Retinal implant
A retinal implant, also known as a retinal prosthesis or bionic eye, is an implantable medical device designed to partially restore vision in patients with severe outer retinal degenerative diseases, such as retinitis pigmentosa and age-related macular degeneration, by electrically stimulating surviving inner retinal neurons to bypass damaged photoreceptors.[1] These devices convert light patterns captured by an external camera or directly via photovoltaic elements into electrical signals delivered through microelectrode arrays positioned on, under, or near the retina, enabling perceptions of light, shapes, motion, and basic objects.[1] Retinal implants represent a multidisciplinary advancement in bioengineering and ophthalmology, with the first human trials in the early 1990s and commercially approved systems emerging in the 2010s.[2] Key types of retinal implants include epiretinal designs, which attach electrodes to the inner retinal surface to stimulate ganglion cells (e.g., the FDA-approved Argus II system); subretinal implants, placed beneath the retina to target bipolar cells (e.g., the CE-marked Alpha AMS and the investigational PRIMA system[3]); and suprachoroidal implants, positioned in the suprachoroidal space for broader stimulation (e.g., the Bionic Vision Australia 44/49-channel device).[1] As of 2025, clinical outcomes demonstrate sustained long-term efficacy in restoring functional vision, with patients achieving improvements in tasks like square localization (up to 100%), motion direction detection (50-100%), and grating visual acuity (up to 3.33 cycles/degree), though achieved visual acuities remain limited to around 20/460-20/565, classifying users as legally blind.[4] Safety profiles are generally favorable, with most adverse events—such as inflammation, erosion, or hypotony—being treatable and occurring primarily within the first year post-implantation, leading to low explantation rates (1-3% across studies).[4] Beyond technical performance, retinal implants have shown notable enhancements in quality of life, including better mobility, orientation, and performance in daily activities, with 65-80% of users reporting positive or mildly positive impacts sustained over 3-5 years.[4] Ongoing research focuses on improving pixel density, biocompatibility with advanced materials like graphene, wireless power delivery, and integration with artificial intelligence for signal processing to enhance resolution and usability.[1] Despite these advances, challenges persist, including surgical risks, limited field of view, the need for extensive rehabilitation, and competition from emerging therapies like optogenetics and gene editing, underscoring the implants' role as a bridge technology for profound vision loss.[4]Fundamentals
Definition and purpose
A retinal implant, also known as a retinal prosthesis or bionic eye, is an implantable electronic device designed to restore partial vision by electrically stimulating surviving inner retinal neurons, such as bipolar and ganglion cells, thereby bypassing damaged photoreceptors in the outer retina.[2][5] These devices generate visual perceptions known as phosphenes, which are spots of light elicited by neural activation, allowing users to perceive patterns rather than natural imagery.[6] The primary purpose of retinal implants is to provide functional vision to individuals with profound vision loss due to outer retinal degenerations, enabling basic tasks such as light perception, object recognition, navigation, and orientation in daily activities, though they do not achieve full restoration of natural sight.[2][6] They target patients who retain viable inner retinal cells but have lost photoreceptor function, offering a means to improve independence and quality of life where no other treatments exist.[5] In basic operation, many designs use an external camera mounted on glasses to capture visual scenes, which are processed by external or onboard electronics into electrical signals that are transmitted wirelessly to an array of implanted electrodes; these electrodes then deliver targeted pulses to mimic the firing patterns of photoreceptors, conveying information to the brain via the optic nerve. Other designs, such as certain subretinal photovoltaic implants, convert incident light directly into electrical stimulation without an external camera.[6][2]Targeted conditions
Retinal implants primarily target outer retinal degenerative diseases where photoreceptors are lost but the inner retinal layers, including bipolar and ganglion cells, remain relatively intact to facilitate signal transmission to the optic nerve and brain.[7] The most common conditions include retinitis pigmentosa (RP), advanced age-related macular degeneration (AMD), and other inherited dystrophies such as Usher syndrome, which features RP-like retinal degeneration alongside hearing loss.[7][8] For instance, the Argus II epiretinal prosthesis is indicated for severe to profound RP, while subretinal systems like PRIMA address geographic atrophy in late-stage AMD.[9][10] Eligibility criteria for implantation emphasize profound vision loss with preserved inner retinal function, typically requiring bare or no light perception (worse than 2.9 logMAR) in both eyes, a history of prior useful vision, and no comorbidities affecting the optic nerve or central visual pathways.[8] Candidates are generally adults aged 25 years or older, pseudophakic or aphakic (or willing to undergo lens removal), and psychologically stable for intensive rehabilitation and follow-up.[9][8] Conditions like optic atrophy, central nervous system damage, or inability to commit to post-operative care exclude patients, as implants depend on viable inner retinal neurons for efficacy.[7][8] These devices are unsuitable for diseases impacting inner retinal layers or post-retinal structures, such as glaucoma or stroke-related visual loss, because stimulation cannot effectively propagate signals without intact bipolar and ganglion cells.[7] In RP and AMD, degeneration primarily affects the outer retina, preserving the inner layers necessary for prostheses to bypass damaged photoreceptors and elicit phosphene-based perception.[7] Patient demographics predominantly involve individuals with inherited RP, which has a global prevalence of approximately 1 in 4,000, or late-stage AMD, affecting around 200 million people worldwide as of 2025.[11][12] Usher syndrome patients, representing a subset of RP cases, also qualify if inner retinal viability is confirmed, though their dual sensory deficits require additional rehabilitation considerations.[13][7]Historical development
Early research
The foundational research on retinal implants began in the late 1980s and continued through the early 2000s, focusing on demonstrating the feasibility of electrical stimulation to elicit visual responses in degenerated retinas using animal models. Researchers such as Eberhart Zrenner in Germany and Mark S. Humayun in the United States conducted pioneering experiments showing that direct electrical stimulation of the retina could activate surviving inner retinal cells. In animal models including cats and rabbits, these studies elicited cortical evoked potentials, analogous to phosphene generation in humans, confirming that stimulation could bypass damaged photoreceptors and propagate signals to the visual cortex.[14][15][16] During the 1990s, key breakthroughs emerged from in vitro and ex vivo studies that validated the safety and efficacy of retinal stimulation without causing tissue damage. Experiments on isolated rabbit and chicken retinas demonstrated selective activation of retinal ganglion cells using bipolar electrodes, with thresholds low enough to suggest practical implant viability. Concurrently, the development of microelectrode arrays leveraged silicon-based microelectromechanical systems (MEMS) technology, enabling precise, multi-site stimulation on flexible substrates suitable for retinal curvature. These arrays, fabricated with photolithographic techniques, allowed for higher electrode densities and reduced impedance, paving the way for more focal visual percepts.[17][18] Influential collaborative efforts accelerated progress, notably the U.S. Department of Energy's Artificial Sight Basic Research Program, initiated in the late 1990s, which funded multi-institutional teams across national laboratories and universities to integrate engineering, neuroscience, and ophthalmology expertise. This program supported advancements in device prototyping and biocompatibility testing. Additionally, early intellectual property, including contributions from Mark S. Humayun in the mid-1990s on epiretinal visual prostheses, outlined methods for chronic implantation and patterned stimulation to mimic visual input.[19][20] A major challenge addressed in this era was determining safe stimulation parameters to prevent electrochemical damage or thermal effects on delicate retinal tissue. Biophysical modeling, informed by finite element simulations and empirical thresholds from animal studies, established charge density limits of approximately 100-200 µC/cm² for platinum electrodes, balancing efficacy with safety margins derived from neural response data. These guidelines, rooted in general principles of neural prosthetics, ensured that pulse durations and amplitudes avoided faradaic reactions while achieving reliable neural activation.[21][22]Major advancements and approvals
The first human implantation of a retinal prosthesis occurred in 2002 with the Argus I device developed by Second Sight Medical Products, which featured a 4x4 array of 16 platinum electrodes and was tested in six patients with retinitis pigmentosa (RP), enabling them to perceive basic light and motion.[23][24] Building on this foundation, the Argus II epiretinal prosthesis, also from Second Sight, advanced the technology with a 60-electrode array and received CE marking in Europe in 2011, followed by approval from the U.S. Food and Drug Administration (FDA) in 2013 as the first retinal prosthesis granted humanitarian device exemption for treating severe to profound RP in adults aged 25 and older.[25][26] In parallel, subretinal approaches progressed with the Alpha IMS implant from Retina Implant AG, which underwent its first multicenter clinical trial starting in 2010 and involving implants in patients with inherited retinal degenerations across European sites from 2012 to 2016, culminating in CE marking in 2013 for restoring low-vision functionality in blind individuals. However, Retina Implant AG filed for insolvency in 2016, leading to the discontinuation of the Alpha IMS and subsequent Alpha AMS devices, affecting long-term patient support.[27][28] Another key milestone was the CE marking in 2016 for the IRIS II epiretinal system by Pixium Vision, designed specifically for patients with outer retinal degeneration due to RP, marking it as the third bionic retinal device approved in Europe at the time. Pixium Vision later prioritized subretinal technologies like PRIMA, with IRIS II seeing limited adoption.[29][28] However, commercial challenges emerged, as Second Sight faced severe financial difficulties leading to near-bankruptcy in 2020, which resulted in the discontinuation of manufacturing and support for the Argus II by 2019, leaving implanted patients without ongoing technical assistance or parts.[30] Global efforts also shifted toward addressing age-related macular degeneration (AMD), exemplified by Pixium Vision's PRIMA subretinal photovoltaic implant, with the first human implantations occurring in 2018 as part of a feasibility study in patients with advanced dry AMD.[31]Classification of implants
Epiretinal implants
Epiretinal implants are positioned on the inner surface of the retina, adhering to the vitreoretinal interface to directly stimulate retinal ganglion cells (RGCs). This placement allows the electrodes to bypass damaged photoreceptors and bipolar cells, delivering electrical pulses to the RGCs that transmit signals via the optic nerve to the brain. To ensure stable contact and prevent migration, the electrode array is secured using retinal tacks, which are small titanium or silicone devices inserted through the sclera to anchor the implant to the retinal surface.[32][33][34] A prominent example of epiretinal implants is the Argus II system, developed by Second Sight Medical Products, which features an array of 60 platinum-iridium electrodes, each approximately 200 µm in diameter, positioned over the macula. The system relies on an external camera mounted on glasses to capture visual information, which is processed and wirelessly transmitted to the implant for stimulation. The Argus II, approved by the FDA in 2013 under a humanitarian device exemption, represented the first such device to achieve broad implantation, with over 350 units placed worldwide by 2019 to provide partial vision restoration in end-stage RP cases; production and new implantations were discontinued in 2019.[35][36][37] These implants offer surgical advantages, including a vitrectomy-based procedure that avoids the need for subretinal dissection, reducing risks such as choroidal hemorrhage or retinal detachment compared to deeper placements. The epiretinal location provides stable positioning on the inner retina due to the tack fixation, facilitating consistent stimulation, and allows for relatively easier explantation if complications arise, as the array can be removed without disrupting underlying retinal layers.[34][38][39] Epiretinal implants gained widespread clinical use initially for patients with retinitis pigmentosa (RP), a condition causing progressive photoreceptor loss.[35][36]Subretinal implants
Subretinal implants are positioned in the subretinal space, between the retinal pigment epithelium and the neural retina, to directly stimulate the bipolar cells in the inner nuclear layer. This placement allows the electrodes to interface closely with the surviving inner retinal circuitry, enabling electrical signals to propagate through the preserved retinal network to activate retinal ganglion cells. By targeting bipolar cells upstream in the visual pathway, these implants leverage the retina's intrinsic neural processing for more physiologically accurate signal transmission compared to approaches that bypass this network.[40][41][42] Representative examples include the Alpha IMS and Alpha AMS devices developed by Retina Implant AG, which feature a hermetically sealed, biocompatible microchip measuring approximately 3 mm × 3 mm with 1,500 square electrodes, each 50 × 50 μm in size, integrated with microphotodiodes and amplifiers for direct light-to-stimulation conversion. Although Retina Implant AG ceased operations in 2019 due to funding challenges and insufficient clinical breakthroughs, its technology has influenced subsequent photovoltaic designs. Another notable device is the PRIMA implant from Pixium Vision (now Science Corporation), a wireless photovoltaic array spanning 2 mm × 2 mm with 378 pixels designed specifically for atrophic age-related macular degeneration, where projected infrared light from external glasses activates the pixels to stimulate underlying bipolar cells without the need for percutaneous connections. In a 2025 multicenter clinical trial with 38 participants, 81% showed clinically meaningful visual acuity improvement (≥0.2 logMAR) at 12 months post-implantation. These designs emphasize compact, high-density electrode arrays to mimic photoreceptor function in the subretinal space.[43][44][45][46][47] Advantages of subretinal implants include their ability to utilize the retina's natural amplification and contrast enhancement mechanisms through preserved bipolar and horizontal cell interactions, potentially yielding more structured phosphene patterns. The close proximity to the photoreceptor layer also facilitates higher spatial resolution by minimizing signal diffusion and allowing for lower stimulation thresholds, which can enhance perceptual quality in patients with intact inner retinal layers.[48][39][49] The development of subretinal implants began with early human trials by Retina Implant AG in 2005, focusing on patients with retinitis pigmentosa and other outer retinal degenerations like choroideremia, where the first implants restored basic light perception and object recognition. Subsequent trials in 2010 expanded to the Alpha IMS device, demonstrating improved visual function such as letter recognition in blinded participants.[50][51][47]Suprachoroidal and emerging types
Suprachoroidal retinal implants position electrode arrays in the suprachoroidal space, located between the sclera and choroid, to stimulate retinal ganglion cell (RGC) axons indirectly from the outer layers of the eye.[52] This approach allows electrical pulses to propagate toward the inner retina without direct contact, potentially preserving residual photoreceptor function in cases of advanced degeneration.[53] A prominent example is the suprachoroidal prosthesis developed by Bionic Vision Technologies, which evolved from an initial 24-channel prototype tested in a first-in-human trial starting in 2014 to a second-generation 44-channel device implanted in clinical trials from 2018 onward. In Australian phase II trials as of 2025, the device demonstrated stable implantation over 5 years with no serious adverse events and functional gains such as improved object recognition and mobility in participants with retinitis pigmentosa. The 44-channel version features a hermetically sealed receiver unit and an extraocular electrode array inserted via a scleral incision, enabling wireless transmission of visual data from an external camera.[52][54][53][55][56] Key advantages of suprachoroidal placement include reduced surgical risk, as no vitrectomy or intraocular manipulation is required, leading to shorter operative times and lower complication rates compared to epiretinal or subretinal methods.[52] It also supports a broader field of view by accommodating larger electrode arrays and is particularly suitable for patients with profound outer retinal degeneration, where inner retinal layers remain viable for indirect stimulation.[56] In Australian clinical trials during the 2020s, the 44-channel suprachoroidal prosthesis demonstrated stable implantation with no serious adverse events over two years and provided functional gains, such as improved object recognition and mobility, in participants with retinitis pigmentosa.[55][56] Emerging retinal implant types extend beyond traditional placements by targeting alternative neural pathways or integrating biological enhancements. Intracortical visual prostheses, such as the Orion system now developed by Cortigent (formerly Second Sight Medical Products), stimulate the visual cortex directly via electrode arrays implanted in the occipital lobe, bypassing the retina and optic nerve entirely to restore form vision in profoundly blind individuals.[57] First human implants occurred in 2018. By 2025, the European clinical trial was completed, with application for approval submitted, showing participants perceiving phosphenes and basic shapes, and positive 5-year stability data.[58][37] Optogenetic hybrid approaches combine genetic delivery of light-sensitive proteins (opsins) to RGCs with implantable devices that provide supplementary electrical stimulation, aiming to lower optogenetic activation thresholds and enhance high-frequency responses for more natural vision restoration.[59] These systems, tested in preclinical models, express opsins via viral vectors to make neurons responsive to safe light levels, with hybrid stimulation extending device longevity by reducing power demands.[60] Fully wireless photovoltaic systems represent another innovation, using subretinal pixel arrays that convert projected near-infrared light into electrical currents without onboard batteries or inductive powering, enabling untethered operation and high pixel densities for improved resolution.[10] The PRIMA implant, for instance, consists of a 2x2 mm microarray with 378 pixels, implanted under the retina to stimulate bipolar cells in patients with geographic atrophy, with 81% of trial participants achieving clinically meaningful visual acuity improvements in a 2025 study.[61][46] These designs prioritize minimal invasiveness and scalability, with prototypes demonstrating stable performance in degenerative models.[62]Design and technology
Electrode arrays and stimulation
Electrode arrays in retinal implants are engineered for biocompatibility, flexibility, and precise neural interfacing to deliver targeted electrical stimulation to surviving retinal cells. Electrodes are typically fabricated from platinum-iridium alloys due to their excellent charge transfer properties and resistance to corrosion in physiological environments, while substrates employ flexible polymers like polyimide or parylene to conform to the curved retinal surface without causing mechanical damage. Silicone is also used in some designs for its durability and insulation qualities. These material choices ensure chronic implantation safety, with polyimide offering superior flexibility for high-density arrays compared to more rigid alternatives like silicone.[63][64][65] Array configurations vary to balance resolution and surgical feasibility, commonly featuring 60 to 1,500 electrodes arranged in rectangular or hexagonal grids with inter-electrode spacings of 100 to 500 µm. For example, early clinical systems like the Argus II employ a 6x10 grid of 60 platinum electrodes, each 225 µm in diameter and spaced 575 µm apart, providing basic pattern recognition. Higher-density prototypes aim for over 1,000 pixels to approach usable vision thresholds, such as arrays with 512 channels at 60 µm pitch for enhanced spatial resolution in preclinical models. Electrode placement varies across implant types, with epiretinal designs positioned on the inner retinal surface and subretinal ones inserted beneath the retina.[66][67][68] Stimulation protocols rely on biphasic current pulses to safely activate neurons while minimizing tissue damage and electrode degradation. These pulses typically follow a cathodic-first (negative then positive) waveform, with durations of 0.25 to 1 ms per phase and amplitudes of 0.1 to 1 mA, delivering charge-balanced stimulation to evoke localized phosphenes without net charge accumulation. The total charge per phase, calculated as Q = I \times t, must remain below material-specific limits—such as approximately 50 nC for platinum electrodes or up to 1 mC/cm² for iridium oxide coatings—to avoid electrolysis and irreversible faradaic reactions. Frequencies are set between 20 and 50 Hz to produce flicker-free percepts, mimicking natural temporal vision dynamics.[69][70][71] Physiological adaptation accounts for the implant's target cells and retinal architecture, with epiretinal systems directly stimulating retinal ganglion cells (RGCs) via extracellular currents that propagate along axons, potentially causing overlapping activation. Subretinal implants preferentially target bipolar cells through closer proximity, enabling more focal responses that preserve retinotopic mapping—the spatial organization mirroring the visual field's projection onto the retina. Electrode arrays are wired to maintain this retinotopy, ensuring stimulated patterns correspond to external visual inputs for coherent perception.[72][73][74] Recent advancements include ultra-flexible electrode cuffs and polyimide-based arrays that enhance retinal conformance and reduce migration risks, alongside optoelectronic hybrids integrating micro-LEDs for combined electrical and optogenetic stimulation to improve specificity and efficiency. These innovations, such as liquid-metal-embedded 3D microelectrode arrays, boost charge injection capacities up to 72 mC/cm² while maintaining biocompatibility. As of 2025, upgrades to subretinal photovoltaic implants have reduced pixel sizes to 22 µm, enabling grating acuities up to 28 µm and approaching natural resolution limits.[75][41][76][77]Power supply and image processing
Retinal implants primarily rely on inductive coupling for power delivery, where an external coil transmits radiofrequency (RF) energy to an implanted receiver coil, enabling wireless operation without internal batteries. In systems like the Argus II, this involves a carrier frequency of 3.156 MHz for power and data transmission, delivering less than 1.2 watts to the implant while maintaining efficiency over short distances of up to 25 mm.[78] Emerging photovoltaic approaches, such as those in subretinal prostheses, utilize pulsed near-infrared (NIR) light (880–915 nm) projected from external goggles to directly power silicon photodiodes in each pixel, eliminating the need for inductive links or batteries and achieving power densities of 0.2–10 mW/mm² within ocular safety limits.[62] Visual input is captured by an external camera mounted on video glasses and fed to a video processing unit (VPU) for real-time conversion into stimulation signals. The VPU digitizes the camera feed, applies filters for edge detection (e.g., using an Inverse Gaussian filter for emphasis) and contrast enhancement (e.g., via Gaussian filtering and histogram equalization), and downsamples the image to match the electrode array's resolution, such as a 6×10 pixel grid for the Argus II's 60 electrodes.[79][78] Key algorithms include grayscale conversion to simplify the image for phosphene generation, typically using the luminance formula I(x,y) = 0.299R + 0.587G + 0.114B, which weights red, green, and blue channels according to human perception.[78] Phosphene mapping then aligns processed pixels to electrode positions via retinotopic lookup tables, preserving spatial organization where phosphene locations correspond to retinal electrode placements, often spanning 0°–60° eccentricity in the visual field.[80] Challenges in power supply and processing include managing heat dissipation to limit tissue temperature rises below 2°C, constrained by safe power budgets of 10–50 mW for multi-electrode arrays to avoid thermal damage.[81] Data rates must also support high-resolution stimulation, reaching up to 1 Mbps for advanced systems, though current implementations like the Argus II operate at 700 kb/s to balance power efficiency.[81]Implantation and operation
Surgical procedure
Preoperative preparation for retinal implant surgery involves comprehensive retinal assessment using optical coherence tomography (OCT) and fundus photography to evaluate the integrity and position of the retina, ensuring suitability for device placement. Patients undergo general anesthesia, with antibiotic prophylaxis administered intravenously to minimize infection risk.[28][82] The surgical procedure typically lasts 2 to 4 hours and begins with a pars plana vitrectomy to clear the vitreous humor and provide access to the retina. For epiretinal implants, such as the Argus II, a 360-degree conjunctival peritomy is performed, followed by creation of a sclerotomy incision; the electrode array is then inserted into the vitreous cavity and secured to the macular surface using a retinal tack or suture. In subretinal approaches, like the Alpha IMS or photovoltaic implants such as PRIMA, a retinotomy is made to create a subretinal bleb, and the array is advanced under the retina via a surgical glide or tool after scleral pocket formation; for PRIMA, the array is placed under the macula without mechanical fixation, using a standard vitrectomy approach lasting under 2 hours.[28][7][82][46] Suprachoroidal implantation, as in the Bionic Vision Australia device, involves a scleral flap or incision to access the suprachoroidal space, allowing array placement without vitrectomy.[28][7][82] During surgery, the electrode array is implanted directly onto or near the retina, while the external receiver coil and electronics case are positioned under the conjunctiva or sclera, often secured with sutures and a silicone band to prevent migration. Intraoperative testing of electrical stimulation is conducted to verify array functionality and phosphene generation before wound closure.[28][82][7] Postoperatively, patients receive topical corticosteroids and antibiotics to monitor and manage inflammation, with intraocular pressure and wound healing assessed through regular examinations. Sclerotomy sites are closed with sutures and covered to reduce infection risks, such as endophthalmitis. Hospital monitoring typically lasts several days, followed by outpatient follow-up.[28][82][7]Patient training and rehabilitation
Following implantation of a retinal prosthesis, initial activation typically occurs 1-4 weeks post-surgery, after surgical recovery, to allow healing and minimize risks such as inflammation or discomfort. During this phase, the device is gradually powered on with a ramp-up of electrical stimulation to prevent sensory overload, starting with low-intensity pulses that elicit basic phosphenes—perceived spots of light. Clinicians perform software calibration to adjust parameters like phosphene brightness, contrast, and threshold, tailoring the output to the patient's individual neural responses and ensuring stable perception without discomfort. This process is critical for epiretinal systems like the Argus II, where external components such as glasses-mounted cameras are fitted and tested for alignment with the patient's gaze.[83][84] Patient training protocols emphasize structured rehabilitation to build functional use of prosthetic vision, often spanning 3-6 months of intensive sessions combining clinic-based and home practice. Computer-based tasks form the core, including letter or symbol recognition, contrast detection, and basic pattern identification to map phosphene patterns to visual cues; for instance, in Argus II programs, patients practice visual scanning and object tracking using customized software interfaces. Mobility training integrates these skills into real-world navigation, such as obstacle avoidance courses or timed up-and-go tests, frequently aided by white canes or harnesses for safety, as seen in protocols employing virtual reality systems like CAREN for simulated environments. Subretinal implants, such as PRIMA, incorporate similar expert-led sessions focused on habituating to projected light patterns, with assessments of visual function to refine device settings. Overall, these programs progress from controlled settings to independent activities, supported by rehabilitation kits that include practice aids like audio-guided exercises.[84] Adaptation presents significant challenges, as patients must learn to interpret coarse, low-resolution phosphene arrays—often resembling scattered spots or blobs that approximate edges rather than detailed images—requiring cognitive remapping of visual inputs long absent due to blindness. This process can induce fatigue, disorientation, or frustration, particularly when expectations of natural vision clash with the device's limitations, necessitating psychological support through counseling to manage emotional adjustment and maintain motivation. For example, in epiretinal prostheses, head movements must be exaggerated to scan scenes, while subretinal systems like Alpha IMS demand acclimation to fixed-field perceptions without external cameras. Ongoing support from low-vision specialists helps mitigate these hurdles, fostering gradual proficiency in daily tasks.[8][51] Rehabilitation outcomes demonstrate that, with consistent training, a majority of patients—approximately 70-80% in key cohorts—achieve basic navigation abilities, such as independent mobility in familiar environments or simple object localization, enhancing quality of life. Tools like head-mounted display simulations during practice sessions accelerate this adaptation by replicating phosphene views in controlled scenarios, allowing repeated exposure without real-world risks. Long-term adherence to device use and rehab correlates with sustained gains, though individual variability depends on factors like pre-implantation light perception and cognitive resilience.[43]Performance and limitations
Visual acuity and spatial resolution
Visual acuity in retinal implants is fundamentally limited by the density of electrodes, which determines the spatial resolution of the induced phosphene patterns. For the Argus II epiretinal implant (discontinued in 2019) with 60 electrodes spaced 575 μm apart, the achieved visual acuity is equivalent to 20/1260 on the Snellen scale, corresponding to a grating acuity of approximately 0.3 cycles per degree (range 0.22–0.34 cpd in trials).[85][86][87] This performance falls short of the theoretical maximum resolution, estimated as ≈ (290 μm/degree) / (2 × electrode spacing in μm), yielding about 0.25 cycles per degree for the Argus II's 575 μm spacing, assuming a standard retinal-to-visual angle conversion of roughly 290 μm per degree.[78] The perceptual quality of vision is shaped by phosphene characteristics, where each electrode elicits a spot of light typically 0.5–2° in angular size. Electrode spacing governs the overall field of view, limited to about 20° diagonally in early devices like the Argus II (11° × 19°). Additionally, the dynamic range remains poor, supporting only 4–8 grayscale levels due to constraints in stimulation amplitude and patient thresholds.[88][20][89] Key factors influencing acuity include electrode count and stimulation precision, which together restrict patients to perceiving basic shapes and motion while failing to resolve fine details such as facial features. Patient training can modestly enhance perceptual interpretation of these coarse patterns. Newer implants targeting 200+ electrodes, such as the PRIMA subretinal system with 378 pixels of 100 μm, achieve ~20/500 acuity; developments with smaller pixels (e.g., 22 μm) target 20/200 by increasing density. As of 2025, PRIMA trials report mean visual acuity of ~20/420 at 12 months.[87][45][46]Safety profile and complications
Retinal implants are designed with biocompatible materials to minimize adverse tissue reactions, such as corrosion-resistant coatings like titanium nitride (TiN) on electrodes, which exhibit no cytotoxicity or irritation in standard biocompatibility tests.[90] These materials promote chemical stability and reduce inflammatory responses, enabling long-term integration with retinal tissue. Tissue responses, including glial scarring, are generally minimal, with glial encapsulation limited to thicknesses of approximately 5-10 µm around implant sites in neural interfaces, though retinal-specific scarring can vary and may contribute to reduced efficacy over time.[91] Surgical implantation of retinal prostheses carries risks inherent to vitreoretinal procedures, including retinal detachment occurring in about 6.7% of cases for rhegmatogenous types in epiretinal systems like the Argus II. Infections, such as presumed endophthalmitis, affect around 10% of patients, while vitreous hemorrhage and device-specific issues like tack migration necessitating re-tacking occur in 6.7% of implantations. These complications are typically manageable with standard ophthalmic interventions, with no reported loss of eyes in long-term follow-up studies.[92][92][92] Over extended periods, electrode degradation can manifest as impedance changes, with some subretinal implants showing initial increases of up to two orders of magnitude in the first few weeks due to tissue adaptation, though suprachoroidal and epiretinal systems often maintain stable impedances (5-10 kΩ) for 3-5 years. Immune rejection remains rare, but encapsulation by fibrous tissue is possible, potentially insulating electrodes and altering stimulation efficacy. In representative trials, such as with the Argus II, device failures leading to loss of function occurred in about 6.7% of cases after 4 years, primarily from non-electrode components like antennas.[93][94][95] Ongoing monitoring of retinal implants involves annual optical coherence tomography (OCT) to assess retinal thickness and implant position, alongside electroretinography (ERG) to evaluate functional integrity and detect early changes in retinal response. Explantation rates are approximately 10%, often due to persistent issues like conjunctival erosion or hypotony, as observed in clinical trials where devices were removed without severe sequelae.[96][97][92]Clinical evidence
Key clinical trials
The Argus II epiretinal prosthesis underwent Phase I/II clinical trials from 2007 to 2009, enrolling 30 patients with retinitis pigmentosa (RP) who had bare or no light perception. The primary efficacy endpoint focused on object localization using a square detection task, where 96% of subjects (27 out of 28 testable patients) demonstrated significantly better performance with the device activated compared to deactivated, as measured by reduced distance from the target center. This trial established a favorable safety profile with one explantation due to recurrent erosion, but no other device- or surgery-related serious adverse events leading to explantation, paving the way for the FDA pivotal trial completed in 2012, which supported approval in 2013 for profound vision loss due to RP.[85] The Alpha IMS subretinal implant was evaluated in a European clinical trial initiated in 2012 and spanning through 2016, with initial results reported from the first 9 patients (8 with RP, 1 with cone-rod dystrophy) blind from hereditary retinal degenerations. Patients achieved prosthetic visual acuity up to 20/546 Snellen equivalent using Landolt C-rings for letter recognition, enabling basic object shape discrimination and motion detection in some cases. Overall, the trial demonstrated sustained implant functionality for up to 9 months in the early cohort, with some complications such as subretinal bleeding and optic nerve touch, but no severe long-term device-related issues.[98][51] The PRIMA photovoltaic subretinal implant's feasibility trial for atrophic age-related macular degeneration (AMD) ran from 2018 to 2021, involving 5 patients with central vision loss due to geographic atrophy. Implanted under the fovea, the device projected infrared images via external glasses, resulting in improved navigation tasks such as obstacle avoidance and orientation, with all participants reporting enhanced functional vision for daily activities like reading large print or identifying objects. The trial supported the pathway to CE marking in 2021, with primary safety endpoints met as no serious adverse events occurred, and prosthetic acuity gains averaged 0.4 logMAR in responders. As of 2025, expanded trials (PRIMAvera) with 38 patients confirm sustained 12-month improvements in visual acuity (mean 0.51 logMAR gain) and reading ability in 84%.[99][100][46] The Bionic Vision Australia suprachoroidal retinal prosthesis pilot trial, conducted from 2014 to 2018, implanted 3 patients with end-stage RP using a 20-electrode array. Participants achieved 72%-100% accuracy in light localisation tasks with the device on, significantly outperforming off states, which supported basic orientation and motion perception. The trial confirmed surgical feasibility with stable electrode placement and superficial skin infections resolved with antibiotics, but no intraocular infections or migrations, establishing proof-of-concept for suprachoroidal stimulation in preserving residual peripheral vision.[101][102]Long-term outcomes
Long-term follow-up studies of the Argus II retinal prosthesis have demonstrated substantial device durability, with 80% of implanted patients (24 out of 30) retaining functioning systems at five years post-implantation, and visual performance remaining stable as measured by consistent superiority in system-ON versus system-OFF conditions on tasks such as square localization and grating visual acuity.[92] Electrode deactivation, often due to impedance changes or tissue interactions, occurs at rates of approximately 20% after three years in simulated models, though clinical data show 2 device failures (6.7%) and gradual impedance decreases without widespread failure.[103][92] Functional gains persist over multiple years, including significant improvements in mobility such as better orientation and navigation in high-contrast environments, with patients showing enhanced performance in walking toward targets and along lines, enabling greater independence for those previously reliant on guides.[87] Psychological benefits are also evident, with quality-of-life assessments indicating reduced emotional distress, including lower depression scores associated with restored light perception and basic visual cues that alleviate isolation.[104] Quality-of-life metrics, such as the National Eye Institute Visual Function Questionnaire-25 (NEI-VFQ-25), reveal sustained gains in daily living activities, with patients reporting improved ability to perform tasks like identifying objects and navigating spaces, contributing to overall well-being.[105] The device's cost-effectiveness is supported by analyses showing an incremental cost of around $150,000 per implant, offset by returns on investment through increased patient independence and reduced long-term care needs, with favorable quality-adjusted life year ratios compared to standard care.[35][106] Despite these benefits, limitations include no progressive improvement in visual acuity over time, as initial gains plateau, and explantation rates of 10% at five years, with higher rates in some extended post-approval cohorts due to diminishing perceived benefits or complications like erosion.[92][36] For the Alpha IMS, long-term data up to 12 months in expanded cohorts (29 patients) showed sustained low-vision functions with 72% meeting primary endpoints for daily activities, though commercial availability limited further follow-up.[51] The PRIMA implant demonstrated stability at 12 months in the 2025 PRIMAvera trial, with 81% of 32 completers achieving meaningful visual acuity improvements and no loss of residual vision, supporting long-term efficacy for AMD.[46] The Bionic Vision Australia 44-channel device, in a 2024 trial with 4 patients, showed substantial functional vision improvements sustained over 2.5 years, including better object recognition and mobility, with all electrodes functional and no serious adverse events.[56]Current landscape
Commercially available devices
As of 2025, a limited number of retinal implants have achieved regulatory approval and been made commercially available to patients, primarily for those with retinitis pigmentosa (RP) or advanced age-related macular degeneration (AMD). These devices, developed by pioneering companies in neuroprosthetics, have collectively resulted in approximately 500 implants worldwide, with the majority performed in specialized centers such as Bascom Palmer Eye Institute in the United States and Moorfields Eye Hospital in the United Kingdom.[5][35][107] The Argus II Retinal Prosthesis System, developed by Second Sight Medical Products (now under Cortigent), is an epiretinal implant approved by the FDA in 2013 under a humanitarian device exemption and receiving CE marking in 2011. It features a 60-electrode array attached to the retina, powered by an external camera and processing unit worn on glasses, enabling patients to perceive light, motion, and basic shapes. Over 350 patients received the implant globally before manufacturing was discontinued in 2020 due to financial challenges, though legacy devices remain active with limited support from Cortigent, and no new units are being produced.[36][108] The Alpha AMS, a subretinal implant from Retina Implant AG, obtained CE marking in 2013 and was designed for RP patients, incorporating 1,600 light-sensitive pixels directly under the retina to convert light into electrical signals without external cameras. Limited commercial availability followed the company's bankruptcy in 2016, with fewer than 50 implants performed primarily in Europe; the device demonstrated improved visual function in clinical users but is no longer supported or manufactured.[109][39] The IRIS II (also referred to as IRIS V in some contexts), an epiretinal system by Pixium Vision, received CE marking in 2016 for RP patients in Europe, utilizing a 150-electrode array with an implantable stimulator and external camera for enhanced object recognition and mobility. Approximately 30-40 implants were completed before Pixium's bankruptcy in 2024, after which assets were acquired by Science Corporation; the device is no longer commercially available, though existing implants continue to function.[110][39] Access to these devices remains constrained by high costs, ranging from $100,000 to $200,000 including surgery and rehabilitation, with coverage available in select European healthcare systems but limited in the U.S. to investigational or humanitarian pathways. No new FDA approvals for retinal implants have occurred since 2013, and current availability relies on legacy systems in expert centers. Meanwhile, the PRIMA subretinal implant by Science Corporation, targeting AMD, is nearing CE marking following a June 2025 application submission and positive trial data, but it is not yet commercially accessible.[35][84][3]| Device | Type | Regulatory Approvals | Manufacturer (Status) | Approximate Implants | Target Condition |
|---|---|---|---|---|---|
| Argus II | Epiretinal | FDA (2013), CE (2011) | Cortigent (legacy support) | >350 | RP |
| Alpha AMS | Subretinal | CE (2013) | Retina Implant AG (bankrupt 2016) | <50 | RP |
| IRIS II | Epiretinal | CE (2016) | Pixium Vision (bankrupt 2024) | 30-40 | RP |