingly, our results demonstrate that the steady-state ratio of ON to OFF cell responses are maximized with inter-stimulus intervals of 150–275 ms (Fig. 7b), which correspond to a similar range of rates (~4–7 Hz). The similarity in the range of intervals suggests that psychophysical preferences may result from the different temporal properties of ON and OFF cells in response to repetitive stimuli shown here. Also, the maximum ON/OFF response ratio (Fig. 7b) suggests that information from the ON system may somehow be more useful to downstream visual centers. This is consistent with preferential reporting of ON responses in human trials (Humayun et al. 1996; Humayun et al. 2003; Fujikado et al. 2007; Naycheva et al. 2012). In addition, the rates that maximize ON responses may be more preferable because electric responses of ON types have better levels of correlations to their light responses (Im and Fried, 2015). In other words, it may be preferable to maximize ON cell responses which have a strong correlation with their light responses while minimizing OFF cell responses which have a poor correlation with light responses and may therefore hamper percept quality.

A second intriguing correlation arises between our results here and previous psychophysical testing (Pérez Fornos et al. 2012) in which some subjects (n = 4/9) described bright percepts (ON responses) at the termination of a stimulus train. This type of response is consistent with the continuous reset of spiking in ON cells followed by the prolonged burst of spikes that occurs only after the termination of stimuli (Figs. 2 and 3). The fact that users reported the increased brightness at the stimulus offset only for higher stimulation rates such as 20 pulses per second (pps) and 60 pps but not for 5 pps is also highly consistent with the ON cell responses seen here.

Potential limitations of network-mediated activation

While several of the parallels described above are intriguing, there are several factors that may limit the applicability of the results found here to human clinical trials. First, several differences between the normal and degenerate retina may alter the responses to repetitive stimulation. For example, because we were not able to unequivocally identify the contribution of photoreceptors to electrical responses, the possibility exists that some or all of our findings will be sensitive to the level of retinal degeneration. However, even if photoreceptors do not contribute significantly to the responses observed here, there are physiological changes known to occur in the retina during degeneration (Strettoi and Pignatelli, 2000; Strettoi et al. 2002; Gargini et al. 2007; Stasheff 2008; Borowska et al. 2011; Menzler et al. 2014) and these too might influence the responses we observed. Thus further testing in the degenerate retina and/or the use of pharmacological agents to block specific pathways will be necessary to establish whether the present findings persist during retinal degeneration.

Our results also raise the possibility that retinal prostheses may perform better if some photoreceptor inner segments survive degeneration (Li et al. 1995; Milam et al. 1998; Lin et al. 2009) as they are the likely portion activated by electric stimulation. This raises the possibility that the retinas of those subjects who demonstrated the best clinical outcomes have more remaining photoreceptors, or at least inner segments. However, it is not known whether those subjects have any remaining outer retinal cell bodies. Subretinal implantation has been shown to damage photoreceptors although photoreceptor nuclei can be observed up to 5 weeks after implantation (Lorach et al. 2014), suggesting there may be some variability in the sensitivity to implantation. Further testing will be necessary to determine whether different shapes or coatings of implanted electrodes (Butterwick et al. 2009) can minimize the deterioration of the outer retina.

Acknowledgments

We thank Seung Woo Lee, Alex E. Hadjinicolaou, Vineeth Raghuram, and Donald K. Eddington for helpful discussion and review of the manuscript. This work was supported in part by the VA Boston Healthcare System (1I01RX000350-01A1) and by the NIH (R01EY019967, R01EY023651).

Footnotes

Conflict of interest: The authors declare no competing financial interests.

References

  1. Ahuja AK, Dorn JD, McMahon MJ, Dagnelie G, da Cruz L, Stanga P, Humayun MS, Greenberg RJ. Blind subjects implanted with the Argus II retinal prosthesis are able to improve performance in a spatial-motor task. Br J Ophthalmol. 2011;95:539–543. doi: 10.1136/bjo.2010.179622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amthor FR, Takahashi ES, Oyster CW. Morphologies of rabbit retinal ganglion cells with concentric receptive fields. J Comp Neurol. 1989;280:72–96. doi: 10.1002/cne.902800107. [DOI] [PubMed] [Google Scholar]
  3. Barlow HB, Hill RM, Levick WR. Retinal ganglion cells responding selectively to direction and speed of image motion in the rabbit. J Physiol. 1964;73:377–407. doi: 10.1113/jphysiol.1964.sp007463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boinagrov D, Pangratz-Fuehrer S, Goetz G, Palanker D. Selectivity of direct and network-mediated stimulation of the retinal ganglion cells with epi-, sub- and intraretinal electrodes. J Neural Eng. 2014;11:026008. doi: 10.1088/1741-2560/11/2/026008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Borowska J, Trenholm S, Awatramani GB. An intrinsic neural oscillator in the degenerating mouse retina. J Neurosci. 2011;31:5000–5012. doi: 10.1523/JNEUROSCI.5800-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Butterwick A, Huie P, Jones BW, Marc RE, Marmor M, Palanker D. Effect of shape and coating of a subretinal prosthesis on its integration with the retina. Exp Eye Res. 2009;88:22–29. doi: 10.1016/j.exer.2008.09.018. [DOI] [PubMed] [Google Scholar]
  7. Cai C, Ren Q, Desai NJ, Rizzo JF, III, Fried SI. Response variability to high rates of electric stimulation in retinal ganglion cells. J Neurophysiol. 2011;106:153–162. doi: 10.1152/jn.00956.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cai C, Twyford P, Fried S. The response of retinal neurons to high-frequency stimulation. J Neural Eng. 2013;10:036009. doi: 10.1088/1741-2560/10/3/036009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Caldwell J, Daw N. New properties of rabbit retinal ganglion cells. J Physiol. 1978;276:257–276. doi: 10.1113/jphysiol.1978.sp012232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cameron MA, Suaning GJ, Lovell NH, Morley JW. Electrical stimulation of inner retinal neurons in wild-type and retinally degenerate (rd/rd) mice. PLOS ONE. 2013;8:e68882. doi: 10.1371/journal.pone.0068882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chuang AT, Margo CE, Greenberg PB. Retinal implants: a systematic review. Br J Ophthalmol. 2014;98:852–856. doi: 10.1136/bjophthalmol-2013-303708. [DOI] [PubMed] [Google Scholar]
  12. da Cruz L, Coley BF, Dorn J, Merlini F, Filley E, Christopher P, Chen FK, Wuyyuru V, Sahel J, Stanga P, Humayun M, Greenberg RJ, Dagnelie G. The Argus II epiretinal prosthesis system allows letter and word reading and long-term function in patients with profound vision loss. Br J Ophthalmol. 2013;97:632–636. doi: 10.1136/bjophthalmol-2012-301525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. DeVries SH, Baylor DA. Mosaic arrangement of ganglion cell receptive fields in rabbit retina. J Neurophysiol. 1997;78:2048–2060. doi: 10.1152/jn.1997.78.4.2048. [DOI] [PubMed] [Google Scholar]
  14. Eickenscheidt M, Jenkner M, Thewes R, Fromherz P, Zeck G. Electrical stimulation of retinal neurons in epiretinal and subretinal configuration using a multicapacitor array. J Neurophysiol. 2012;107:2742–2755. doi: 10.1152/jn.00909.2011. [DOI] [PubMed] [Google Scholar]
  15. Famiglietti EV, Kolb H. A bistratified amacrine cell and synaptic circuitry in the inner plexiform layer of the retina. Brain Res. 1975;84:293–300. doi: 10.1016/0006-8993(75)90983-x. [DOI] [PubMed] [Google Scholar]
  16. Freeman DK, Eddington DK, Rizzo JF, III, Fried SI. Selective activation of neuronal targets with sinusoidal electric stimulation. J Neurophysiol. 2010;104:2778–2791. doi: 10.1152/jn.00551.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Freeman DK, Fried SI. Multiple components of ganglion cell desensitization in response to prosthetic stimulation. J Neural Eng. 2011;8:016008. doi: 10.1088/1741-2560/8/1/016008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fried SI, Hsueh H, Werblin FS. A method for generating precise temporal patterns of retinal spiking using prosthetic stimulation. J Neurophysiol. 2006;95:970–978. doi: 10.1152/jn.00849.2005. [DOI] [PubMed] [Google Scholar]
  19. Fujikado T, Morimoto T, Kanda H, Kusaka S, Nakauchi K, Ozawa M, Matsushita K, Sakaguchi H, Ikuno Y, Kamei M, Tano Y. Evaluation of phosphenes elicited by extraocular stimulation in normals and by suprachoroidal-transretinal stimulation in patients with retinitis pigmentosa. Graefe’s Arch Clin Exp Ophthalmol. 2007;245:1411–1419. doi: 10.1007/s00417-007-0563-z. [DOI] [PubMed] [Google Scholar]
  20. Gargini C, Terzibasi E, Mazzoni F, Strettoi E. Retinal organization in the retinal degeneration 10 (rd10) mutant mouse: a morphological and ERG study. J Comp Neurol. 2007;500:222–238. doi: 10.1002/cne.21144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Greenberg R. PhD thesis. Baltimore MD: Johns Hopkins University; 1998. Analysis of electrical stimulation of the vertebrate retina – work towards a retinal prosthesis. [Google Scholar]
  22. Ho AC, Humayun MS, Dorn JD, da Cruz L, Dagnelie G, Handa J, Barale PO, Sahel JA, Stanga PE, Hafezi F, Safran AB, Salzmann J, Santos A, Birch D, Spence R, Cideciyan AV, de Juan E, Duncan JL, Eliott D, Fawzi A, de Koo LCO, Brown GC, Haller JA, Regillo CD, Del Priore LV, Arditi A, Geruschat DR, Greenberg RJ. Long-Term Results from an Epiretinal Prosthesis to Restore Sight to the Blind. Ophthalmol. 2015 doi: 10.1016/j.ophtha.2015.04.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hu EH, Bloomfield SA. Gap junctional coupling underlies the short-latency spike synchrony of retinal α ganglion cells. J Neurosci. 2003;23:6768–6777. doi: 10.1523/JNEUROSCI.23-17-06768.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Humayun MS, de Juan E, Jr, Dagnelie G, Greenberg RJ, Propst RH, Phillips DH. Visual perception elicited by electrical stimulation of retina in blind humans. Arch Ophthalmol. 1996;114:40–46. doi: 10.1001/archopht.1996.01100130038006. [DOI] [PubMed] [Google Scholar]
  25. Humayun MS, Prince M, de Juan E, Jr, Barron Y, Moskowitz M, Klock IB, Milam AH. Morphometric analysis of the extramacular retina from postmortem eyes with retinitis pigmentosa. Invest Ophthalmol Vis Sci. 40:143–148. [PubMed] [Google Scholar]
  26. Humayun MS, Weiland JD, Fujii GY, Greenberg R, Williamson R, Little J, Mech B, Cimmarusti V, Van Boemel G, Dagnelie G, de Juan E., Jr Visual perception in a blind subject with a chronic microelectronic retinal prosthesis. Vision Res. 2003;43:2573–2581. doi: 10.1016/s0042-6989(03)00457-7. [DOI] [PubMed] [Google Scholar]
  27. Humayun MS, Dorn JD, da Cruz L, Dagnelie G, Sahel JA, Stanga PE, Cideciyan AV, Duncan JL, Eliott D, Filley E, Ho AC, Santos A, Safran AB, Arditi A, Del Priore LV, Greenberg RJ. Interim results from the international trial of second sight’s visual prosthesis. Ophthalmol. 2012;119:779–788. doi: 10.1016/j.ophtha.2011.09.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Im M, Fried SI. Indirect activation elicits strong correlations between light and electrical responses in ON but not OFF retinal ganglion cells. J Physiol. 2015;593:3577–3596. doi: 10.1113/JP270606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Jensen RJ, Ziv OR, Rizzo JF. Responses of rabbit retinal ganglion cells to electrical stimulation with an epiretinal electrode. J Neural Eng. 2005;2:S16–S21. doi: 10.1088/1741-2560/2/1/003. [DOI] [PubMed] [Google Scholar]
  30. Jensen RJ, Rizzo JF., III Responses of ganglion cells to repetitive electrical stimulation of the retina. J Neural Eng. 2007;4:S1–S6. doi: 10.1088/1741-2560/4/1/S01. [DOI] [PubMed] [Google Scholar]
  31. Jensen RJ, Rizzo JF., III Activation of retinal ganglion cells in wild-type and rd1 mice through electrical stimulation of the retinal neural network. Vision Res. 2008;48:1562–1568. doi: 10.1016/j.visres.2008.04.016. [DOI] [PubMed] [Google Scholar]
  32. Jensen RJ, Ziv OR, Rizzo JF, III, Scribner D, Johnson L. Spatiotemporal aspects of pulsed electrical stimuli on the responses of rabbit retinal ganglion cells. Exp Eye Res. 2009;89:972–979. doi: 10.1016/j.exer.2009.08.015. [DOI] [PubMed] [Google Scholar]
  33. Kotecha A, Zhong J, Stewart D, da Cruz L. The Argus II prosthesis facilitates reaching and grasping tasks: a case series. BMC Ophthalmol. 2014;14:71. doi: 10.1186/1471-2415-14-71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mazzoni F, Novelli E, Strettoi E. Retinal ganglion cells survive and maintain normal dendritic morphology in a mouse model of inherited photoreceptor degeneration. J Neurosci. 2008;28:14282–14292. doi: 10.1523/JNEUROSCI.4968-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lee SW, Eddington DK, Fried SI. Responses to pulsatile subretinal electric stimulation: Effects of amplitude and duration. J Neurophysiol. 2013;109:1954–1968. doi: 10.1152/jn.00293.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Li ZY, Kljavin IJ, Milam AH. Rod photoreceptor neurite sprouting in retinitis pigmentosa. J Neurosci. 1995;15:5429–5438. doi: 10.1523/JNEUROSCI.15-08-05429.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Liang Z, Freed MA. The ON pathway rectifies the OFF pathway of the mammalian retina. J Neurosci. 2010;30:5533–5543. doi: 10.1523/JNEUROSCI.4733-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lin B, Masland RH, Strettoi E. Remodeling of cone photoreceptor cells after rod degeneration in rd mice. Exp Eye Res. 2009;88:589–599. doi: 10.1016/j.exer.2008.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lorach H, Goetz G, Mandel Y, Lei X, Kamins TI, Mathieson K, Huie P, Dalal R, Harris JS, Palanker D. Performance of photovoltaic arrays in-vivo and characteristics of prosthetic vision in animals with retinal degeneration. Vis Res. 2014 doi: 10.1016/j.visres.2014.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lorach H, Goetz G, Smith R, Lei X, Mandel Y, Kamins T, Mathieson K, Huie P, Harris J, Sher A, Palanker D. Photovoltaic restoration of sight with high visual acuity. Nat Med. 2015 doi: 10.1038/nm.3851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Margalit E, Babai N, Luo J, Thoreson WB. Inner and outer retinal mechanisms engaged by epiretinal stimulation in normal and rd mice. Vis Neurosci. 2011;28:145–154. doi: 10.1017/S0952523810000489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Margolis DJ, Detwiler PB. Different mechanisms generate maintained activity in ON and OFF retinal ganglion cells. J Neurosci. 2007;27:5994–6005. doi: 10.1523/JNEUROSCI.0130-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Masland RM. The fundamental plan of the retina. Nat Neurosci. 2001;4:877–886. doi: 10.1038/nn0901-877. [DOI] [PubMed] [Google Scholar]
  44. Menzler J, Channappa L, Zeck G. Rhythmic ganglion cell activity in bleached and blind adult mouse retinas. PLOS ONE. 2014;9:e106047. doi: 10.1371/journal.pone.0106047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Milam AH, Li ZY, Fariss RN. Histopathology of the human retina in retinitis pigmentosa. Prog Retin Eye Res. 1998;17:175–205. doi: 10.1016/s1350-9462(97)00012-8. [DOI] [PubMed] [Google Scholar]
  46. Naycheva L, Schatz A, Röck T, Willmann G, Messias A, Bartz-Schmidt KU, Zrenner E, Gekeler F. Phosphene thresholds elicited by transcorneal electrical stimulation in healthy subjects and patients with retinal diseases. Invest Ophthalmol Vis Sci. 2012;53:7440–7448. doi: 10.1167/iovs.12-9612. [DOI] [PubMed] [Google Scholar]
  47. Ölveczky BP, Baccus SA, Meister M. Segregation of object and background motion in the retina. Nature. 2003;423:401–408. doi: 10.1038/nature01652. [DOI] [PubMed] [Google Scholar]
  48. Pérez Fornos A, Sommerhalder J, da Cruz L, Sahel JA, Mohand-Said S, Hafezi F, Pelizzone M. Temporal properties of visual perception on electrical stimulation of the retina. Invest Ophthalmol Vis Sci. 2012;53:2720–2731. doi: 10.1167/iovs.11-9344. [DOI] [PubMed] [Google Scholar]
  49. Ratliff CP, Borghuis BG, Kao YH, Sterling P, Balasubramanian V. Retina is structured to process an excess of darkness in natural scenes. Proc Natl Acad Sci. 2010;107:17368–17373. doi: 10.1073/pnas.1005846107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Raviola E, Dacheux RF. Excitatory dyad synapse in rabbit retina. Proc Natl Acad Sci. 1987;84:7324–7328. doi: 10.1073/pnas.84.20.7324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rizzo JF, III, Wyatt J, Loewenstein J, Kelly S, Shire D. Perceptual efficacy of electrical stimulation of human retina with a microelectrode array during short-term surgical trials. Invest Ophthalmol Vis Sci. 2003;44:5362–5369. doi: 10.1167/iovs.02-0817. [DOI] [PubMed] [Google Scholar]
  52. Rockhill RL, Daly FJ, MacNeil MA, Brown SP, Masland RH. The diversity of ganglion cells in a mammalian retina. J Neurosci. 2002;22:3831–3843. doi: 10.1523/JNEUROSCI.22-09-03831.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sanes JR, Masland RH. The types of retinal ganglion cells: Current status and implications for neuronal classification. Annu Rev Neurosci. 2015 doi: 10.1146/annurev-neuro-071714-034120. [DOI] [PubMed] [Google Scholar]
  54. Santos A, Humayun MS, de Juan E, Jr, Greenberg RJ, Marsh MJ, Klock IB, Milam AH. Preservation of the inner retina in retinitis pigmentosa. Arch Ophthalmol. 1997;115:511–515. doi: 10.1001/archopht.1997.01100150513011. [DOI] [PubMed] [Google Scholar]
  55. Sekirnjak C, Hottowy P, Sher A, Dabrowki W, Litke AM, Chichilnisky EJ. High-resolution electrical stimulation of primate retina for epiretinal implant design. J Neurosci. 2008;28:4446–4456. doi: 10.1523/JNEUROSCI.5138-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Shivdasani MN, Sinclair NC, Dimitrov PN, Varsamidis M, Ayton LN, Luu CD, Perera T, McDermott HJ, Blamey PJ. Factors affecting perceptual thresholds in a suprachoroidal retinal prosthesis. Invest Ophthalmol Vis Sci. 2014;55:6467–6481. doi: 10.1167/iovs.14-14396. [DOI] [PubMed] [Google Scholar]
  57. Stasheff SF. Emergence of sustained spontaneous hyperactivity and temporary preservation of OFF responses in ganglion cells of the retinal degeneration (rd1) mouse. J Neurophysiol. 2008;99:1408–1421. doi: 10.1152/jn.00144.2007. [DOI] [PubMed] [Google Scholar]
  58. Stingl K, Bartz-Schmidt KU, Besch D, Braun A, Bruckmann A, Gekeler F, Greppmaier U, Hipp S, Hörtdörfer G, Kernstock C, Koitschev A, Kusnyerik A, Sachs H, Schatz A, Stingl KT, Peters T, Wilhelm B, Zrenner E. Artificial vision with wirelessly powered subretinal electronic implant alpha-IMS. Proc R Soc B. 2013;280:20130077. doi: 10.1098/rspb.2013.0077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Stingl K, Bartz-Schmidt KU, Besch D, Chee CK, Cottrial C, Gekeler F, Groppe M, Jackson TL, MacLaren RE, Koitschev A, Kusnyerik A, Neffendorf J, Nemeth J, Naeem MAN, Peters T, Sachs H, Simpson A, Singh MS, Wilhelm B, Wong D, Zrenner E. Subretinal visual implant Alpha-IMS – Clinical trial interim report. Vis Res. 2015 doi: 10.1016/j.visres.2015.03.001. [DOI] [PubMed] [Google Scholar]
  60. Stone JL, Barlow WE, Humayun MS, de Juan E, Jr, Milam AH. Morphometric analysis of macular photoreceptors and ganglion cells in retinas with retinitis pigmentosa. Arch Ophthalmol. 1992;110:1634–1639. doi: 10.1001/archopht.1992.01080230134038. [DOI] [PubMed] [Google Scholar]
  61. Strettoi E, Pignatelli V. Modification of retinal neurons in a mouse model of retinitis pigmentosa. Proc Natl Acad Sci. 2000;97:11020–11025. doi: 10.1073/pnas.190291097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Strettoi E, Porciatti V, Falsini B, Pignatelli V, Rossi C. Morphological and functional abnormalities in the inner retina of the rd/rd mouse. J Neurosci. 2002;22:5492–5504. doi: 10.1523/JNEUROSCI.22-13-05492.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Tsai D, Morley JW, Suaning GJ, Lovell NH. Direct activation and temporal response properties of rabbit retinal ganglion cells following subretinal stimulation. J Neurophysiol. 2009;102:2982–2993. doi: 10.1152/jn.00545.2009. [DOI] [PubMed] [Google Scholar]
  64. Twyford P, Cai C, Fried S. Differential responses to high-frequency electrical stimulation in ON and OFF retinal ganglion cells. J Neural Eng. 2014;11:025001. doi: 10.1088/1741-2560/11/2/025001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Twyford P, Fried S. The retinal response to sinusoidal electrical stimulation. IEEE Trans Neural Syst Rehabil Eng. 2015 doi: 10.1109/TNSRE.2015.2415811. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Wässle H. Parallel processing in the mammalian retina. Nat Rev Neurosci. 2004;5:747–757. doi: 10.1038/nrn1497. [DOI] [PubMed] [Google Scholar]
  67. Wilke R, Gabel VP, Sachs H, Schmidt KUB, Gekeler F, Besch D, Szurman P, Stett A, Wilhelm B, Peters T, Harscher A, Greppmaier U, Kibbel S, Benav H, Bruckmann A, Stingl K, Kusnyerik A, Zrenner E. Spatial resolution and perception of patterns mediated by a subretinal 16-electrode array in patients blinded by hereditary retinal dystrophies. Invest Ophthalmol Vis Sci. 2011;52:5995–6003. doi: 10.1167/iovs.10-6946. [DOI] [PubMed] [Google Scholar]
  68. Zrenner E. Will retinal implants restore vision? Science. 2002;295:1022–1025. doi: 10.1126/science.1067996. [DOI] [PubMed] [Google Scholar]
  69. Zrenner E, Bartz-Schmidt KU, Benav H, Besch D, Bruckmann A, Gabel V, Gekeler F, Greppmaier U, Harscher A, Kibbel S, Koch J, Kusnyerik A, Peters T, Stingl K, Sachs H, Stett A, Szurman P, Wilhelm B, Wilke R. Subretinal electronic chips allow blind patients to read letters and combine them to words. Proc R Soc B. 2011;278:1489–1497. doi: 10.1098/rspb.2010.1747. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES