Design and validation of a foldable and photovoltaic wide-field epiretinal prosthesis

Abstract

Retinal prostheses have been developed to fight blindness in people affected by outer retinal layer dystrophies. To date, few hundred patients have received a retinal implant. Inspired by intraocular lenses, we have designed a foldable and photovoltaic wide-field epiretinal prosthesis (named POLYRETINA) capable of stimulating wireless retinal ganglion cells. Here we show that within a visual angle of 46.3 degrees, POLYRETINA embeds 2215 stimulating pixels, of which 967 are in the central area of 5 mm, it is foldable to allow implantation through a small scleral incision, and it has a hemispherical shape to match the curvature of the eye. We demonstrate that it is not cytotoxic and respects optical and thermal safety standards; accelerated ageing shows a lifetime of at least 2 years. POLYRETINA represents significant progress towards the improvement of both visual acuity and visual field with the same device, a current challenging issue in the field.

Fig. 1

Foldable and photovoltaic wide-field retinal prosthesis. a 3D model of the fabricated PDMS-interface and of the dome-shaped PDMS support. b 3D model of the retinal prosthesis after boding the PDMS-interface to the PDMS support. c Fabricated PDMS–photovoltaic interface with pixels arranged in three areas of different sizes and densities: central area (red), diameter of 5 mm, 967 electrodes in hexagonal arrangement, electrode diameter 80 µm and pitch 150 µm, density 49.25 px mm⁻²; first ring (green), diameter of 8 mm, 559 electrodes in hexagonal arrangement, electrode diameter 130 µm and pitch 250 µm, density 17.43 px mm⁻²; second ring (blue), diameter 12.7 mm, 719 electrodes, electrode diameter 130 µm, density 9.34 px mm⁻². Circles show an enlarged view of the pixels distribution. Scale bar is 2.5 mm. d Picture of POLYRETINA. Four anchoring wings with holes are present for attaching the prosthesis with retinal tacks. e POLYRETINA folded before injection. f Scanning electron microscope image (40° tilted view) of a photovoltaic pixel. Scale bar is 10 µm. g 3D model after epiretinal placement

Fig. 2

Simulated surgical implantation. a Picture sequence of the implantation in a human eye plastic model. The white line in top-right panel shows the incision of 6.5 mm. b Picture of POLYRETINA placed in epiretinal configuration. c Picture sequence of the implantation in a pig eye

Fig. 3

Optimization of the photovoltaic pixel. a Picture of the KPFM measures. b Sketch of the fabricated device. Glass substrates have been coated with a thin film of ITO (200 nm), a thin film of PEDOT:PSS (50 nm), a thin film of P3HT:PCBM (100 nm), and last aluminium (100 nm) or titanium (150 nm). c Representative KPFM map on a Glass/PEDOT:PSS/Blend/Al device obtained by repeating a line scan of 100 nm (vertical direction). The horizontal bar indicates period of dark (black) and light (white). The bottom panel shows the average potential fluctuation during time; each point is the average potential in a single line scan. d Surface potential variations (voltage in light—voltage in dark) for 6 different architectures. Each bar is the mean (±s.e.m.) of at least N = 3 devices, in which at least n = 3 electrodes/points has been measured and averaged. ITO/PEDOT:PSS/Blend/Al: 0.2106 ± 0.0092 V, N = 5, n = 3; PEDOT:PSS/Blend/Al: 0.2259 ± 0.0085 V, N = 5, n = 3; Blend/Al: 0.1334 ± 0.0090 V, N = 3, n = 3; ITO/PEDOT:PSS/Blend: 0.0128 ± 0.0032 V, N = 3, n = 3; PEDOT:PSS/Blend: 0.0091 ± 0.0025 V, N = 3, n = 4; Blend: 0.0052 ± 0.0007 V, N = 3, n = 4. One-way ANOVA, p < 0.0001, F = 177.9. e Surface potential variations with/without a bottom PDMS layer and with Al or Ti top contacts of 100 and 150 µm in diameter. Each point is the mean (±s.e.m.) of at least N = 3 devices, in which at least n = 3 electrodes has been measured and averaged. PEDOT:PSS/Blend/Al-100 µm: 0.1984 ± 0.0043 V, N = 3, n = 3; PEDOT:PSS/Blend/Al-150 µm: 0.2232 ± 0.0082 V, N = 3, n = 3; PDMS/PEDOT:PSS/Blend/Al-100 µm: 0.1927 ± 0.0115 V, N = 5, n = 3; PDMS/PEDOT:PSS/Blend/Al-150 µm: 0.2163 ± 0.0150 V, N = 5, n = 3; PDMS/PEDOT:PSS/Blend/Ti-100 µm: 0.1055 ± 0.0063 V, N = 3, n = 6; PDMS/PEDOT:PSS/Blend/Ti-150 µm: 0.1342 ± 0.0068 V, N = 3, n = 3. f Representative AFM images of PEDOT:PSS/Blend, PEDOT:PSS/Blend/Al, and PEDOT:PSS/Blend/Ti surfaces

Fig. 4

Characterization of the photo-current and photo-voltage. a Drawing of the experimental setup for the measure of PC and PV; the light pulse comes from the bottom. b, c Examples of PC density (b) and PV (c) measures obtained from 1 electrode (diameter 100 µm) at maximal light intensity (565 nm, 943.98 µW mm⁻²) and for increasing pulse durations (10, 50, 100, and 200 ms). Horizontal bars represent the light pulses. d,e, Mean (±s.e.m) PC density (d) and PV(e) measured upon illumination with 10 ms pulses at increasing light intensities. f, g, Mean (±s.e.m) PC density (f) and PV (g) measured for increasing light intensities (12.75, 111.11, 225.00, 430.56, 616.67, 785.65, and 943.98 µW mm⁻²) and pulse durations (10, 50, 100, and 200 ms). In panels d to g, the PC density and PV on every device (N = 3) has been measured for all electrodes (n = 6) and data have been averaged

Fig. 5

High-frequency train stimulation. a Mean PV trace obtained at maximal light intensity (565 nm, 10 ms, 943.98 µW mm⁻²). The trace is the mean of N = 6 devices; in which n = 6 electrodes have been measured and averaged. The horizontal bars represent the light pulse. The dotted lines highlight the discharging rate of the electrode. b Evolution of the PC density peaks during 1000 stimuli delivered at 1 Hz (10 ms, 943.98 µW mm⁻²). Each point is the mean (±s.e.m.) of N = 3 devices, in which n = 6 electrodes have been measured and averaged. c Representative PV recording upon 10 pulses at 10 Hz (565 nm, 10 ms, 943.98 µW/mm²). d Representative PV recording upon 20 pulses at 20 Hz (565 nm, 10 ms, 943.98 µW mm⁻²). e Evolution of the PC density peaks normalized to the first pulse. Each point is the mean ± s.e.m. of N = 10 devices, in which n = 6 electrodes have been measured and averaged. f Evolution of the PC density peaks normalized to the first pulse. Each point is the mean ± s.e.m. of N = 8 devices, in which n = 6 electrodes have been measured and averaged. g PC generated with 320,000 stimuli delivered at 20 Hz (565 nm, 10 ms, 943.98 µW mm⁻²). Each point is the mean ± s.d. of n = 2 electrodes from N = 1 device

Fig. 6

Evaluation ex vivo with retinal explants. a Sketch of the recording set-up together with a picture of a retinal explant over the PDMS–photovoltaic interface with the metal electrode used for recordings. Scale bar is 100 µm. b Representative single-sweep recording from a retinal ganglion cell over PDMS–photovoltaic interface upon 10-ms illumination at 1081.7 µW mm⁻². The red dotted line is the threshold set for spike detection. The green bar represents the light pulse. The blue insert shows a magnification of the period around the light pulse. The asterisk indicates the over-threshold spike detected, while the gray arrows are the on-set and off-set stimulation artifacts. c Mean (±s.e.m.) firing rate (circles) and firing probability (squares) of SL spikes, computed across all the recorded cells (n = 39, 10 sweeps each) on the PDMS–photovoltaic interface. For each cell, the probability has been defined as the percentage of sweeps with at least a SL spike over the 10 consecutive trials. d Mean (±s.e.m.) latency (circles) and jitter (squares) of the first spike occurring in the 10 ms window after the light onset, computed across all the recorded cells (n = 39, 10 sweeps each) on the PDMS–photovoltaic interface. For each cell, the mean latency and jitter has been computed over the ten consecutive trials. e, f Mean (±s.e.m.) firing rate of medium (e) and long (f) latency spikes, computed across all the recorded cells (n = 39, ten sweeps each) on the PDMS–photovoltaic interface. In panels c–f values have been plotted up to 3 mW mm⁻², while the full profiles are shown in Supplementary Fig. 3c–f

Fig. 7

Spatial confinement of the prosthetic stimulation. a Sketch of the experimental setting. The green circle corresponds to the area illuminated around the central pixel. Gray circles represent the illuminated pixel and the six surrounding ones. The voltage has been measured in nine positions (red dots) for each direction (D1, D2, and D3), all cantered in the center of the illuminated pixel. b Picture during recordings. The light spot is visible (brighter area). The scale bar is 100 µm. c Voltage spreading colour map generated by interpolating the experimental measures with a triangulation-based linear interpolation. At each point ten consecutive recordings have been averaged and the voltage peaks have been normalized with respect to the value obtained in the central pixel (position 1 in a). The green circle is the illuminated area, while the gray circles represent the pixels. d Mean (±s.e.m.) normalized PV peaks from n = 4 pixels. For each pixel, the data from the three directions have been averaged. The red line shows a Gaussian fitting, while the blue dotted line represents the normalized voltage profile obtained by FEA simulations. The gray dotted lines show the FWHM value for the simulated profile. e FEA simulations for three beam sizes, normalized to the potential corresponding to the illumination of the single central pixel. f FEA simulations for various patterns of activation normalized to the potential corresponding to the illumination of the single central pixel

Fig. 8

Lifetime of the retinal prosthesis. a Pictures of the titanium cathodes before (left column) and after (right column) bonding on the dome-shaped PDMS support. The top row is without SU-8 rigid platforms, while the bottom row is with SU-8 rigid platforms. b Picture of a KPFM measure on bonded prostheses integrating SU-8 rigid platforms. c Comparison of KPFM measures on bonded prostheses integrating SU-8 rigid platforms (99.35 ± 25.26 mV, mean ± s.d., n = 15; electrode diameter 80 µm) with respect to measures on PDMS-interface bonded to a planar glass substrate (105.50 ± 17.79 mV, mean ± s.d., n = 36; electrode diameter 100 µm). d Sketch of the accelerated ageing tests. KPFM measures have been performed at the beginning of the experiment, then prostheses have been immersed in saline solution at 87 °C and 100% humidity for 135 h, after that KPFM has been repeated, and on for four cycles. e Quantification (mean ± s.d., N = 4 prostheses, n = 4 electrodes per prosthesis) of the surface potential changes (voltage in light—voltage in dark) during accelerated ageing tests over a simulated period of 24 months (months: 0, 110.5 ± 33.53 mV; 6, 108.5 ± 33.37 mV; 12, 109.8 ± 44.59 mV; 18, 103.8 ± 25.73 mV; 24, 111.1 ± 35.48 mV)

Fig. 9

Temperature variation during operation. a The top surface of POLYRETINA has been imaged with a thermal camera while pulsed illumination has been provided from the bottom, as in the epiretinal configuration. The camera has been focused on the top electrodes and a ROI has been selected to measure the changes in surface temperature (cyan circle). Electrodes show higher value of baseline temperature because the metallic surface reflects part of the IR light used for the measurement. b Mean (±s.d.) changes in surface temperature measured in N = 4 prostheses. Data have been plotted has difference with respect to the baseline temperature measured for 5 min before pulsed illumination. The green bar represents the period of 2 h when light pulses have been applied (10 ms pulses, 20 Hz repetition rate, 1.22 mW mm⁻²). The dotted red line represents the maximal allowed temperature increase. c Mean (±s.e.m.) changes in surface temperature measured on the electrodes (left, N = 4 prostheses) or on the polymer area (right, N = 4 prostheses). For each prosthesis, n = 3 electrodes/areas have been sampled and averaged. d Mean (±s.d.) changes in surface temperature in the average surface, the electrode area or the polymer area are not significantly different (1.24 ± 0.29, 1.23 ± 0.20, 1.31 ± 0.21, respectively; one-way ANOVA, F = 0.0569, p = 0.9451)

Fig. 10

FEA simulation of thermal effects with POLYRETINA. a Temperature increase in the modeled eye with POLYRETINA after 150 s of continuous illumination (CW, 560 nm, 328 µW mm⁻²). The insert shows a larger view of the modeled retina and POLYRETINA. b Time course of the temperature increase in the modeled retina during 150 s of continuous illumination (CW, 560 nm, 328 µW mm⁻²). The simulation frequency has been set to 1 Hz. The solid line is the log Gaussian fit (R² = 0.9934). c Probability of retinal damage as a function of irradiance with (red) and without (black) POLYRETINA. ED50 corresponds to a temperature increase of 12.5 °C. The irradiance has been expressed for pulsed illumination (20% of duty cycle). The solid lines are the Sigmoidal fits (R² = 0.9971 for black and 0.9977 for red)