Flexible photoacoustic apparatus for retinal stimulation

Abstract

A flexible photoacoustic ultrasound transducer film is disclosed for biomedical applications, particularly retinal stimulation and vision restoration. The film comprises an elastomeric polydimethylsiloxane (PDMS) matrix and a carbon-based optical absorber, either as a discrete layer or dispersed throughout the matrix. The film emits ultrasonic waves upon irradiation with pulsed laser light. The ultrasonic waves can have a lateral resolution below 100 µm when illuminated by pulsed laser light (1 µJ to 10 µJ pulse energy), while maintaining a surface temperature rise under 1 °C. Embodiments include a PDMS / candle-soot / PDMS sandwich structure or a carbon-nanotube-loaded PDMS composite.

Description

[0001] Flexible Photoacoustic Apparatus for Retinal Stimulation

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS This Application claims the benefit of and priority to U.S. Provisional Application Ser. No. 63 / 748,739, filed January 23, 2025, the content of which is incorporated herein by reference in its entirety.

[0003] FIELD OF INVENTION

[0004] The present disclosure pertains to biomedical devices, specifically to flexible photoacoustic technologies for retinal stimulation and vision restoration.

[0005] BACKGROUND

[0006] The restoration of visual function through targeted stimulation of remaining retinal neurons has emerged as a significant objective in biomedical engineering, driven by the widespread incidence of photoreceptor degeneration in conditions such as retinitis pigmentosa and age-related macular degeneration. As photoreceptors deteriorate, the downstream retinal circuitry remains intact, offering an opportunity to restore perception by directly engaging inner retinal cells. Research in this field spans implantable electrode arrays, optogenetic approaches, and focused ultrasound methods, each aiming to translate external stimuli into neural activation of retinal ganglion cells. These efforts reflect a broader commitment to leveraging advances in materials science, optics, and acoustics to develop modalities that can interface with delicate ocular tissues without disrupting the native architecture of the eye. Clinically, there is a clear demand for stimulation strategies that deliver high spatial precision over large retinal areas while minimizing surgical complexity and device footprint. Optimal stimulation platforms would avoid rigid substrates and bulky connectors, instead employing conformal, biocompatible interfaces that can be applied subretinally, epiretinally, or even in wearable form. Near-infrared or visible light sources offer the promise of non-invasive power delivery, but require effective transduction mechanisms to generate localized mechanical or electrical cues at defined sites within the retina. Simplified device packaging and streamlined implantation workflows are necessary to broaden patient access and accommodate variable clinical settings. Existing electrical prostheses are challenged by limited pixel density, suboptimal electrode -tissue coupling, andmechanical mismatch between rigid microfabricated substrates and soft retinal tissue, leading to chronic inflammation and other tolerance issues. Optogenetic gene therapies involve viral delivery and raise regulatory and safety concerns that may slow clinical adoption.

[0007] Conventional focused ultrasound techniques face diffraction-limited spot sizes and often require acoustic pressures that exceed ophthalmic safety thresholds, compromising their ability to achieve desired resolution (e.g., sub-100 μm targeting) within the retina. These limitations underscore the need for a new class of retinal stimulation that combine precise spatial control, gentle tissue coupling, and compliance with safety standards for both acoustic and thermal exposure.

[0008] SUMMARY

[0009] Systems and methods of the invention provide a thin, flexible transducer capable of converting pulsed light into ultrasound pressures sufficient to activate mechanosensitive retinal cells with lateral resolution below 100 μm, while maintaining temperature rises under clinical safety limits. Such transducer films can conform to the curvature of the eye, span extensive retinal areas, and integrate without rigid electronics or tethered power leads.

[0010] Furthermore, they can be compatible with near-infrared wavelengths and provide effective ocular penetration, minimal off-target stimulation, and seamless biocompatibility.

[0011] In one embodiment, a flexible photoacoustic ultrasound transducer film includes an elastomeric polydimethylsiloxane (PDMS) matrix and a carbon-based optical absorber disposed either as a discrete layer sandwiched between opposed PDMS portions or dispersed throughout the PDMS. The film has a total thickness not greater than about 200 μm, not greater than about 150 μm, not greater than about 115 μm, or not greater than about 100 μm and, when illuminated by pulsed laser light with individual pulse energies between about 1 pJ and about 10 pJ, emits ultrasonic waves with a lateral spread of less than about 100 μm while limiting the surface temperature rise to less than about 1 °C. In some implementations the absorber is a candle-soot layer of approximately 3 μm thickness between two PDMS layers formed from a 5:1 pre-polymer / curing-agent mixture, each spin-coated at about 500 rpm and cured at about 110 °C, with an optional hydrophilic surface modification produced by one-minute oxygen plasma exposure. In other implementations the absorber comprises carbon nanotubes dispersed at about 15 wt % in the PDMS matrix to yield a composite thickness of about 40 μm. The film may be optimized for pulsed illumination at approximately 1030 nmto generate ultrasonic waves having a central frequency in the range of about 30 MHz to about 60 MHz.

[0012] In another embodiment, a skin-conformal wearable ultrasound patch integrates the flexible photoacoustic film along with an optical coupling designed to receive external pulsed laser radiation and direct the radiation onto the carbon-based absorber. The patch can also include an adhesive or elastomeric support that attaches securely to irregular bodily surfaces while enabling on-body ultrasound generation without relying on rigid piezoelectric elements. In certain configurations, the optical coupling can consist of a flexible optical fiber with an approximate diameter of 200 μm, terminating near the absorber. Additionally, the patch can incorporate a hydrogel coating on the body-facing side to acoustically match impedance between the film and the skin.

[0013] In another embodiment, a photoacoustic retinal implant can comprise the flexible photoacoustic film having a maximum thickness of about 200 μm and may be configured for subretinal placement to generate ultrasonic waves that stimulate adjacent retinal neurons when illuminated by pulsed laser light delivered through ocular media. In particular implementations the implant can use either the carbon-nanotube-loaded PDMS film of about 40 μm thickness or the PDMS / candle-soot / PDMS sandwich film of about 115 pm thickness. In another embodiment, a method for neuromodulating retinal ganglion cells is provided comprising implanting, in a subretinal location, a flexible photoacoustic film with an elastomeric PDMS matrix and a carbon-based optical absorber selected from candle soot or carbon nanotubes, the film having a thickness not exceeding about 200 μm; directing toward a selected region of the film pulsed laser light of approximately 1030 nm wavelength, individual pulse energies between about 1 μJ and about 10 μJ, a repetition rate between about 1 kHz and about 6 kHz, and a focal spot diameter of about 50 μm; and thereby generating localized ultrasonic fields with lateral dimensions below about 100 μm to activate retinal ganglion cells while limiting the temperature increase at the film-tissue interface to less than about 1 °C.

[0014] In various embodiments, ultrasound stimulation of retinal nerves using flexible photoacoustic films described herein may be augmented with sonogenetic methods such as those discussed in Ji, et al., Potential of ultrasound stimulation and sonogenetics in vision restoration: a narrative review, Neural Regen Res, 2025 Dec 1;20(12):3501-3516, the content of which is incorporated herein by reference. As discussed, below, sonogenetics augmentation can include pharmacological and / or genetic manipulations to inhibit or enhance expression ofultrasound or mechanical stimulation sensitive channels in target regions to modulate the strength of ultrasound-induced neurostimulation.

[0015] In some cases the laser can be delivered in bursts of 5 ms to 30 ms to control firing rates, and retinal integrity can be monitored by optical coherence tomography at intervals for at least four months. Ultrasonic generation can produce peak acoustic pressures at the film surface of not more than about 0.6 MPa and operates within FDA-specified ophthalmic safety limits for mechanical index and spatial-peak temporal-average intensity. In another embodiment, a method for fabricating a PDMS / candle-soot / PDMS photoacoustic sandwich film is provided comprising depositing a continuous candle-soot carbon layer of about 3 μm thickness by exposing the cured PDMS to a candle flame for about 20 seconds, spin-coating a first layer of uncured PDMS onto candle soot layer at about 500 rpm; curing the first layer at approximately 110 °C; removing, inverting, and reattaching the spin-coated and cured candle soot layer to the substrate; spin-coating a second layer of uncured PDMS over the exposed side of the soot layer at about 500 rpm; and curing the second layer to obtain a flexible sandwich film of about 115 pm total thickness that emits ultrasonic waves upon pulsed laser irradiation. In another embodiment, a method for producing a carbon-nanotube-loaded PDMS photoacoustic film suitable for subretinal implantation comprises dispersing carbon nanotubes in PDMS pre-polymer to a concentration of about 15 wt % by ultrasonic sonication for approximately five minutes; degassing the dispersion for about 30 minutes; spin-coating the degassed dispersion onto a substrate at about 500 rpm; curing the coated layer at approximately 110 °C; and peeling the cured composite to yield a flexible film of about 40 μm thickness that generates ultrasonic waves when irradiated with pulsed laser light.

[0016] BRIEF DESCRIPTION OF DRAWINGS FIG. 1 shows a schematic and characterization of Characterization of flexible photoacoustic films according to various embodiments.

[0017] FIG. 2 shows photoacoustic modulation of ex vivo wild type and degenerated retinae in various embodiments.

[0018] FIG. 3 shows RGC responses under different laser burst durations and laser repetition rates in various embodiments.

[0019] FIG. 4 shows spatial distribution of RGC modulation upon photoacoustic stimulation in various embodiments.FIG. 5 shows in vivo photoacoustic implant biocompatibility in various embodiments. FIG. 6 shows superior colliculus activation following photoacoustic stimulation of in vivo LE retinae in various embodiments.

[0020] FIG. 7 shows raw data from a tensile test used to calculate Young’s Modulus of various photoacoustic films according to certain embodiments.

[0021] FIG. 8 shows the ultrasound field generated by delivering laser pulses with a 200-pm optical fiber to photoacoustic films according to certain embodiments.

[0022] FIG. 9 shows transmittance (Tl) of a 200-nm-thick candle soot layer of a photoacoustic film according to certain embodiments.

[0023] FIG. 10 shows a characterization of a PDMS-CNT photoacoustic film according to certain embodiments.

[0024] FIG. 11 shows a thermocouple measurement setup according to certain embodiments. FIG. 12 shows temperature variation during in vivo stimulation conditions according to certain embodiments.

[0025] FIG. 13 shows raw voltage recording on the MEA electrode closest to the center of the laser beam with and without PDMS / CS / PDMS film in certain experimental embodiments.

[0026] FIG. 14 shows RGC cell count per stimulation site in certain experimental embodiments.

[0027] FIG. 15 shows an estimation of laser power density exiting the laser injector in certain experimental embodiments.

[0028] FIG. 16 shows an estimation of the laser density on the retina or implant in certain experimental embodiments.

[0029] FIG. 17 shows superior colliculus activation following photoacoustic retinal stimulation of in vivo LE retinae in certain experimental embodiments.

[0030] DETAILED DESCRIPTION

[0031] Retinal degenerative diseases of photoreceptors are a leading cause of blindness with no effective treatment. Retinal prostheses aim to restore sight by stimulating remaining retinal cells. Described herein are photoacoustic retinal stimulation technology including a polydimethylsiloxane and carbon-based flexible film that can convert near-infrared laser pulses into a localized acoustic field with at least 56-μm lateral resolution or better, aiming at high-precision acoustic stimulation of mechanosensitive retinal cells. This photoacoustic stimulation can result in robust and localized modulation of retinal ganglion cell activity inboth wild-type and degenerated ex vivo retinae. In examples discussed below, millimetersized photoacoustic film was implanted in the rat subretinal space and pulsed laser stimulation generated neural modulation in vivo along the visual pathway to the superior colliculus, as measured by functional ultrasound imaging. The biosafety of the film was confirmed by the absence of short-term adverse effects under optical coherence tomography retinal imaging, while local thermal increases were measured below 1 °C. These findings demonstrate the potential of the described photoacoustic stimulation systems and methods for high-acuity visual restoration over a large field of view in blind patients.

[0032] Flexible photoacoustic films are disclosed herein that efficiently generate acoustic waves. The generated photoacoustic waves can be used to stimulate activity in retinal cells ex vivo and in vivo, thereby activating downstream visual pathways in vivo. This RGC stimulation can be demonstrated both ex vivo and in vivo on degenerated retinae as seen in the examples. These results therefore provide evidence that the degenerated retina maintains ultrasound sensitivity, allowing RGC activity modulation and the consecutive stimulation of higher visual areas for visual restoration.

[0033] In various embodiments, an acoustic pressure of 0.05 MPa can be sufficient to elicit responses in the superior colliculus, which is two orders of magnitude lower than the pressure thresholds previously reported for in vivo transducer-based ultrasound retinal stimulation. Pressure thresholds of photoacoustic stimulation correspond to mechanical indexes (MI) below 0.03 (PDMS / CS / PDMS film) and 0.1 (PDMS-CNT film) and spatial peak temporal average intensities (ISPTA) below 0.06 and 0.9 mW.cm-2as shown in Table 1 below.

[0034] Temperature increases can be below 1 °C at the film surface (FIG. 1g, FIG. 12). These metrics align with FDA safety guidelines for ultrasonic ophthalmic devices, making PA implants of the invention a promising option for safe ultrasound stimulation of the retina.

[0035] Restoring meaningful vision requires high spatial resolution over a large area of the retina. To achieve this, each individual stimulation source (pixel) should produce tightly confined fields with sub-20-μm26lateral resolution, and maximize pixel density to 2500 px.mm-2in the macula, which covers a 25 mm2area in humans. Among current electrical retinal prostheses, complex 3D honeycomb photovoltaic devices and optimized laser stimulation sequences show potential to generate 20-pm-wide electric fields with dense arrays of thousands of electrodes, but each rigid implant can only cover at best 9 mm2of the retina. Ultrasound retinal stimulation offers the advantage of non-invasive stimulation over an area greater than the macula, but the pixel density of transducer 2D-arrays is very limited (currently 16 x 169) regardless of the lateral resolution of the stimulation field, which remainsabove 80 μm (in silico data, at 20 MHz9). The photoacoustic film demonstrated herein offers the potential to meet both of these needs due to its spatially continuous nature and the multiplexing capability of photons. In various embodiments, a 50-μm-diameter laser spot can generate a 56-μm lateral ultrasound field, and RGC activity can be locally modulated around the laser spot in both healthy and degenerated retinae. Photoacoustic films herein can also provide high visual acuity using smaller laser spot sizes in various embodiments. Moreover, in some embodiments, pixel numbers can exceed one million by using digital micromirror devices to project a laser pattern onto a continuous film implant.

[0036] In various embodiments, the films described herein can be combined with the use of various sonogenetic techniques to further modulate ultrasound-based neurostimulation.

[0037] Sonogenetics uses ultrasound to non-invasively activate specific neurons through the induction of calcium-permeable mechanosensitive ion channels, including TRPP1 / 2, TRPC1, and Piezo 1. This activation initiates a sequence of intracellular events, commencing with the rapid accumulation of calcium ions. Subsequent activation of calcium-sensitive sodium channels, including TRPM4, leads to cell membrane depolarization and the further opening of voltage-gated calcium channels, notably T-type channels, which can lead to larger neuronal responses.

[0038] Sonogenetic intervention can be enhanced by genetic tools that facilitate the targeted expression of ultrasound- sensing proteins within specific cell types to increase cellular sensitivity to ultrasound, enabling the activation of neurons at lower ultrasound intensities and reducing the overall energy required for stimulation. Sonogenetics applications exist across different categories, incorporating mechanisms based on mechanical force, cavitation effects, and thermal effects. Examples include the large conductance mechanosensitive ion channel (MscE) family, known for its role in converting physical forces at the cellular membrane into electrophysiological activities. Mutants such as MscL-G22s can be selectively activated by low-frequency ultrasound, providing a tool for specific and controlled neuromodulation at lower sound pressures.

[0039] Examples:

[0040] Fabrication and characterization of flexible photoacoustic films The working principle of photoacoustic retinal stimulation is illustrated in FIG. 1. FIG. la shows a 1030-nm pulsed laser delivered to the back of the eye onto the retina, illuminating a subretinally-implanted flexible photoacoustic (PA) film. Laser absorption by the film produces transient heating, which causes thermal expansion and compression of thematerial, thereby generating pulsed ultrasound. The generated ultrasound then stimulates the mechanosensitive retina.

[0041] The PA film illustrated in FIG. lb is composed of candle soot (CS) as the absorber material, sandwiched between two layers of polydimethylsiloxane (PDMS), which serves as the thermal expansion material. The film has a Young's modulus of 2.12 ± 0.10 MPa to minimize the immune response once implanted (FIG. 7). Upon excitation with 4.2-ns laser pulses at 7 μJ per pulse, the PDMS / CS / PDMS film emitted ultrasound pulses with peak-to-peak pressure of 146.2 kPa measured 0.9 mm away (FIG. 1c). At the surface of the film, the conversion efficiency was estimated at 26 kPa.μJ-1based on the distance-dependent pressure profile measured (FIG. 8). PA signals were found to have a central frequency of 42.2 MHz and -6 dB bandwidth ranging from 29.6 to 59.9 MHz. This central frequency has been demonstrated to activate ex vivo salamander retinae with a lower intensity threshold compared to lower acoustic frequencies. These data suggest that the PDMS / CS / PDMS film is a promising photoacoustic converter for retinal stimulation.

[0042] The spatial distribution of the ultrasound field generated by the PDMS / CS / PDMS film was further mapped by PA field microscopy (FIG. Id). A 50-μm optical fiber was attached to the PA film to ensure a 50-μm-diameter illumination area. The axial pressure profile shows that the maximum acoustic signal is generated at the surface of the film upon illumination (Z = 0 pm) and attenuates to 50% of its peak value at Z = 140 μm (FIG. 1d, right). The lateral width (W) of the acoustic field, quantified by the full width at half maximum, measures W = 56 μm at Z = 0 μm and increases with axial depth to W = 124 μm at Z = 100 μm (FIG. 1e). These results confirm that under a confined illumination, the PA film produces a highly localized, 56-μm lateral ultrasound field comparable to the size of illumination, opening up potential for retinal stimulation with sub-100-μm resolution.

[0043] Measured acoustic pressure exhibited a linear relation with the incident laser energy per pulse (FIG. If), which indicates that the output pressure can be precisely modulated by adjusting the input laser energy.

[0044] To ensure that the laser light absorption by the designed PA film is not associated with a substantial and detrimental temperature increase, the temperature was measured at the surface of the PA film. The tested laser conditions were consistent with those employed in the following ex vivo retinal stimulation experiments. Aa maximum temperature rise of 0.52 ± 0.09 °C (FIG. 1g) was observed. The baseline change due to cumulative thermal effects was 0.21 °C after 40 s. This value is an order of magnitude below the temperature increase neededfor thermal neural modulation. Therefore, the film is unlikely to thermally modulate retinal neuronal activity, avoiding any unintended neurostimulation.

[0045] FIG. la shows the working principle of the flexible photoacoustic (PA) film.

[0046] Illumination of the PA film with a nanosecond pulsed laser produces an ultrasound emission. CS: candle soot, PDMS: polydimethylsiloxane, RPE: retinal pigment epithelium, RGC: retinal ganglion cells, BPC: bipolar cells. FIG. lb shows a photograph of an exemplary PDMS / CS / PDMS film with a three-layer design held by a tweezer. FIG. 1c shows a characterization of the PA film in the temporal domain and frequency domain measured 0.9 mm away from the film surface. FIG. 1d shows a mapping of the ultrasound field generated by the PDMS / CS / PDMS film upon illumination through a 50-μm optical fiber. Center. measured distribution of the generated US field. The side lobe on the right of the field is due to the slight tilted angle when the optical fiber was put in contact with the sample film. The top dotted line: interface between water and the film. Top and right: normalized lateral and axial profiles of the PA field, respectively, measured along the vertical and lower horizontal dashed lines in the center panel. The amplitude of the acoustic signal was normalized to the maximum amplitude measured in the field. FIG. 1e shows full width at half maximum of the lateral profile as a function of the axial position Z extracted from FIG. Id. FIG. If shows peak-to-peak pressure of the PA signal as a function of laser energy per pulse measured from a PDMS / CS / PDMS film by a hydrophone. The pressure was normalized to the maximum pressure in all measurements. N = 3 for each data point. The fitted line: linear fitting: y = 0.105x, R2= 0.9945. FIG. 1g shows temperature increase at the surface of the PA film following illumination with a 200-μm laser spot. N = 3 for each data point, mean (thick central line) ± SD (shaded area). The dots indicate laser on. Laser parameters: energy of 10 μJ per pulse, repetition rate of 3 kHz (laser power density P = 0.95 W / mm2), and burst duration 50 ms, delivered every 1 s over 40 s. The baseline change due to cumulative thermal effect was obtained by logistic curve fitting to the data (central fitted line).

[0047] Photoacoustic modulation of the ex vivo retina

[0048] To evaluate retinal responses following photoacoustic stimulation, the activity of retinal ganglion cells (RGCs) was recorded from ex vivo retinae of wild-type Long-Evans (LE) rats on a multi -electrode array (n = 4 rats). The PDMS / CS / PDMS film was placed on the ex vivo retina against the photoreceptor layer. It was photoactivated by a 1030-nm pulsed laser delivered through a 200-pm optical fiber (FIG. 2a), which was successively moved at different positions between stimulations. 4.2-ns laser pulses were applied at a repetition rateof 1.9 kHz (every 520 μs) for a burst duration of db= 10 ms, with a pulse energy of 10 pJ (FIG. 2b. top; power density P = 0.27 W.mm-2), yielding an estimated peak-to-peak ultrasound pressure of 0.12 MPa.

[0049] Photoacoustic stimulation evoked robust RGC responses in healthy LE retinae (FIG.

[0050] 2b, 2c, top two panels). Individual RGCs were considered responsive or modulated if their firing rate significantly increased (excited response) or decreased (inhibited response) relative to baseline (FIG. 2c, 2e). 100 LE RGCs (78%) exhibited an alteration in activity upon PA stimulation (FIG. 2e) out of the 129 spontaneously active RGCs on the electrodes contained in a 300-μm-radius area centered on the laser spot. Responsive RGC activity was mainly increased (92% of responding RGCs, FIG. 2d, left). RGCs with an increased activity had a mean response firing rate of 66 ± 3.7 Hz (FIG. 2f) and had a mean response latency of 51 ± 34.2 ms (FIG. 2g). The response latency was inversely correlated with the firing rate (FIG.

[0051] 2h).

[0052] To investigate the potential of such photoacoustic stimulation for restoring vision, ex vivo retinae from blind P23H rats (n = 4 rats) were then stimulated. Similarly to LE retinae, though with a lower fraction, 89 out of 229 P23H RGCs (39%) exhibited robust responses to photoacoustic stimulation (FIG. 2c-e), predominantly with increased activity (93% of responding RGCs). Compared to LE RGCs, P23H RGC firing rate was significantly lower following stimulation (29 ± 2.88 Hz, FIG. 2f) and response latency was significantly increased (89 ± 65 ms, FIG. 2g). Only 36% of P23H RGCs had response latencies below 45 ms12, which was significantly fewer than LE cells (39%, FIG. 2g). The reduction in shortlatency responses suggests that photoreceptors may partly contribute to this early retinal mechanosensitivity. These results on the P23H rat retina demonstrate the ex vivo efficacy of photoacoustic stimulation in modulating retinal ganglion cells activity in degenerated retinae.

[0053] To further define the origin of the PA responses, glutamatergic blockers (rs)-CPP and CNQX were bath-applied to P23H retinae (FIG. 2i). PA-induced responses were nearly completely abolished and recovered following the washout of the blockers (FIG. 2i). These results suggest that the main mechanosensitive cells are upstream of RGCs, consistent with previous studies, and that glutamate neurotransmission is required to transfer the mechanosensitive signal to the RGCs.

[0054] To exclude the possibility that the healthy LE retina was responding to the pulsed laser light20,21, off-film stimulation was performed by applying laser pulses to the bare LE. Using identical laser conditions, only 3.6 ± 0.9% of RGCs showed a modified activity following off-film stimulation, compared to 77 ± 20% following on-film photoacousticstimulation (FIG. 2j). Furthermore, the PA film was found to absorb more than 99% of the on-film applied laser energy (FIG. 9). These results confirm that the observed RGC responses upon laser activation of the photoacoustic film were not caused by a direct light activation of the healthy retina.

[0055] FIG. 2a shows the ex vivo retina placed on a multi-electrode array (MEA) with the photoacoustic film on the outer layer. FIG. 2b Top shows: a schematic of the laser sequence for photoacoustic stimulation. Laser pulses, with an energy of Ep= 10 p.l per pulse and duration dpulse= 4.2 ns, were delivered at a repetition frequency frep= 1.9 kHz during a single burst of duration db= 10 ms. Each laser pulse is converted by the PA film into an acoustic wave with a duration dUS= 36 ns. FIG. 2b Bottom shows: Example high-pass filtered MEA recording from a single electrode displaying elicited spikes following a photoacoustic stimulation. Shaded area: laser on. Inset: zoom in of action potentials following stimulation. FIG. 2c shows examples of Long Evans (LE) and P23H RGC responses to photoacoustic stimulation. The top two (LE) and bottom two (P23H) lines: mean firing rate. Shaded areas: 99% bootstrapped CI from 1000 samples. Red shaded area: PA stimulation. FIG. 2d shows heatmaps of normalized firing rates for responsive LE RGCs (left, n = 100) and P23H RGCs (right, n = 88). Shaded area: PA stimulation. Dashed black line: 45-ms cutoff for slow and fast latency responses. Cells with excited responses display an increase in firing rate after photoacoustic stimulation, and cells with inhibited responses display a decrease in firing rate. FIG. 2e shows the percentage of cells modulated by photoacoustic stimulation per stimulation site. LE: 74% (4 rats, n = 10 stimulation sites), P23H: 39 %, (4 rats, n = 12 stimulation sites). *** p < 0.001, Mann Whitney U test. FIG. 2f shows firing rates of LE and P23H RGCs during baseline (basal) and following stimulation (stim). Mean firing rates: LE: frbasal= 14 ± 1.0 Hz, frstim= 66 ± 3.7 Hz (n = 82 RGCs within stimulation range, p < 0.001, Wilcoxon signed-rank); P23H: frbasal= 10 ± 1.2 Hz, frstim= 29 ± 2.88 Hz (n = 72 RGCs within stimulation range, p < 0.001, Wilcoxon signed-rank), (g) Latencies of RGC responses for LE (51 ms + 34.2 ms, mean ± standard deviation) and P23H (89 ms ± 65 ms). Dashed black line: 45-ms cutoff for slow and fast latency responses. FIG. 2h shows the firing rate of modulated RGCs as a function of response latency. Firing rate and response latency of excited RGCs were correlated for LE (r = -0.40, p < 0.001, Pearson correlation) and P23H (r = -0.59, p < 0.001, Pearson correlation) RGCs. FIG. 2i shows glutamate blockers (rs)-CPP and CNQX abolished RGC responses to photoacoustic stimulation in P23H retinae. The population firing rate of RGCs (n = 44 cells, 2 retinae) is compared between baseline (basal) and stimulation (stim), before blocker admission (no blocker), following blocker admission ((rs)-CPP+CNQX) and after washout (washout) with Ringer medium. Firing rates per condition, no blockers: frbasal= 12 ± 13 Hz, frstim= 26 ± 29 Hz (p < 0.001); (rs)-CPP+CNQX: frbasal= 6 ± 6 Hz, frs= 6 ± 4 Hz (p =0.749); washout: frbasal= 8 ± 9 Hz, frstim= 21 ± 27 Hz (p = 0.021). Comparison of the population stimulation firing rate frstim after blocker admission and frstim before admission (p < 0.001, Wilcoxon signed-rank), and frstim after washout (p < 0.001, Wilcoxon signed-rank). FIG. 2j shows off-film laser light stimulation. Percentage of LE RGCs modulated by direct laser stimulation on the retina (“off film”) compared to photoacoustic stimulation (“on film”). Laser parameters: Ep= 10 pj per pulse, frep= 3.5 kHz, db= 10 ms. Off film: 3.6% ± 0.9% (n = 56 cells, 2 retinae), on film: 77% ± 20% (n = 59 cells, 3 retinae). Statistics: * p< 0.05, *** p < 0.001. Mann-Whitney U test.

[0056] Dependence of RGC response on laser conditions RGC responses to photoacoustic stimulation was further investigated using different laser repetition rates (10 pJ per pulse, frep1= 1.9 kHz with P1= 0.27 mW.mm-2and frep2= 3.5 kHz with P2= 0.52 mW.mm-2) and burst durations (db= 5 - 30 ms). In LE retinae, the firing rates of cells with excited responses increased with burst durations up to db= 25 ms for frepi, while at frep2 they plateaued up to db= 15 ms and decreased with longer burst durations (FIG.

[0057] 3a and 3b, left). P23H RGCs with excited responses also showed an increase in firing rate with longer burst durations for frep1(FIG. 3a and FIG. 3b, right), while firing rate increasedfor burst durations up to db= 20 ms then decreased with longer burst durations for frep2(FIG.

[0058] 3b, right). LE RGC firing rates were significantly higher than those for P23H RGCs, up to 2.8-fold during db= 25 ms and up to 4.7-fold during db= 20 ms with frep1and frep2, respectively (FIG. 3b). These results suggest that the degenerated retina requires higher thresholds for photoacoustic stimulation, consistent with previous findings concluding on a higher acoustic stimulation threshold in degenerated retinae compared to wild-type retinae. LE RGC response latencies did not increase with burst duration or repetition rate (FIG. 3c), nor did P23H RGC latencies.

[0059] FIG. 3a shows example LE (left) and P23H (right) cells showing increased maximum firing rate (fr) with increased burst duration (recorded at frep1= 1.9 kHz). Lighter shadesindicate longer burst durations (db= 5-30 ms). Vertical lines: laser onset. In FIG. 3b, maximum firing rate as a function of burst duration for LE and P23H RGCs during stimulation with repetition frequencies frep1= 1.9 kHz (dashed line) and frep2= 3.5 kHz (solid line) are shown. Data are plotted as mean + SE. In LE RGCs (left panel), the firing rate was positively correlated with burst duration for frep1(r = 0.91, p = 0.01, Pearson R). In P23HRGCs (right panel), the firing rate was positively correlated during both frep1(r = 0.996, p < 0.001) and frep2(r = 0.811, p = 0.05). With frep1, for db= 5 ms and 20 ms, the maximum firing rate of LE RGCs was 2.8- and 1.3-fold higher, respectively, than for P23H RGCs (p < 0.001 for all conditions, Mann- Whitney U-test). With frep2, for db= 5 ms, the maximum firing rate of LE RGCs is 4.7-fold higher than for P23H RGCs (p < 0.001, Mann-Whitney U-test). FIG.

[0060] 3c shows response latency as a function of burst duration for LE and P23H RGCs showed no significant correlation (LE: p = 0.70 and p = 0.19 for frep1and frep2, respectively; P23H: p = 0.79 and p=0.61, Pearson R). In both FIG. 3b and FIG. 3c, the dashed lines: frep1= 1.9 kHz. (P1= 0.27 W.mm-2). Solid lines: frep2= 3.5 kHz (P2= 0.52 W.mm-2). Dataset for FIGS. 3b and 3c: for LE, n = 244 cells, recorded from 4 retinae. For P23H, n = 104 cells, recorded from 4 retinae.

[0061] Spatial resolution of ex vivo photoacoustic retinal stimulation To investigate the spatial resolution of the photoacoustic retinal stimulation, multiple positions on the film were sequentially targeted by moving the laser-delivering 200-pm fiber to different sites above the photoacoustic film and the responsive retinal cells were mapped (FIG. 4a). The laser repetition rate was set at frep1= 1.9 kHz for LE retinae (n = 11 sites) and frep2= 3.5 kHz for P23H retinae (n = 6 sites), to account for the higher modulation threshold previously described for P23H retinae.

[0062] To assess the spatial distribution of PA-modulated RGCs with excited responses, the maximum RGC firing rate relative to the stimulation site was mapped (FIG. 4b). For both LE and P23H RGCs, the maximum firing rates were located within an area slightly larger than the laser spot (< 400 pm from center) and were negatively correlated with the distance from the laser spot (FIG. 4c). Furthermore, the percentage of responsive RGCs decreased when increasing the distance from the center of the laser spot (FIG. 4d). 73% of LE RGCs and 70% of P23H RGCs were modulated within a 100-pm distance, compared to 14% of LE RGCs and 6% of P23H RGCs at a 400-pm distance. These results indicate that stimulation with the PA film induces a localized response and demonstrate the potential for a high spatial resolution in photoacoustic stimulation.

[0063] FIG. 4a shows different populations of RGCs were modulated by moving the laser fiber at different sites on the film. Example of a P23H retina stimulated at three sites; modulated cells at each stimulation site are grouped by color. The 300-pm-diameter laser spots are marked by dashed circles. FIG. 4b shows RGC firing rate of modulated cells normalized to maximum firing rate, mapped relative to the stimulation site for LE (left, 4retinae, 11 stimulation sites) and P23H (right, 4 retinae, 6 stimulation sites). RGC maximum firing rates were averaged across all recorded cells at the same distance relative to the center of the laser spot (LE: n = 576 and P23H: n = 157 RGCs). Data were smoothed using convolution with a 100-pm gaussian kernel. Dashed circle: 300-pm-diameter laser spot. The shift between the maximum firing rate and the laser spot may be due to uncertainty in the laser spot coordinates, due to the 100-pm pitch of the MEA used for indirect measurement of the exact laser position. FIG. 4c shows the maximum firing rate for individual cells as a function of distance from the laser for LE (left) and P23H (right) RGCs. The response firing rate is negatively correlated with distance (LE: r = -0.310, p < 0.001. P23H: r = -0.268, p < 0.05, Pearson R). Each circle represents an individual cell. FIG. 4d shows the percentage of RGCs modulated as a function of distance from the laser spot.

[0064] In vivo safety of the photoacoustic implants

[0065] To test the feasibility of photoacoustic stimulation and the biosafety of the film in vivo, 1-mm-diameter PA films were chronically implanted in LE and P23H rats. Two designs of PA films were used for in vivo experiments. The 115-pm-thick PDMS / CS / PDMS film, used in the previous ex vivo experiments, was optimized for high photoacoustic conversion efficiency and easy handling. A thinner 40-pm-thick uniformly mixed PDMS-CNT film (characterized in FIG. 10) was designed to approach the 30-pm thickness of the clinically tested PRIMA photovoltaic implant, which has shown no long-term adverse effects aside from minor retinal thinning in patients.

[0066] After implantation, eye fundus imaging confirmed the correct positioning of the implant near the optic nerve and the overall integrity of the retina (FIG. 5a and 5b). No complications, such as retinal tearing after implantation or major inflammation after 7 days post-implantation (dpi), were observed on the OCT images and eye fundus exams. Although the presence of glial cells between the implant and the retina is expected, stimulation efficacy should not be affected, as acoustic attenuation is below 10 dB.mm-1.

[0067] On OCT, the average retinal thickness was 174.0 ± 2.3 pm for LE rats (n = 13) and 72.3 ± 2.3 pm for degenerated P23H rats (n = 8). At the PDMS / CS / PDMS implant position, LE retinal thickness decreased to 123.2 ± 2.9 pm at 15 dpi, 121.5 ± 4.2 pm at 30 dpi, and 107.3 ± 2.0 pm at 90 dpi (FIG. 5c). In LE rats implanted with the PDMS-CNT implant, similar values of 113.0 ± 4.8 pm at 15 dpi and 105.8 pm at 30 dpi, were measured (FIG. 5c). The decrease in the retinal thickness above the implant in LE rats was due to photoreceptor degeneration (FIG. 5a, right), caused by the physical separation of photoreceptors from theretinal pigment epithelium, as previously reported with other prostheses. In implanted P23H rats, retinal thickness above the implant remained stable and comparable to the neighboring area for up to four months for both PDMS / CS / PDMS and PDMS-CNT implants (FIG. 5d). These observations indicate that the photoacoustic implants have no intrinsic short-term toxicity on the wild-type and degenerated retinae.

[0068] FIG 5a shows eye fundus (left) and OCT images (middle and right) of an LE rat retina with a subretinal PDMS / CS / PDMS implant (dotted line) at 7, 15, and 90 days postimplantation (dpi). OCT images were taken along the dotted line shown in the left panel. In the zoomed-in OCT images, GCL: retinal ganglion cell layer, INL: inner nuclear layer, PRL: photoreceptor layer. RPE: retinal pigmented epithelium. The PR layer has degenerated above the implant. Right inset: zoom on the OCT image at 90 dpi. FIG. 5b shows the same elements as FIG. 5a but for a P23H rat. FIG. 5c shows the mean LE retinal thickness above PDMS / CS / PDMS and PDMS-CNT implants over time. Control: mean retinal thickness next to the implant at 15 dpi. Thickness at 15 dpi and later is significantly lower than control thickness (**p < 0.01, Wilcoxon Signed- Rank test). At 15 dpi, the thickness above PDMS / CS / PDMS implants is not statistically different from the thickness above PDMS-CNT implants (p = 0.16, Mann-Whitney U test). Between 15 dpi and 90 dpi, the decrease in thickness above PDMS / CS / PDMS implant (123.2 ± 2.9 pm to 107.3 ± 2.0 pm) is not significant (p = 0.25, Wilcoxon Signed-Rank test). FIG. 5d shows the same as elements as FIG. 5c but for P23H rats. The difference of retinal thickness above both implants is not statistically significant (p = 0.16 at 15 dpi and 0.32 at 30 dpi, Mann- Whitney U-test). Retinal thickness is stable up to 120 dpi for both PDMS-CNT (p = 0.18, one-way ANOVA) and PDMS / CS / PDMS implants (p = 0.51).

[0069] Photoacoustic retinal stimulation in vivo

[0070] The in vivo photoacoustic stimulation of the degenerated retina was then examined using subretinal PDMS / CS / PDMS and PDMS-CNT implants in LE rats. As indicated above, photoreceptors degenerate above the implant in LE rats, creating a localized model of retinal degeneration at the photoacoustic stimulation site (FIGS. 5a and 5c). Activation of the visual pathway was assessed in the contralateral Superior Colliculus (cSC) using functional ultrasound imaging (fUS), which measures relative changes in cerebral blood volume (rCBV) triggered by neuronal excitation (FIG. 6a). To verify the coordinates of the eSC, its activation was measured using control full-field white light stimulation of the implanted eye (P = 0.02 mW.mm2). This generated a large rCBV response area in the cSC (FIGS. 6c and 6i), which was used as a reference area in further analyses (FIG. 6f). A large increase in the amplitude ofthe rCBV was measured, averaged on a 29-by-29-pixel (29 pixels = 302 ± 10 pm) square region of interest (ROI; FIG. 6f), which was centered on the maximum responses in the cSC. The same ROI was subsequently used to compute the averaged rCBV for all of the stimulation conditions (FIG. 6f). To better define the expected size of the activated area for PA stimulation, a 400-μm-diameter spot of 595-nm laser light (P = 0.21 mW.mm-2) was focused onto the healthy retina next to the implant (FIG. 6b, iii). It similarly triggered an increase in rCBV in the cSC (FIG. 6c, ii), with an activated area representing 32 ± 11 % (n = 6) of the area activated by the full-field white light (FIG. 6f). In the same 29-by-29-pixel ROI, the rCBV amplitude and response kinetics were similar to those generated by the fullfield white light stimulation (FIG. 6g, orange).

[0071] Photoacoustic stimulation on the implants was then carried out (FIG. 6b, ii; FIG. 6d, top). The 1030-nm laser delivered eight 125-ms bursts during 2 s, repeated every 15 s, for a total of 15 stimulations per recording (FIG. 6d). Laser power densities were P = 0.29 ± 0.06 W.mm-2(mean ± SD) for PDMS / CS / PDMS implants and P = 0.39 ± 0.12 W.mm-2for PDMS-CNT implants. The estimated peak-to-peak acoustic pressures at the surface of the implants were 0.05 MPa and 0.15 MPa for PDMS / CS / PDMS and PDMS-CNT implants, respectively. Activation was observed in a large cSC area following each photoacoustic stimulation (FIG.

[0072] 6d-e) using both types of implants (FIG. 6c, iii). Photoacoustic stimulation with the PDMS / CS / PDMS and PDMS-CNT implants activated cSC areas measuring 25 ± 4 % (n = 4) and 38 ± 10 % (n = 5) of the full-field-white-light-activated area, respectively (FIG. 6f). Furthermore, in the 29-by-29-pixel ROI, the averaged rCBV amplitudes for the PDMS / CS / PDMS (FIG. 6g) and PDMS-CNT (FIG. 6g) implant stimulations reached similar amplitudes to that generated with the 400-μm-diameter spot of the 565-nm laser light on the healthy retina (FIG. 6g). The size of the activated areas generated by stimulation with the PDMS / CS / PDMS and PDMS-CNT implants, and with the 595-nm-laser spot, were not significantly different (FIG. 6f), By contrast, a direct 1030-nm-light stimulation of the healthy retina (FIG. 6c, iv; P = 0.56 ± 0.21 W.mm-2, 400-pm-diameter spot) did not generate a significant cSC activation (FIG. 6f). The averaged rCBV was negligible in the ROI, and statistically different from all of the other measurements (FIG. 6g). Taken together, these results show that the photoacoustic stimulation of the degenerated retina elicits a robust and local activation of the visual pathway downstream of the retina, with an amplitude and size comparable to a visible light stimulation of the retina.

[0073] FIG. 6a shows a setup for in vivo eye stimulation and fUS recordings. FIG. 6b shows eye fundus images of a 1-mm PA implant (z, yellow dotted circle) and 400-pm-diameter laserspots (ii: pulsed 1030-nm laser on the implant, iii: continuous 595-nm laser on the retina) used for laser and photoacoustic stimulation. FIG. 6c shows functional ultrasound imaging in the coronal plane (left hemisphere, AP, -6.5 mm from the bregma). The correlation map displays the relationship between relative cerebral blood volume (rCBV) and the stimuli. Active pixels reflect regions of activated neurons in the contralateral superior colliculus (eSC) for a single recording (15 stimulations). FIG. 6d Top: Laser sequence for photoacoustic stimulation (repetition rate frep= 6.1 kHz). FIG. 6d Bottom: example normalized CB V trace for pulsed 1030-nm photoacoustic stimulation on a PDMS-CNT implant (measured in 29-by-29-pixel region of interest of the eSC, chosen as the area with peak correlation to full-field-white-light stimulation). FIG. 6e shows average rCBV for a single session with 15 stimulations (same data as for FIG. 6d). Mean rCBV ± 99% CI. The laser sequence starts at 0 s. Shaded area: PA stimulation. FIG. 6f shows the activated area following stimulation, relative to the area activated by full-field white light stimulation (rel. activated area), for different stimuli: full-field white light stimulation (6 rats, n = 6 recordings), 1030-nm laser stimulation on the retina (3 rats, n = 4 recordings), 595-nm laser stimulation on the retina (3 rats, n = 6 recordings), photoacoustic stimulation with PDMS / CS / PDMS implant (3 rats, n = 4 recordings), and photoacoustic stimulation with PDMS-CNT implant (2 rats, n = 5 recordings). The activated area is measured by counting the number of pixels on correlation maps such as FIG. 6c. Circles on the graph mark the ratio for individual recordings. Statistics: p-values vs white light stimulation: * p < 0.05, ** p < 0.01, Wilcoxon Signed-Rank test. PDMS / CS / PDMS vs 595 nm: p = 0.91, PDMS-CNT vs 595 nm: p = 0.792, PDMS / CS / PDMS vs PDMS-CNT: p = 0.56, Mann- Whitney U test. FIG. 6g shows the mean rCBV responses of individual rats following stimulation with a laser (595-nm and 1030-nm on retina) and photoacoustic stimulation, for all rats. Same data as FIG. 6f. Horizontal bars denote significant elevation with respect to the baseline (e.g., no overlap of CI with basal CI). No significant difference in rCBV following photoacoustic stimulation between both implant types was found (e.g., overlapping confidence intervals). Peak rCBV values: white light, 0.26 (at 2.65 s); 595 nm, 0.20 (at 2.69 s); PDMS-CNT, 0.18 (at 3.24 s); PDMS / CS / PDMS, 0.13 (at 3.18 s); 1030-nm laser on retina, 0.02 (at 2.19 s). Shaded areas: 95% bootstrapped CI. Shaded column: PA stimulation.Design of the photoacoustic film

[0074] The laser wavelength and the PA materials were both optimized for safety and performance. Light wavelengths ranging from 500 nm to 1150 nm have maximum transmission in the human eye media. A nanosecond laser with a 1030-nm wavelength was chosen to maximize transmission to the retina, while avoiding triggering responses in photoreceptors, as AMD patients may retain peripheral vision. CS and CNT were selected as the absorber material due to their high photoacoustic conversion efficiency, accessibility, and lower safety concerns compared to lead-containing materials. For the thermal expansion material, PDMS was identified as the best option due to its transparency, high Griineisen parameter, excellent biocompatibility, and stability. The PDMS mixing ratio was adjusted to 5:1 to increase the Young’s modulus, thereby enhancing photoacoustic conversion efficiency.

[0075] Fabrication of the photoacoustic films

[0076] To fabricate the 115-pm-thick PDMS / CS / PDMS film, a uniform layer of candle soot was flame- synthesized and deposited onto a glass slide for 20 s, achieving a thickness of approximately 4 pm. Subsequently, a degassed PDMS mixture of silicone elastomer base and curing agent (Sylgard 184, Dow Corning Corporation, USA) with mix ratios of 5:1 was spin-coated at 500 rpm onto the candle soot layer, and cured at 110 °C for 15 minutes. The resulting cured film was then detached from the glass slide, inverted, and reattached. Another layer of PDMS mixture was spin-coated at 500 rpm and cured at 110 °C for 15 minutes. Both sides of the film were treated with oxygen plasma for 1 min to make the implant surface hydrophilic. The film was cut into smaller areas (5 x 5 cm2) and stored in distilled water before use to avoid reversion to their hydrophobic state. Biopsy punches (Kaimedical) of 1 mm and 1.5 mm were used to cut individual photoacoustic films for ex vivo and in vivo experiments.

[0077] To fabricate the 40-pm-thick PDMS-CNT film, PDMS was prepared at mix ratios of 10:1. Subsequently, a 15%\\ t of CNT (<8 nm OD, 2-5 nm ID, length 0.5-2 pm, VWR, Inc., USA) was mixed with the PDMS, with the addition of IPA to facilitate CNT dissolution. The resulting mixture underwent a 5-minute sonication process, followed by a 30-minute degassing step to eliminate bubbles and IPA. The prepared mixture was then spin-coated onto a glass substrate at 500 rpm for 5 minutes. The coated substrate was cured at 110 °C for 15 minutes.Characterization of the photoacoustic properties of the films

[0078] The photoacoustic properties of the films were characterized with a 40-μm needle hydrophone system (NH0040, Precision Acoustics Inc., UK) or an 85-μm needle hydrophone system (HGL-0085, Onda Corporation, USA). Illumination was provided by a Q-switched diode-pumped laser with a pulse width of 8 ns (RPMC, wavelength 1030 nm, repetition frequency 2.9 kHz, USA), delivered to one side of the film via a multimode fiber with a 200-μm core (FT200UMT, Thorlabs, USA). On the other side of the film, the hydrophone was mounted on a 3D stage and aligned with the area illuminated by the optical fiber. The signals were amplified with a pulser-receiver (Olympus, Model 5073PR, USA) and then recorded via a digital oscilloscope (Rigol, DS4024, USA).

[0079] Mapping the photoacoustic pressure field

[0080] Photoacoustic field microscopy was used to map the generated ultrasound field, as previously reported. Here a 1064-nm pulsed laser (OPOLETTE 355 LD, OPOTEK, pulse duration 5 ns) was used as the pump beam. A continuous wave 1310-nm laser (1310LD-4-0-0, AeroDIODE Corporation) served as the probe. A piece of PDMS / CS / PDMS film was mounted on a 50-μm optical fiber (FG050UGA, Thorlabs), and the 1064-nm laser was delivered to the film sample to generate the photoacoustic signals. A translation stage (ProScan III, Prior) was used to scan the generated ultrasound field. Under a single ns pulse, the PA-induced refractive index change was detected as the imaging contrast.

[0081] Temperature measurements

[0082] A J-type thermocouple with a 200-μm tip was set against the PDMS / CS / PDMS film inside 3% agarose gel, typically used for mimicking tissue. The PDMS / CS / PDMS film wasattached to a 200-μm optical fiber to assure the alignment between the illuminated area on the film and the thermocouple tip. A Q-switched diode-pumped laser with a pulse width of 4.5 ns (RPMC, wavelength 1030 nm, USA) was used to illuminate the film. The temperature on the film was recorded with a 2 kHz sampling rate from 10 recordings. Average data was computed from the 3 recordings that showed the highest temperature rise. Photos of the setup are shown in FIG. 11. No transient temperature events faster than the 2 kHz acquisition frequency are expected to occur (FIG. 12).Animals

[0083] Wild-type Long-Evans male rats aged between 2 and 8 months were obtained from Janvier Laboratories. P23H male and female transgenic rats (9-14 months old) were raised locally. P23H rats serve as a model for autosomal dominant retinitis pigmentosa.

[0084] Ex vivo experiments

[0085] Ex vivo retina preparation and blockers

[0086] The following procedures were carried out under dim red light. Animals were dark adapted for 30 minutes, then anesthetized with CO2 and euthanized by cervical dislocation. The eyes were enucleated and hemisected in carboxygenated (95% O2, 5% CO2) Ringer medium containing (in mM): 125 NaCl, 2.5 KCl, 1 MgCl2, 1.25 NaH2PO4, 20 glucose, 26 NaHCO3, 1 CaCl2and 0.5 L-Glutamine at pH 7.4. The medium was continuously perfused in the recording chamber at a speed of 1.5 mL.min-1and was kept around 37 °C.

[0087] Isolated retinae were placed on a dialysis membrane (Spectra / Por® 650 kD dialysis membrane, Spectrum) coated with poly-L-lysine (0.1%, Sigma), with the photoacoustic film between the dialysis membrane and the retina, and with photoreceptors against the film. The retinae were pressed against a multi-electrode array (MEA) (MEA256 iR-ITO; MultiChannel Systems, Reutlingen, Germany) with a custom 3D-printed piece.

[0088] AMPA / kainate glutamate receptor antagonist 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX, 20 pM, Tocris Bioscience) and NMDA glutamate receptor antagonist (RS)-3-(2-carboxypiperazin-4-yl)-propyl-l-phosphonic acid ((RS)-CPP, 10 pM, Tocris Bioscience) were bath-applied through the perfusion line.

[0089] Ex vivo photoacoustic retinal stimulation

[0090] Photoacoustic stimulations were done with a 1030-nm, 4.2-ns-pulsed laser (One DPSS, Bright Solutions) delivered through a 200-pm-core 0.22 NA multimode SMA / SMA optic fiber (Thorlabs Inc, USA., ref M25L01). A second 200-pm-core was connected to the first fiber using a fixed attenuator (Thorlabs Inc., USA, ref FA26M) to control the power density. The optical fiber was inserted into a custom 3D-printed holder incorporated in a motorized XYZ stage with 0.5-nm precision (Sensapex, uMp-3 micromanipulator). It waslowered above the PA film at a ~90° angle and placed at a fixed (~ 1 mm) distance above the PA film. The illumination spot was measured ~300 μm in diameter on the film using ImageJ and the MEA electrode pitch as reference. A low power 650-nm guiding beam (FIBERCHECK, Laser Components) was used to calibrate the beam position relative to the MEA.Laser pulse repetition rate and the laser burst trains were controlled using a Teensy microcontroller custom written software (C++, Java, Python). In a typical stimulation, the laser delivered 10 pJ pulses with a repetition frequency frepof 1.9 kHz or 3.5 kHz during a single 5-ms to 30-ms burst, repeated at 1 Hz for 40 bursts. PA film integrity was confirmed by the lack of photoelectric effect in the MEA recordings (FIG. 13).

[0091] Analysis of MEA recordings

[0092] MEA raw traces were recorded through the MEA software (MC Rack, Multichannel Systems). Spikes were sorted using SpyKING CIRCUS, manually curated using phy, and attributed to individual cells. Spikes were referenced relative to stimulus onset and grouped across trials in bins using a sliding window (bin width = 20 ms, increments = 5 ms). Cell activity in each bin was estimated using bootstrap resampling (n = 1000 resamples, 99% confidence intervals), and considered significantly increased or decreased if there were no overlapping confidence intervals compared to baseline (200 - 100 ms before stimulus onset).

[0093] RGCs were considered responsive or modulated if their firing rate was significantly increased or decreased compared to baseline for at least 15 ms consecutively, and response latency was defined as the first bin of this series. Noise clusters were filtered from the cell clusters by excluding cells with response latencies below 5 ms. A 300-pm-diameter area was illuminated by the 1030-nm laser (200- m fiber) during photoacoustic stimulation. For quantifying responsive cells and dose responses (FIGS. 2 and 3), only cells within 300-pm of the center of the illuminated area (“stimulation site”) were included. The number of recorded cells per stimulation site is shown in FIG. 14. Cells with a response latency above 250 ms were excluded, as they were likely not a result from direct stimulation.

[0094] To analyze the relation between cell modulation and distance from the stimulation area (FIG. 4a), local averages in firing rate were calculated by assigning the firing rate of each cluster to a bin in a grid with a spacing of 25 pm, and smoothing the resulting averages using a convolution with a gaussian kernel (sigma = 100 pm).

[0095] In vivo experiments

[0096] Successful implantation was defined as good positioning of the 1 -mm-diameter film in the subretinal space, with no complications due to surgery. N = 8 adult (9-10 mo) P23H rats were successfully implanted and used for biocompatibility studies. N = 7 adult Long-Evans rats were successfully implanted at 8 weeks of age and used for photoacoustic stimulation.Surgery procedures for chronic subretinal implantation

[0097] A 1-mm-diameter PA film was surgically placed in the subretinal space in the central region next to the optic nerve, as previously described. Briefly, a small sclerotomy was performed on the dorsal sclera tangential to the cornea. A gel of sodium chondroitin sulfatesodium hyaluronate (Viscoat Alcon) was injected in the sclerotomy to generate a retinal detachment. The implant was then inserted below the detached retina in the subretinal space, targeting a location adjacent to the optic disk.

[0098] Ocular imaging

[0099] Eye fundus imaging (MICRON® IV, Phoenix, USA) and optical coherence tomography (Bioptigen® OCT system, Leica microsystems, Germany) were performed at 7 and 15 days post-implantation (dpi) for all rats to monitor inflammation and confirm correct implantation. Additional imaging was conducted at 30, 60, 90 and 120 dpi for rats that did not undergo prior retinal stimulation.

[0100] Cranial window acute surgery

[0101] Anesthesia was provided with an intraperitoneal injection of 40 mg.kg-1ketamine (Axience, France) and 0.14 mg.kg-1medetomidine (Domitor®, Vétoquinol, France) diluted in sodium chloride. The animal was placed on a stereotaxic frame to perform a left craniotomy. Drops of ocular gel (Lubrithal®, Dechra, France) were applied and the eyes were then covered with a black cloth for dark adaptation. A rectangular piece of bone was removed from Bregma -3 mm to -8 mm.

[0102] Retinal stimulation and brain imaging

[0103] Retinal stimulation was performed 26 - 40 days after implantation surgery for rats with PDMS / CS / PDMS implants and 23 - 29 days after implantation surgery for PDMS-CNT implants, and immediately after the cranial window surgery. Anesthesia was re-administered every 45 min with one-third of the initial dose, up to a maximum of 5 injections. At the end of the experiment, the animals were euthanized using an intracardiac injection (Exagon®, Axience, France).

[0104] For full field light stimulations with a white LED source, the light power on the retina was estimated to be ~0.02 mW.mm-2, based on the power entering the pupil of 1.2 mW. The choice of the stimulation protocol was informed by prior retina studies using functionalultrasound imaging. Each 1.8-s stimulation sequence consisted of 6 evenly spaced 300-ms illuminations (LED on), repeated 15 times.

[0105] For 595-nm, 1030-nm laser stimulation and photoacoustic stimulation, focused laser spots were aimed using a laser injector from the MICRON 810-nm Image-Guided Laser modality combined with a MICRON® III camera (Phoenix, USA). A low power 650-nm guiding beam (FIBERCHECK, Laser Components) was coupled to the injector to safely choose the area to stimulate. The rat’s implanted eye was covered in ocular gel (Lubrithal®, Dechra, France) and in contact with the camera lens. Stimulation sequences for all 3 modalities were identical.

[0106] For 595-nm (continuous) laser stimulation, power density on retina was 26 pW in a 400 ± 26 -pm-diameter laser spot. For photoacoustic stimulation, the same 1030-nm pulsed laser used in the ex vivo experiments was employed. The laser energy exiting the laser injector was EP= 15 pJ per pulse. To aim at the implant for photoacoustic stimulation, the laser focal spot was not placed on the optical axis of the injector lens, which resulted in a loss of power. All the laser diameter at 1 / e2(DL) and laser power density P were estimated from average intensity profiles extracted with Fiji / ImageJ (FIGS. 15 and 16) and are expressed as the mean ± standard deviation.

[0107] For 1030-nm laser stimulation on the retina: DL= 470 ± 70 μm. P = 0.56 ± 0.21 W.mm-2,

[0108] for 1030-nm photoacoustic stimulation with PDMS-CNT implants: DL= 360 ± 60 μm. P = 0.39 ± 0.12 W.mm-2, for 1030-nm photoacoustic stimulation with PDMS / CS / PDMS implants: DL= 410 ± 45 μm. P = 0.29 ± 0.06 W.mm-2.

[0109] The pupil of the eye of interest was dilated with a tropicamide-based eye drop solution (Mydriaticum®, Thea, France) before the first recording. Body temperature was monitored with a rectal probe and maintained using a heating blanket. Respiratory and heart rates were continuously monitored (TCMT, Minerve, France). After local application of lidocaine (4 mg.kg-1, Laocaïne®, MSD, France), the thinned skull was exposed and covered with ultrasound gel.

[0110] The rats were scanned with a system dedicated to small animal ultrasound neuroimaging (Iconeus, Paris, France). Doppler vascular images were obtained using the Ultrafast Compound Doppler Imaging technique. The probe was positioned coronally at Bregma - 6.5 mm in order to measure the cerebral blood volume (CBV) in the contralateral superior colliculus. Each frame was a compound plane wave frame resulting from the coherent summation of backscattered echoes obtained after successive tilted plane wavesemissions. Then, the CBV signal was extracted from the tissue signal by filtering the image stacks with a dedicated spatiotemporal filter using Singular Value Decomposition. Each transcranial Power Doppler image was obtained from 200 compounded frames acquired at 500 Hz frame rate.

[0111] Analysis of in vivo experiments

[0112] Analysis of OCT images

[0113] Mean retinal thickness next to and above the implant were measured with ImageJ on OCT images (diametral slices). The number of rats imaged at 30 dpi and later was lower than the number imaged at 7 dpi and 15 dpi because rats were used for terminal retinal stimulation recordings starting at 23 dpi.

[0114] Analysis of functional ultrasound imaging recordings The correlation map of the CBV variations and the laser sequence for stimulation was computed by the manufacturer’s proprietary IcoStudio software (vl.5.2). A delay of either 2 or 3 s was computed in the calculation of the correlation to account for vascular delay (the chosen value maximizes the correlation). In correlation map displays (FIG. 6c), only significant pixels with a correlation threshold greater than 0.2 are shown. Maps with a correlation threshold of 0.1 are shown in FIG. 17. A 29-by-29-pixel (29 pixels = 302 ± 10 pm) region of interest (ROI) was defined for each animal, centered on the peak intensity of the correlation map for full-field-white-light stimulation. Relative CBV variations (rCBV) were extracted in this ROI for all stimulation types. For each recording (15 laser stimulations), the cerebral blood flow (CBV) was normalized into a relative steady-state value (rCBV) and calculated as the following: rCBV = (CBV(t) - CBVo) / CBVo, where CBV(t) is the power doppler value t seconds after the start of laser sequence, and CBVo is the baseline in the ROI. The baseline was defined as the mean power doppler value 5 seconds before the start of the laser sequence. The data was bootstrapped to calculate confidence intervals.Properties of the photoacoustic film

[0115] Young’s Modulus of PDMS / CS / PDMS photoacoustic film

[0116] The Young’s Modulus E of the PDMS / CS / PDMS film was evaluated using a tensile test (FIG. 7) and the following equation:

[0117] „ Stress FL

[0118] E = - Strain = — AAL,

[0119] where F is the exerted force (N), L is the original length (m), A is the cross-sectional area (m2), and AL is the change in the length (m).

[0120] Following the measurements, the PDMS film’s Young’s modulus was calculated to be 2.12 ± 0.10 MPa. It is two orders of magnitude lower than silicon-based implants (200-300 GPa), but remains several orders of magnitude higher than the retina’s Young's modulus, which ranges between 0.5 kPa1and 25 kPa2. Therefore, the PA film can provide better biocompatibility and induce less immune response than current implants.

[0121] FIG. 7 shows raw data from the tensile test used to calculate Young’s Modulus. A rectangular piece of PDMS / CS / PDMS film was subjected to stretching, and deformation and force were measured.

[0122] Peak pressure and energy conversion efficiency of the PDMS / CS / PDMS film Pressure measurements were performed with a hydrophone (HGL-0085, Onda Corporation, USA), following the illumination of the photoacoustic (PA) film with a 1030-nm laser delivering 8-ns pulses with an energy of 7 pJ per pulse. A peak pressure of 146.2 kPa was measured at 900-pm away from the PDMS / CS / PDMS film, resulting in a photoacoustic conversion efficiency of 21 kPa.μJ-1at 900-μm away from the PDMS / CS / PDMS film.

[0123] FIG. 8 shows the ultrasound field generated by delivering laser pulses with a 200-pm optical fiber. FIG. 8a Top: Acoustic wavefront at the surface of the film (line). The dotted line shows the acoustic propagation direction. FIG. 8a Bottom: acoustic amplitude as a function of distance (pm) from the candle soot (CS) layer. The film surface is 60-pm away from the CS layer. FIG. 8b shows the same as 8a, but the acoustic wavefront is 100-pm away from the surface of the film. FIG. 8c shows the same as 8a, but the acoustic wavefront is 900-pm away from the surface of the film. FIG. 8d Top: Acoustic field. FIG. 8d Bottom: lateral beam width as a function of axial distance from the film surface. Lateral beam width -150 pm at Z = 200pm. As a reference, the average retinal thickness is 174.0 ± 2.3 pm in Long Evans rats and 72.3 ± 2.3 pm in P23H rats (FIGS. 5c and 5d).

[0124] In order to estimate the photoacoustic conversion efficiency at the surface of the film, the PA field generated with a 200-pm fiber was mapped. The PA signal generated at the surface of the film had an amplitude of 1.41. At 100- and 900- m away, the amplitude of the PA signal decayed to 1.39 and 1.12, respectively. Therefore, the photoacoustic conversion efficiency at the surface of the film is estimated to be 26 kPa.μJ-1.

[0125] The energy conversion efficiency ECE is calculated using:

[0126] ECE = EA / EO,

[0127] where E0is the optical energy (energy per pulse, E0= 7 μJ), and EA is the acoustic energy given by:

[0128]

[0129] given a laser spot area A of 200-pm diameter (m2), the density of water p = 998 kg. m-3, the speed of sound in water c = 1480 m.s1, and the peak-to-peak pressure of the acoustic wave (Pa).

[0130] Using the estimated photoacoustic conversion efficiency of 26 kPa.μJ-1at the film’s surface and optical energy, an energy conversion efficiency value of ECE= 5.1 ×10-5% was found when applying a surface laser energy of 22 mJ.cnT2.

[0131] Optical properties of the CS layer

[0132] To investigate the possibility of light leakage following the passage of laser pulses through the PDMS / CS / PDMS film, the transmitted power was measured using a power meter positioned behind the film. The measurements indicated negligible power transmission. Absorption was expected to occur mainly at the CS layer. In order to measure a transmittance value in the detectable range of the spectrophotometer (UV-1900i from Shimadzu), a 200-nm-thick layer of CS was deposited on glass using flame deposition (same protocol as for PDMS / CS / PDMS implant fabrication). The theoretical transmittance of the 4-pm-thick CS layer was then calculated. At 1030 nm, transmission was measured at Ti = 9.0 % for a CS layer thickness of Li = 200 nm (FIG. 9). According to Beer-Lambert law, the expected light transmission T(L) at 1030 nm for a CS layer thickness of L = 4 pm is:

[0133] T(L) = lO^ULi ogh).

[0134] resulting in T(L) = 1.1×10-19%.Adsorption of the CS in the PDMS may affect how compact the layer is, and increase transmittance compared to the theoretically expected values. FIG. 9 shows the transmittance (Ti) of a 200-nm-thick candle soot layer.

[0135] Characterization of PDMS-CNT film for in vivo photoacoustic retinal stimulation A second type of PA film was developed for the in vivo experiments: a 40-pm thick PDMS-CNT film (FIG. 10a). As with the PDMS / CS / PDMS film, photoacoustic signals were generated by delivering a pulsed 1030-nm laser and recorded with a hydrophone set 900-pm away from the film. The PDMS-CNT film emitted ultrasound with a peak pressure of 133 kPa for a laser energy of 10 pJ per pulse, resulting in photoacoustic conversion efficiency of 13.3 kPa.μJ-1at 900-μm-away from the film surface. The PDMS-CNT film provides a central frequency of 10.9 MHz (compared to 42.2 MHz for the PDMS / CS / PDMS film) and -6 dB bandwidth of 5.9 to 15.8 MHz (FIG. 10b).

[0136] Using a PDMS-CNT photoacoustic emitter with a central frequency of 10 MHz, a decay factor of 5 at 900-pm away from the emitter has been reported. Therefore, the photoacoustic conversion efficiency at the surface of the film may be estimated to be 66 kPa.μJ-1, and the energy conversion efficiency is ECE = 8.4 xlO-4% for a surface energy of 32 mJ.cnT2.

[0137] In rats, the in vivo distance between the inner retina and the PA implant can be below 100 pm4. In the inner retina, the peak pressures of the acoustic waves generated with the PDMS-CNT and PDMS / CS / PDMS implants may therefore differ up to a factor 2.5 (66 kPa.pJ1vs 26 kPa.μJ-1). Still, significant and similar superior colliculus activation was found during PA stimulation with both implants (FIG. 6). The lack of measured difference between stimulation with both implants could be due to limitations of the recording method or saturation of superior colliculus responses due to strong stimulation.

[0138] FIG. 10 shows the characterization of the PDMS-CNT photoacoustic film. FIG. 10a shows CNT-embedded PDMS with a thickness of 40 pm. FIG. 10b shows PA performance in the temporal domain (black) and frequency domain (red) of the photoacoustic films corresponding to films shown in FIG. 10a.Mechanical index and spatial peak temporal average intensity

[0139] FDA safety regulations for ultrasound systems and transducers for ophthalmic uses prescribe a mechanical index (MI) below 0.23, and a spatial peak temporal average intensity ( SPTA) below 50 mW.cm-2. These are defined as follows:

[0140]

[0141] MI = NPP / where NPP is the negative peak pulse (acoustic) pressure (MPa) and / is the acoustic frequency (MHz);

[0142] - ISPTA = I * frep, 'N & frep is the laser repetition frequency (Hz), and / is the pulse intensity integral (W.cm-2);

[0143] I = J" p2(t)dt I pc, where p is the density of water (kg.m-3), c is the speed of sound (m.s1), and p is the peak-to-peak pressure of the acoustic wave (Pa).

[0144] First the upper bound values for MI and ISPTA were calculated based on the laser parameters that produce the strongest PA stimulation, i.e. those used for ex vivo photoacoustic stimulation: laser energy E = 10 pJ per pulse, delivered by a 200-μm-diameter fiber, with a repetition frequency of 3.5 kHz. NPP and peak-to-peak pressure values at the surface of the film are estimated from the experimentally measured photoacoustic conversion efficiency, as above.

[0145] Note that to establish these upper bound values, it is assumed that the laser spot diameter on the film during photoacoustic retinal stimulation is identical to that used for establishing the photoacoustic conversion efficiency of the PA film (do — 200 pm). In practice, the laser spot in ex vivo studies can be closer to di = 300 pm in diameter (mean laser energy density P = 0.52 W.mm-2). At constant laser energy per pulse, both NPP and p are expected to decrease as laser spot diameter increases. For a rough estimate, it could be considered that NPP and p values are linearly correlated to laser energy density, and so are to be divided by (di I do)2= 2.25 to approximate experimental values.

[0146] For both PA films, the calculated upper-bound MI and ISPTA values are below FDA thresholds (Tabe 1). The MI and ISPTA values for the PDMS / CS / PDMS film are 3- and 17-fold lower, respectively, than that of the PDMS-CNT film, making the former a safer option. Second, the lower bound values for MI and ISPTA were estimated (Table 1). In vivo, the mean laser spot size was 410 pm for PA stimulation with the PDMS / CS / PDMS film (mean laser energy density P = 0.39 W.mm-2) and 360 pm (P = 0.29 W.mm-2) for stimulation with the PDMS-CNT film. Taking the laser parameters for in vivo stimulation, and assuming that NPP and p values are linearly correlated with laser energy density, lower bound values for MI and ISPTA, were calculated which are 4- to 6-fold lower than the upper bound values.Table 1

[0147] Laser spot f NPP

[0148] PA film MI 1 (mW. cm'2) ISPTA (mW.cm-2) diameter (pm) (MHz) (MPa)

[0149] PDMS / CS 200 {ex vivo) 42.2 0.174 0.06 1.61E-05 0.056 / PDMS 410 {in vivo) 42.2 0.029 0.01 1.88E-06 0.011

[0150] PDMS- 200 {ex vivo) 10.9 0.330 0.10 2.68E-04 0.939 CNT 360 {in vivo) 10.9 0.071 0.02 4.10E-05 0.250

[0151]

[0152] Ultrasound characterization of photoacoustic films. NPP: negative peak pressure, MI: mechanical index, I: ultrasound intensity, ISPTA: spatial peak temporal average intensity. Ex vivo: upper bound values calculated using the laser parameters for ex vivo photoacoustic retinal stimulation and assuming a 200-pm laser spot diameter. In vivo lower bound values calculated using the laser parameters for in vivo photoacoustic retinal stimulation. The listed values comply with FDA thresholds5for MI and ISPTA. MI < 0.23 and ISPTA< 50 mW.cm-2.

[0153] In the retina and, more generally, the eye, plasma formation due to high-energy laser pulses is a concern. In this example, the peak laser power, defined as surface pulse energy divided by pulse duration, is Ppeak - IO’ W.mm'2. This value is three orders of magnitude below the threshold for plasma formation on the cornea, lens, and retina (Ppeak ~ 108W.mm'2for 6-ns laser pulses12).Temperature increase

[0154] Thermocouple measurement setup

[0155] FIG. 11 shows an experimental thermocouple measurement setup. FIG. Ila shows a schematic of the setup. FIG. lib shows a photo of small PDMS / CS / PDMS film placed on the tip of a 200- pm optical fiber to facilitate alignment with the 200-pm thermocouple. FIG. 11c shows a photo of aligned fiber, PA film, and thermocouple sensor in 3% agarose gel.

[0156] Transient temperature events with fast laser repetition rates FDA safety guidelines for ultrasound systems and transducers used in ophthalmic devices set the maximum local temperature increase to 1 °C. Temperature increases with the laser stimulation parameters used ex vivo (FIG. 1g) and in vivo (FIG. 12) have been measured to be below 1°C at the film surface using a thermocouple.

[0157] Transient temperature events in the 0.1-1 ms range could theoretically activate heatsensitive TRPV1 and TRPV2 channels, which have activation thresholds of 43 °C and 52 °C, respectively. However, the thermocouple’s acquisition frequency of 2 kHz is not sufficient to capture those. These transient peaks would have to be induced by the individual laser pulses, which is incompatible with the absorber-to-cell distances in the system. Indeed, the transient component of the temperature rise, induced by laser pulses with a repetition frequency frep, propagates over a distance driven by the thermal diffusion length. = √(D / frep), where D is the thermal diffusivity of the medium. Considering Dwater = 0.14 mitf.s'1(diffusivity for both pure and carbon-loaded PDMS is in the range of 0.1 - 0.2 mm2.s ) and repetition frequencies of 1.9 kHz and 6.1 kHz (the minimum and maximum laser repetition frequencies used in this example), then the thermal diffusion length μ = 8.6 μm and 4.8 μm, respectively. With the PDMS / CS / PDMS film, the minimum distance of cells to the CS layer is 50 pm. Therefore, transient temperature events are not expected to activate heat-sensitive channels.

[0158] FIG. 12 shows temperature variation during in vivo stimulation conditions.

[0159] Temperature variation (AT) at the film surface during 1030-nm laser irradiation. P = 0.34 W.mm'2. Red lines: laser ON. The same stimulus paradigm as for implant stimulation in vivo. Maximum temperature increase (AT) of 0.64 °C.Photoelectric effect of laser on MEA

[0160] In this study, a multi-electrode array (MEA) was used to measure retinal ganglion cell activity. Whether the laser resulted in the electrical signals due to the photoelectric effect was investigated. First, the effect of the laser transmitted through the PA film was examined. The configuration of the measurement was the same as for ex vivo experiments, except with no retina (FIG. 13a, top). Following the photoactivation of the PA film, a low frequency electrical signal was measured by the MEA at the onset of photoacoustic stimulation (Fig. 13a, bottom). This signal is too slow to be attributed to photoelectric effect and is possibly an indirect detection of induced temperature variations. When the 1030-nm pulsed laser directly illuminates the MEA (FIGS 13b- 13d), it generates a strong photoelectric signal, with individual voltage peaks for each laser pulse (FIG 13d). At comparable energy density, the photoelectric signal is much stronger than the slow wave signal generated by the PA film. The lack of photoelectric effect when the MEA is covered by the PA film is coherent with the expected low light transmission of the film.

[0161] FIG. 13 shows raw voltage recording on the MEA electrode closest to the center of the laser beam with and without PDMS / CS / PDMS film. FIG. 13a Top: setup with photoacoustic (PA) film between laser and multielectrode array (MEA). FIG 13a Bottom: voltage recording from the MEA electrode on which the laser is centered (greatest signal amplitude on MEA). horizontal line: laser on. Laser parameters: energy per pulse E = 10 μJ, repetition rate frep= 2.94 kHz, burst duration db = 30 ms,. FIG. 13b Top: setup with the laser directly illuminating the MEA. Bottom: same as FIG. 13a. Same laser parameters as FIG. 13a. Electrode saturation occurs for voltage signals greater than 3.4 μV. FIG. 13c has the same setup as FIG. 13b. Laser parameters: E = 2 μJ per pulse, db - 15 ms. E divided by 5 and db divided by 2 to avoid electrode saturation and obtain voltage signals of comparable amplitude to FIG. 13a. FIG. 13d shows a magnified version of the X axis of FIG. 13c. In 5 ms, the laser generates 14-15 optical pulses at 2.94 kHz. 14 voltage peaks are observed in the MEA recording.Ex vivo RGC count per stimulation site

[0162] FIG. 14 shows RGC cell count per stimulation site. Number of RGCs with baseline activity in a 600-pm-diameter area (the “stimulation area”) centered on the 300-pm-diameter area illuminated by the laser during stimulation. Each dot represents an individual stimulation area. LE: 10 stimulation areas from 4 retinas, n = 11 ± 5 cells per stimulation area (mean ± S. D.). P23H: 12 stimulation areas from 5 retinas, n = 7 ± 5 cells per stimulation area.

[0163] In vivo - laser irradiance calculation

[0164] Characterization of the laser beam exiting the laser injector The laser injector from the MICRON 810-nm Image-Guided Laser modality is designed to project a laser spot of similar size to the diameter of the optical fiber used for delivery at the focal point of the injector lens. At the focal plane (7 mm from the injector lens, equivalent to the average diameter of a rat’s eye), the measured laser beam radius, delivered with a 200-μm-diameter fiber, was wi = 162 pm.

[0165] A continuous laser with a repetition frequency frep= 6.1 kHz was applied to the laser injector. The power exiting the laser injector Po was measured with a power meter. The resulting energy per pulse Epo was calculated using Epo = Po / frep= 15 μJ / pulse.

[0166] When the laser beam exits the injector along the optical axis (“on-axis” laser beam), e.g. through the center of the injector’s lens, the power is concentrated at the laser focal spot. When the laser beam exits off the axis (“off-axis” laser beam, with r the distance to the optical axis), the laser focal spot holds only a fraction of the total laser power. The ratio Rp(r) = I(r) / 1(0), representing the laser intensity at the focal spot for on- and off-axis configurations, was measured from images of the laser spot exiting the injector (FIG. 15a) using Image J by extracting the integral pixel value of the laser profile (cutoff at 1 / e2of maximum). The resulting off-axis energy per pulse can then be calculated with: Ep(r) = Rp(r) * Po.

[0167] For a given laser spot, the beam radius at 1 / e2is calculated using Gaussian interpolation of the laser intensity profile (FIG. 15b). In the on-axis configuration (r = 0), laser radius is w1 = 162 μm. Resulting power density is P1 = frep* Ep1 / (πw12), with Ep1= 15 μJ per pulse. Pi = 1.11 W.mm’2. In the off-axis configuration (r = 292 pm), w2 = 120 μm. The measured intensity ratio is RP(r) = 0.28 (FIG. 15c). As a result, for r - 292 pm, energy per pulse at focal point is Ep2= RP(r) * Ep1= 4.2 μJ and power density is P2 = Ep2 / frep= 0.57 W.mm-2. In this example, the power density is approximately halved in the off-center position.FIG. 15 shows an estimation of laser power density exiting the laser injector. FIG. 15a shows images of the 1030-nm laser beam exiting the injector. Left: on- axis beam, r = 0. Right: off-axis beam, r = 292 pm. Scale bar: 500 pm. FIG. 15b shows laser intensity profiles at the focal spot of an on-axis and off-axis 1030-nm laser beam. Experimental profile (continuous line) and Gaussian interpolations (dotted lines). FIG. 15c shows integrals I1 and I2 (cutoff at 1 / e2of maximum) of the experimental laser profiles, respectively on-axis and off-axis. In this specific example, Rp= 0.28.

[0168] Characterization of the laser beam on the retina and on the implant In the previous section, the laser power density was estimated at the focal point when the laser beam exits the laser injector off-axis. In practice, the rat retina is not in the focal plane of the injector lens during in vivo stimulations. When the PA implant or the rat retina is closer to the injector lens than the focal planes, the diameter D of the laser beam on the implant or retina will be larger than the laser spot diameter at the focal plane, whether the beam is off- or on-axis (FIG. 16a). This will further reduce the laser power density.

[0169] When using a laser at repetition rate frep= 6.1 kHz and an energy per pulse Epi= 15 pj / pulse, the resulting laser power density at the focal plane (in an on-axis configuration) is P1 = frep* Ep1 / (π*w12) = 1 W.mm-2, with wi = 162 pm as described in the previous section.

[0170] The assumptions made to calculate laser power density (W.mm-2) during stimulation on the retina or on the implant were the following:

[0171] - for Ep(r)

[0172] no light absorption in the eye;

[0173] Ep= 15 pJ / per pulse when. when the laser beam is on-axis;

[0174] no reflection of laser light on the implant. All the injected light is considered absorbed by the implant and converted into acoustic or thermal energy;

[0175] The size of the imaged laser spot (on camera) is equal to the size of the spot on the PA implant.

[0176] In addition to optical aberrations due to the injector lens, there may also be spherical aberrations due to the biological lens.

[0177] A laser beam radius of weye= 195 pm was measured from eye fundus images with the laser on (FIG. 6b(ii), FIG. 16b). This suggests that the retina is in front of the injector lens’ focal point. When using the same laser repetition rate (frep= 6.1 kHz) and energy per pulse (Epi= 15 pJ per pulse) as the previous paragraph, the resulting power density is Peye= 0.69 W.mm-2.Control experiments with a 595-nm and a 1030-nm laser used to directly stimulate the retina were all done in an on-axis configuration. Experiments on implants had to be performed in off-axis configurations to align the laser on the 1 -mm-diameter implant.

[0178] FIG. 16 shows estimation of the laser density on the retina or implant. FIG. 16a shows schematics of an injected laser beam where the injector lens focal plane is behind the implant. f= 7 mm, d <f. Laser spot diameter on implant D = 2weye. FIG. 16b shows laser intensity profiles (solid line) obtained from eye fundus, and gaussian fits (dotted line) of on-axis laser beam at focal plane and on a rat retina. Laser diameter D = 2w, with laser beam radius a 1 / e2.

[0179] Neglect of laser absorption in the eye

[0180] Before reaching the implant, the 1030-nm laser light goes through the aqueous humor, the lens and the vitreous. In humans, more than 90% of 1030-nm light is transmitted through the aqueous humor, and more than 80 % is transmitted through the vitreous. Given that the thicknesses of the aqueous humor and the vitreous arc much smaller in rats than in humans, the total light absorption in a rat’s eye is therefore low enough (~ 20%) that it can be assumed that no light was absorbed by the rat eyeball for experimental purposes.

[0181] Superior colliculus activation following photoacoustic retinal stimulation of in vivo LE retinae.

[0182] Pixels with a correlation threshold increase greater than 0.2 between the different stimulations and the relative cerebral blood volume variations (rCBV) are displayed on the correlation map in FIG. 6c. For lower correlation values, the increase of rCBV compared to the baseline is not significant. FIG. 17 shows additional correlation maps (bottom line) with a 0.1 minimum pixel correlation threshold. With this lowered threshold, pixels appear in the contralateral Superior Colliculus (eSC) of the brain for 1030-nm laser stimulation of the retina. This suggests that for higher laser energy levels, the eSC may significantly respond to infrared pulsed stimulation.

[0183] FIG. 17 shows superior colliculus activation following photoacoustic retinal stimulation of in vivo LE retinae. Brain slice of one rat (coronal plane, left hemisphere, AP Bregma -6.5 mm) with correlation maps displaying eSC activation for a single recording (15 stimulations). Top row: correlation threshold between rCBV and the laser sequence is 0.2 (same as FIG. 6c). Bottom row: correlation threshold is 0.1.

Claims

1. CLAIMSWhat is claimed is:

1. A flexible photoacoustic ultrasound transducer film comprising:a matrix formed of polydimethylsiloxane (PDMS) and a carbon-based optical absorber that is disposed as a discrete layer sandwiched between opposed portions of the PDMS matrix or dispersed throughout the PDMS matrix;wherein the film has a total thickness of about 200 pm or less and is configured, when illuminated by pulsed laser light having individual pulse energies between about 1 pJ and about 20 pJ, to emit ultrasonic pulses with a lateral spread of less than about 100 μm while producing a temperature rise at a surface of the film of less than about 1 °C.

2. The film of claim 1, having a photoacoustic conversion efficiency at the film surface of at least about 26 kPa / pJ when illuminated by the pulsed laser light.

3. The film of claim 1, having a Young’s modulus of about 2.12 MPa.

4. The film of claim 1, wherein the carbon-based optical absorber is a candle-soot layer having a thickness of about 3 pm located between first and second PDMS layers to provide a PDMS / candle-soot / PDMS sandwich structure.

5. The film of claim 4, wherein the candle-soot layer has a transmittance at 1030 nm of less than about 1.1×10-19%.

6. The film of claim 4, wherein each PDMS layer is formed from a pre-polymer / curing-agent mixture having a weight ratio of about 5:1 and is spin-coated at about 500 rpm and cured at a temperature of about 110 °C.

7. The film of claim 1, further comprising hydrophilic surface modification produced by exposing at least one surface of the film to oxygen plasma for approximately one minute.

8. The film of claim 1, wherein the carbon-based optical absorber is a plurality of carbon nanotubes (CNTs) dispersed in the PDMS matrix at a concentration of about 15 % wt.

9. The film of claim 8, wherein the CNT-loaded PDMS matrix has a thickness of about 40 μm.

10. The film of claim 1, wherein the pulsed laser light has a wavelength of approximately 1030 nm and a pulse duration of about 4 to 8 ns.

11. The film of claim 1, wherein the emitted ultrasonic waves exhibit a central frequency in a range from about 30 MHz to about 60 MHz and a -6 dB bandwidth ranging from about 29.6 MHz to about 59.9 MHz.

12. The film of claim 1, configured to emit a peak acoustic pressure at the film surface between about 0.05 MPa and about 0.15 MPa when illuminated by the pulsed laser light.

13. A skin-conformal wearable ultrasound patch comprising,the photoacoustic ultrasound transducer film of claim 1 and an optical coupling arranged to receive external pulsed laser radiation and direct the radiation onto the carbon-based optical absorber.

14. The patch of claim 13, wherein the optical coupling is a flexible optical fiber with an approximate diameter of 200 pm that terminates adjacent to the carbon-based optical absorber.

15. A photoacoustic retinal implant comprising, the film of claim 1 and configured for placement in a subretinal space to generate ultrasonic waves that stimulate adjacent retinal neurons when illuminated by pulsed laser light delivered through ocular media.

16. A method for neuromodulating retinal ganglion cells, the method comprising, implanting, in a subretinal location of an eye, a flexible photoacoustic film comprising an elastomeric PDMS matrix and a carbon-based optical absorber selected from the group consisting of candle soot and carbon nanotubes, the film having a thickness not exceeding about 200 pm;directing pulsed laser light toward a selected region of the implanted film to generate, at the selected region, localized ultrasonic fields sufficient to activate retinal ganglion cells while limiting the temperature increase at a film-tissue interface to less than about 1 °C.

17. The method of claim 16, further operable to maintain a mechanical index below about 0.1.

18. The method of claim 16, further operable to maintain a spatial peak temporal average intensity below about 1 mW / cm2.

19. The method of claim 16, wherein the pulsed laser light has a wavelength of about 1030 nm, individual pulse energies between about 1 μJ and about 20 μJ, a repetition rate between about 1 kHz and about 6 kHz, a pulse duration of about 4 to 8 ns, and a focal spot having a diameter of about 50 pm.

20. The method of claim 16, wherein the pulsed laser light is delivered in bursts having a duration between about 5 ms and about 30 ms to control firing rates of the retinal ganglion cells.

21. The method of claim 16, wherein generation of the ultrasonic fields produces peak acoustic pressures at the film surface of between about 0.05 MPa and about 0.15 MPa.

22. A method for fabricating a PDMS / candle-soot / PDMS photoacoustic sandwich film, the method comprising,depositing, onto a substrate, a uniform layer of candle soot at a thickness of about 3 pm spin-coating a first layer of uncured PDMS onto the uniform layer of candle soot at about 500 rpm;curing the spin-coated first layer at approximately 110 °C;detaching from, inverting, and reattaching the cured, PDMS-coated candle soot layer to the substrate;spin-coating a second layer of uncured PDMS over the inverted candle-soot layer at about 500 rpm; andcuring the second PDMS layer at approximately 110 °C to obtain a flexible photoacoustic sandwich film.

23. The method of claim 22, wherein the flexible photoacoustic sandwich film has a total thickness of about 115 pm.

24. The method of claim 22, further comprising treating the flexible photoacoustic sandwich film with oxygen plasma for about one minute to render both surfaces of the film hydrophilic.

25. A method for producing a carbon-nanotube-loaded PDMS photoacoustic film suitable for subretinal implantation, the method comprising,dispersing carbon nanotubes into PDMS pre-polymer to a concentration of about 15 wt % by ultrasonic sonication for approximately five minutes;degassing the dispersion for about 30 minutes;spin-coating the degassed dispersion onto a substrate at about 500 rpm;curing the spin-coated, degassed, dispersion at approximately 110 °C to obtain a flexible photoacoustic film.

26. The method of claim 25, wherein the flexible photoacoustic film has a total thickness of about 40 μm.

Images