Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Avoidance of axonal activation in epiretinal implants using short biphasic pulses

Avoidance of axonal activation in epiretinal implants using short biphasic pulses DE GRUYTER Current Directions in Biomedical Engineering 2022;8(3): 05-08 Paul Werginz*, Andrea Corna and Günther Zeck Avoidance of axonal activation in epiretinal implants using short biphasic pulses https://doi.org/10.1515/cdbme-2022-2002 cells close to a stimulating electrode and cells with their cell bodies far distant but axons that traverse below the stimulating Abstract: Retinal implants allow patients suffering from electrode. Because of concurrent cell activation, the resulting degenerative retinal diseases to regain visual percepts. Despite percept has been reported to be streak-like [2] and therefore an considerable effort in developing and improving retinal accurate pixelized version of the outer world cannot be neuroprostheses in the last decades, the restored vision in created. Several strategies to avoid axonal activation such as patients does not achieve sufficient quality. The elicited long biphasic pulses [3] as well as bi-electrode stimulation [4] percepts do not accurately match the stimulating pixels and have been proposed. Here, we examine the response of RGCs therefore the formation of high acuity vision is hindered. One to epiretinal stimulation for short pulse durations using a of the main obstacles in epiretinal implants is the concurrent computational model. In experiments, short pulses in the range activation of cells close to a stimulating electrode and cells that of 100 μs have been shown to increase the difference of axonal have their axons traversing the electrode. In this study, we use versus focal thresholds [3] and therefore we sought to study computational modeling to examine the effect of pulse responses for even shorter pulses down to 10 μs. Our results duration on the selectivity of focal versus non-focal activation. suggest that shorter pulses increase the axonal/focal threshold Our results suggest that biphasic pulses in the range of 10-20 ratio and therefore may be applicable to generate more focal microseconds can prevent axonal activation while still reliably responses during epiretinal stimulation. activating target neurons. Keywords: retinal implant, electrical stimulation, retinal ganglion cells, microelectrode array 2 Methods To study the response of RGCs to electrical stimulation in- silico, we used morphologically- and biophysically-realistic 1 Introduction multicompartment models of mouse ON-alpha RGCs [5]. Extracellular electric stimulation was applied via disk Today, patients suffering from retinal degeneration have the electrodes from the epiretinal side. Extracellular potentials possibility to regain a rudimentary kind of vision by retinal were calculated based on an analytic approach [6]. The prostheses [1]. Modern neurotechnology allows the remaining, modeled retina consisted of 100 RGCs in a rectangle 500x150 healthy cells in the retina of blind patients to be stimulated μm (Fig. 1A). The z-distance between each cell’s soma center electrically and therefore to convey visual information to the and the electrode was 15 μm. We tested rectangular brain. Aside from the technical and surgical challenges monophasic and biphasic symmetric charge balanced current involved in the development of a retinal implant, many other pulses with durations of 10, 20, 50, 100 and 200 μs (Fig. 1B ). obstacles related to the specific layout and complex Additional simulations were performed with real pulse shapes physiology of the retina have to be overcome. The basic idea applied by a CMOS-based Microelectrode array (MEA) [7] behind a retinal implant is to focally activate target neurons in and measured as described [8] (Fig. 1B2). For the used MEA the vicinity of a stimulating pixel (electrode) and therefore pulse duration was limited to ≥20 μs. Activation threshold was create a pixelized version of the visual scene. In epiretinal computed for each cell in the retina and thresholds were implants, with retinal ganglion cells (RGCs) as target neurons, compared between cells below the electrode (focal, 1 in Fig. a problem of particular interest is the concurrent activation of 1) and cells that had their axons passing the stimulating electrode (axonal, 2 in Fig. 1). Threshold ratio for each cell was computed as Thresh / ThreshMin with ThreshMin being the ______ *Corresponding author: Paul Werginz: TU Wien, Institute of overall minimum threshold across all cells. Biomedical Electronics, Gusshausstraße 27-29, Vienna, Austria, e-mail: paul.werginz@tuwien.ac.at Andrea Corna, Günther Zeck: TU Wien, Institute of Biomedical Electronics, Gusshausstraße 27-29, Vienna, Austria Open Access. © 2022 The Author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 5 a window of opportunity to activate target cells without activating passing axons. Threshold ratios for all pulses tested can be found in Table 1. We also plotted transmembrane voltage over time for two RGCs at different stimulus amplitudes (Fig. 2B). One cell was located directly below the stimulating electrode whereas the other cell had its soma located approximately 300 μm distant of the electrode center. The threshold for focal cell activation was ~53 μA. Stimulus amplitude was increased to 2, 3 and 4x threshold amplitude to examine the resulting spiking activity in the two cells. Even at 3x threshold focal cell activation was possible (Fig. 2B, red traces), i.e., whereas the cell below the electrode initiated a spike the distant cell stayed silent (Fig. 2B, blue traces). Not until 4x threshold amplitude the distant cell fired an action potential. Figure 1: In-silico retina model and stimuli used for simulations. A) 100 RGCs (white circles) were distributed evenly on the retinal surface. The stimulus electrode (1) was 128 μm in diameter. Threshold ratios were compared between cells below locations 1 and 2, respectively. RGC axon direction is indicated by the black arrow. B) Idealized mono- (left) and biphasic (right) pulses (B ) as well as measured current output from a CMOS-based Microelectrode Array(B ). 3 Results This study examines RGC thresholds in response to epiretinal stimulation. Our goal was to investigate the influence of current pulse duration on both, cells close to the stimulating electrode as well as cells far distant that have their axons traversing the stimulating electrode. 3.1 Focal vs. axonal response threshold We compared focal and non-focal (axonal) activation by computing thresholds for all 100 cells in the model retina and further calculating the ratio between each cell’s threshold and the lowest threshold of all RGCs (Fig. 2A). For the 200 μs pulse, thresholds were in the range of 2.2-7.1 μA with threshold ratios rather uniform across the whole retinal surface Figure 2: Short biphasic pulses prevent axonal activation. A) For (Fig. 2A, top). Thus, axonal thresholds were in the same range each of the 100 simulated RGCs activation threshold was as thresholds for cells right below the stimulation electrode computed in response to the idealized 200 (top) and 10 μs and focal stimulation of target cells is not achievable. The (bottom) pulse. Cells were color-coded by their threshold shortest pulse duration tested (10 μs), on the other hand, ratio. The minimum threshold of all RGCs is indicated by the white ‘x’. B) Spiking activity in two cells marked by the blue resulted in a distinct low-threshold region around the and red arrowheads in A (bottom). Transmembrane voltage stimulating electrode and higher axonal thresholds by up to a is plotted over time for each cell’s distal axon for 4 different factor of 15 (Fig. 2A, bottom). On average, threshold ratios for stimulus amplitudes. Arrowheads indicate action potentials. cells far distant of the electrode were 4.1 times higher, creating 6 3.2 Realistic pulse shape and idealized pulses Our results using biphasic current pulses suggested that pulses in the range of 10 μs strongly increase axonal vs. focal threshold ratios and are thus useful to create more focal regions of activation. So far, however, only idealized biphasic pulses were applied which may not be easily generated by neurostimulators. Therefore, we measured the current output Figure 3: Peak current amplitude (left) and stimulation charge from a CMOS-based MEA [7,8] for different pulse durations (right) at threshold for all cells. (Fig. 1B ) and used the measurements as inputs for our simulations. Similar to results from the idealized biphasic µs) for the idealized pulses and from 0.76 (200 µs) to 0.90 nC waveforms also measured waveforms resulted in increased (20 µs) for the measured waveforms. threshold ratios allowing for focal stimulation of RGCs (Table 1). For example, mean threshold ratios for idealized and measured pulses were 5.25 and 4.47, respectively. Monophasic pulses, on the other hand, resulted in substantially 4 Discussion lower threshold ratios indicating that the second balancing pulse in the biphasic pulse configuration is key for preventing We modeled the response of a population of retinal ganglion axonal activation. cells to epiretinal stimulation and found that short biphasic pulses affect the ratio between axonal and focal activation. We Table 1: Threshold ratios for RGC activation. Numbers in red did not investigate the underlying mechanisms but one likely indicate the mean threshold ratio for cells in (1) of Fig. 1A. Numbers factor to contribute to the differential responses for various in black indicate the mean threshold ratio for cells in (2) of Fig. 1A. pulse durations could be the Axon Initial Segment due to its high density of sodium channels [5]. A second potential Mean threshold ratios Pulse duration (μs) mechanism could be the much larger somatic capacitance in Monophasic Biphasic comparison to the axonal capacitance. Our results suggest that aside from the previously tested stimulation strategies to avoid axonal activation such as long duration pulses [3] and bi-electrode stimulation [4] also short 10 2.79/3.14 1.90/7.80 | ----- biphasic pulses <50 μs can focally activate target cells close to 20 2.57/2.89 2.77/5.25 | 2.64/4.47 the stimulating electrode without activating passing axons. To 50 2.23/2.42 2.20/3.38 | 2.25/3.25 test the simulation results in experiments will be challenging 100 1.88/1.94 1.89/2.44 | 1.93/2.39 as generating very short rectangular pulses may be difficult with current MEAs. Especially constraints on power 200 1.58/1.54 1.54/1.72 | 1.60/1.74 consumption have to be considered when implementing short pulses. The used CMOS-based MEA can apply approximately 1.2 nC of charge/phase for an electrode 128x128 μm in size. 3.3 Stimulation charge with short pulses In our simulations, thresholds for cells close to the stimulating electrode were in the range of 0.6 nC when 10 μs biphasic Short pulse durations require higher current levels to generate pulses were used (Fig. 3, right). Thresholds are strongly neural activity, and therefore we asked how the stimulation dependent on the electrode-to-cell distance which in our study charge scales with decreasing pulse duration. We compared was set to 15 μm; a tight retina-MEA interface could therefore peak current and charge injection for all cells at threshold (Fig. allow for the application of short pulses. Our measured data 3). Biphasic pulses required higher peak currents than shows that for pulses shorter than 100 μs the resulting current monophasic pulses for each pulse duration. Measured waveforms can be strongly distorted (Fig. 1B ). However, our waveforms resulted in similar peak current as well as slightly simulations indicate that the non-rectangular waveforms have lower charge levels when compared to the idealized biphasic a similar effect on threshold ratios (Table 1) and therefore pulses. The stimulation charge (calculated during the cathodic could still be beneficial to achieve focal cell activation. phase) increased moderately from 0.80 (200 µs) to 1.22 nC (20 7 effects on the percepts elicited by retinal stimulation. IOVS. Additionally, pulse shape could be optimized by modifying the 2012;53(1):205-214. input voltage waveform. [3] Weitz AC, Nanduri D, Behrend MR, Gonzales-Calle A, Greenberg RJ, et al. Improving the spatial resolution of epiretinal implants by increasing stimulus pulse durations. Sci Transl Med. 2015;7(318):318ra203-318ra203. Author Statement [4] Vilkhu RS, Madugula SS, Grosberg LE, Gogliettino AR, Research funding: Financial support was provided by the Hottowy P, et al. Spatially patterned bi-electrode epiretinal Austrian Science Fund (FWF P35488) and by the European stimulation for axon avoidance at cellular resolution. J Neural Union’s Horizon 2020 research and innovation program under Eng. 2021;18(6):066007 [5] Werginz P, Raghuram V, Fried SI. The relationship between the Marie Skłodowska-Curie grant agreement No 861423 morphological properties and thresholds to extracellular (Entrain Vision). Conflict of interest: Authors state no conflict electric stimulation in α RGCs. J Neural Eng. of interest. 2020;17(4):045015. [6] Newman J. Resistance for flow of current to a disk. J Electrochem Soc. 1966, [7] Bertotti G, Velychko D, Dodel N, Wolansky D, Tillak B, et al. References A CMOS-based sensor array for in-vitro neural tissue interfacing with 4225 recording sites and 1024 stimulation [1] Palanker D, Le Mer Y, Mohand-Said S, Muqit MMK, Sahel sites. 2014 IEEE BioCAS Proceedings. 2014. JA. Photovoltaic restoration of central vision in atrophic age- related macular degeneration. Ophthalmology. [8] Corna A, Ramesh P, Jetter F, Lee MJ, Macke JH, et al. Discrimination of simple objects decoded from the output of 2020;127(8):1097-1104. [2] Nanduri D, Fine I, Horsager A, Boynton GM, Humayun MS, retinal ganglion cells upon sinusoidal electrical stimulation. J Neural Eng. 2021;18(4):046086 et al. Frequency and amplitude modulation have different http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Current Directions in Biomedical Engineering de Gruyter

Avoidance of axonal activation in epiretinal implants using short biphasic pulses

Download PDF
4 pages

Loading next page...
 
/lp/de-gruyter/avoidance-of-axonal-activation-in-epiretinal-implants-using-short-RP15OJSc0H
Publisher
de Gruyter
Copyright
© 2022 The Author(s), published by De Gruyter
eISSN
2364-5504
DOI
10.1515/cdbme-2022-2002
Publisher site
See Article on Publisher Site

Abstract

DE GRUYTER Current Directions in Biomedical Engineering 2022;8(3): 05-08 Paul Werginz*, Andrea Corna and Günther Zeck Avoidance of axonal activation in epiretinal implants using short biphasic pulses https://doi.org/10.1515/cdbme-2022-2002 cells close to a stimulating electrode and cells with their cell bodies far distant but axons that traverse below the stimulating Abstract: Retinal implants allow patients suffering from electrode. Because of concurrent cell activation, the resulting degenerative retinal diseases to regain visual percepts. Despite percept has been reported to be streak-like [2] and therefore an considerable effort in developing and improving retinal accurate pixelized version of the outer world cannot be neuroprostheses in the last decades, the restored vision in created. Several strategies to avoid axonal activation such as patients does not achieve sufficient quality. The elicited long biphasic pulses [3] as well as bi-electrode stimulation [4] percepts do not accurately match the stimulating pixels and have been proposed. Here, we examine the response of RGCs therefore the formation of high acuity vision is hindered. One to epiretinal stimulation for short pulse durations using a of the main obstacles in epiretinal implants is the concurrent computational model. In experiments, short pulses in the range activation of cells close to a stimulating electrode and cells that of 100 μs have been shown to increase the difference of axonal have their axons traversing the electrode. In this study, we use versus focal thresholds [3] and therefore we sought to study computational modeling to examine the effect of pulse responses for even shorter pulses down to 10 μs. Our results duration on the selectivity of focal versus non-focal activation. suggest that shorter pulses increase the axonal/focal threshold Our results suggest that biphasic pulses in the range of 10-20 ratio and therefore may be applicable to generate more focal microseconds can prevent axonal activation while still reliably responses during epiretinal stimulation. activating target neurons. Keywords: retinal implant, electrical stimulation, retinal ganglion cells, microelectrode array 2 Methods To study the response of RGCs to electrical stimulation in- silico, we used morphologically- and biophysically-realistic 1 Introduction multicompartment models of mouse ON-alpha RGCs [5]. Extracellular electric stimulation was applied via disk Today, patients suffering from retinal degeneration have the electrodes from the epiretinal side. Extracellular potentials possibility to regain a rudimentary kind of vision by retinal were calculated based on an analytic approach [6]. The prostheses [1]. Modern neurotechnology allows the remaining, modeled retina consisted of 100 RGCs in a rectangle 500x150 healthy cells in the retina of blind patients to be stimulated μm (Fig. 1A). The z-distance between each cell’s soma center electrically and therefore to convey visual information to the and the electrode was 15 μm. We tested rectangular brain. Aside from the technical and surgical challenges monophasic and biphasic symmetric charge balanced current involved in the development of a retinal implant, many other pulses with durations of 10, 20, 50, 100 and 200 μs (Fig. 1B ). obstacles related to the specific layout and complex Additional simulations were performed with real pulse shapes physiology of the retina have to be overcome. The basic idea applied by a CMOS-based Microelectrode array (MEA) [7] behind a retinal implant is to focally activate target neurons in and measured as described [8] (Fig. 1B2). For the used MEA the vicinity of a stimulating pixel (electrode) and therefore pulse duration was limited to ≥20 μs. Activation threshold was create a pixelized version of the visual scene. In epiretinal computed for each cell in the retina and thresholds were implants, with retinal ganglion cells (RGCs) as target neurons, compared between cells below the electrode (focal, 1 in Fig. a problem of particular interest is the concurrent activation of 1) and cells that had their axons passing the stimulating electrode (axonal, 2 in Fig. 1). Threshold ratio for each cell was computed as Thresh / ThreshMin with ThreshMin being the ______ *Corresponding author: Paul Werginz: TU Wien, Institute of overall minimum threshold across all cells. Biomedical Electronics, Gusshausstraße 27-29, Vienna, Austria, e-mail: paul.werginz@tuwien.ac.at Andrea Corna, Günther Zeck: TU Wien, Institute of Biomedical Electronics, Gusshausstraße 27-29, Vienna, Austria Open Access. © 2022 The Author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 5 a window of opportunity to activate target cells without activating passing axons. Threshold ratios for all pulses tested can be found in Table 1. We also plotted transmembrane voltage over time for two RGCs at different stimulus amplitudes (Fig. 2B). One cell was located directly below the stimulating electrode whereas the other cell had its soma located approximately 300 μm distant of the electrode center. The threshold for focal cell activation was ~53 μA. Stimulus amplitude was increased to 2, 3 and 4x threshold amplitude to examine the resulting spiking activity in the two cells. Even at 3x threshold focal cell activation was possible (Fig. 2B, red traces), i.e., whereas the cell below the electrode initiated a spike the distant cell stayed silent (Fig. 2B, blue traces). Not until 4x threshold amplitude the distant cell fired an action potential. Figure 1: In-silico retina model and stimuli used for simulations. A) 100 RGCs (white circles) were distributed evenly on the retinal surface. The stimulus electrode (1) was 128 μm in diameter. Threshold ratios were compared between cells below locations 1 and 2, respectively. RGC axon direction is indicated by the black arrow. B) Idealized mono- (left) and biphasic (right) pulses (B ) as well as measured current output from a CMOS-based Microelectrode Array(B ). 3 Results This study examines RGC thresholds in response to epiretinal stimulation. Our goal was to investigate the influence of current pulse duration on both, cells close to the stimulating electrode as well as cells far distant that have their axons traversing the stimulating electrode. 3.1 Focal vs. axonal response threshold We compared focal and non-focal (axonal) activation by computing thresholds for all 100 cells in the model retina and further calculating the ratio between each cell’s threshold and the lowest threshold of all RGCs (Fig. 2A). For the 200 μs pulse, thresholds were in the range of 2.2-7.1 μA with threshold ratios rather uniform across the whole retinal surface Figure 2: Short biphasic pulses prevent axonal activation. A) For (Fig. 2A, top). Thus, axonal thresholds were in the same range each of the 100 simulated RGCs activation threshold was as thresholds for cells right below the stimulation electrode computed in response to the idealized 200 (top) and 10 μs and focal stimulation of target cells is not achievable. The (bottom) pulse. Cells were color-coded by their threshold shortest pulse duration tested (10 μs), on the other hand, ratio. The minimum threshold of all RGCs is indicated by the white ‘x’. B) Spiking activity in two cells marked by the blue resulted in a distinct low-threshold region around the and red arrowheads in A (bottom). Transmembrane voltage stimulating electrode and higher axonal thresholds by up to a is plotted over time for each cell’s distal axon for 4 different factor of 15 (Fig. 2A, bottom). On average, threshold ratios for stimulus amplitudes. Arrowheads indicate action potentials. cells far distant of the electrode were 4.1 times higher, creating 6 3.2 Realistic pulse shape and idealized pulses Our results using biphasic current pulses suggested that pulses in the range of 10 μs strongly increase axonal vs. focal threshold ratios and are thus useful to create more focal regions of activation. So far, however, only idealized biphasic pulses were applied which may not be easily generated by neurostimulators. Therefore, we measured the current output Figure 3: Peak current amplitude (left) and stimulation charge from a CMOS-based MEA [7,8] for different pulse durations (right) at threshold for all cells. (Fig. 1B ) and used the measurements as inputs for our simulations. Similar to results from the idealized biphasic µs) for the idealized pulses and from 0.76 (200 µs) to 0.90 nC waveforms also measured waveforms resulted in increased (20 µs) for the measured waveforms. threshold ratios allowing for focal stimulation of RGCs (Table 1). For example, mean threshold ratios for idealized and measured pulses were 5.25 and 4.47, respectively. Monophasic pulses, on the other hand, resulted in substantially 4 Discussion lower threshold ratios indicating that the second balancing pulse in the biphasic pulse configuration is key for preventing We modeled the response of a population of retinal ganglion axonal activation. cells to epiretinal stimulation and found that short biphasic pulses affect the ratio between axonal and focal activation. We Table 1: Threshold ratios for RGC activation. Numbers in red did not investigate the underlying mechanisms but one likely indicate the mean threshold ratio for cells in (1) of Fig. 1A. Numbers factor to contribute to the differential responses for various in black indicate the mean threshold ratio for cells in (2) of Fig. 1A. pulse durations could be the Axon Initial Segment due to its high density of sodium channels [5]. A second potential Mean threshold ratios Pulse duration (μs) mechanism could be the much larger somatic capacitance in Monophasic Biphasic comparison to the axonal capacitance. Our results suggest that aside from the previously tested stimulation strategies to avoid axonal activation such as long duration pulses [3] and bi-electrode stimulation [4] also short 10 2.79/3.14 1.90/7.80 | ----- biphasic pulses <50 μs can focally activate target cells close to 20 2.57/2.89 2.77/5.25 | 2.64/4.47 the stimulating electrode without activating passing axons. To 50 2.23/2.42 2.20/3.38 | 2.25/3.25 test the simulation results in experiments will be challenging 100 1.88/1.94 1.89/2.44 | 1.93/2.39 as generating very short rectangular pulses may be difficult with current MEAs. Especially constraints on power 200 1.58/1.54 1.54/1.72 | 1.60/1.74 consumption have to be considered when implementing short pulses. The used CMOS-based MEA can apply approximately 1.2 nC of charge/phase for an electrode 128x128 μm in size. 3.3 Stimulation charge with short pulses In our simulations, thresholds for cells close to the stimulating electrode were in the range of 0.6 nC when 10 μs biphasic Short pulse durations require higher current levels to generate pulses were used (Fig. 3, right). Thresholds are strongly neural activity, and therefore we asked how the stimulation dependent on the electrode-to-cell distance which in our study charge scales with decreasing pulse duration. We compared was set to 15 μm; a tight retina-MEA interface could therefore peak current and charge injection for all cells at threshold (Fig. allow for the application of short pulses. Our measured data 3). Biphasic pulses required higher peak currents than shows that for pulses shorter than 100 μs the resulting current monophasic pulses for each pulse duration. Measured waveforms can be strongly distorted (Fig. 1B ). However, our waveforms resulted in similar peak current as well as slightly simulations indicate that the non-rectangular waveforms have lower charge levels when compared to the idealized biphasic a similar effect on threshold ratios (Table 1) and therefore pulses. The stimulation charge (calculated during the cathodic could still be beneficial to achieve focal cell activation. phase) increased moderately from 0.80 (200 µs) to 1.22 nC (20 7 effects on the percepts elicited by retinal stimulation. IOVS. Additionally, pulse shape could be optimized by modifying the 2012;53(1):205-214. input voltage waveform. [3] Weitz AC, Nanduri D, Behrend MR, Gonzales-Calle A, Greenberg RJ, et al. Improving the spatial resolution of epiretinal implants by increasing stimulus pulse durations. Sci Transl Med. 2015;7(318):318ra203-318ra203. Author Statement [4] Vilkhu RS, Madugula SS, Grosberg LE, Gogliettino AR, Research funding: Financial support was provided by the Hottowy P, et al. Spatially patterned bi-electrode epiretinal Austrian Science Fund (FWF P35488) and by the European stimulation for axon avoidance at cellular resolution. J Neural Union’s Horizon 2020 research and innovation program under Eng. 2021;18(6):066007 [5] Werginz P, Raghuram V, Fried SI. The relationship between the Marie Skłodowska-Curie grant agreement No 861423 morphological properties and thresholds to extracellular (Entrain Vision). Conflict of interest: Authors state no conflict electric stimulation in α RGCs. J Neural Eng. of interest. 2020;17(4):045015. [6] Newman J. Resistance for flow of current to a disk. J Electrochem Soc. 1966, [7] Bertotti G, Velychko D, Dodel N, Wolansky D, Tillak B, et al. References A CMOS-based sensor array for in-vitro neural tissue interfacing with 4225 recording sites and 1024 stimulation [1] Palanker D, Le Mer Y, Mohand-Said S, Muqit MMK, Sahel sites. 2014 IEEE BioCAS Proceedings. 2014. JA. Photovoltaic restoration of central vision in atrophic age- related macular degeneration. Ophthalmology. [8] Corna A, Ramesh P, Jetter F, Lee MJ, Macke JH, et al. Discrimination of simple objects decoded from the output of 2020;127(8):1097-1104. [2] Nanduri D, Fine I, Horsager A, Boynton GM, Humayun MS, retinal ganglion cells upon sinusoidal electrical stimulation. J Neural Eng. 2021;18(4):046086 et al. Frequency and amplitude modulation have different

Journal

Current Directions in Biomedical Engineering – de Gruyter

Published: Sep 1, 2022

Keywords: retinal implant; electrical stimulation; retinal ganglion cells; microelectrode array

There are no references for this article.