Seeing anew: Mechanisms, outcomes, and challenges in adapting to restored sight via visual prostheses Authors: Katerina Eleonora K Rassia, Lydia Liapi, John S Pezaris Published: 2026 ABSTRACT AbstractRestoring vision through prosthetic devices represents one of the most ambitious frontiers in translational neuroscience. Despite decades of progress, functional outcomes remain highly variable, often falling short of patient expectations, as adapting to restored vision through visual prostheses is a complex process influenced by neural plasticity, perceptual learning, and psychological factors. We propose a five-stage framework that captures the hierarchically dependent yet fragile process of adaptation: Device Activation, Perceptual Calibration, Cognitive Integration, Behavioral Generalization, and Psychosocial Adjustment. Across retinal and cortical prosthesis designs, the earlier stages are frequently achieved, but progression to fluent, real-world use and sustained device acceptance is far less common. The most consistent point of failure lies between Cognitive Integration and Behavioral Generalization: skills acquired in structured training rarely transfer to unconstrained everyday behavior. Technical constraints, insufficient training, and unrealistic expectations frequently stall adaptation at this threshold, leading to frustration, limited daily use, or device abandonment. By examining visual-prosthesis clinical trials, case studies of surgical vision restoration, and evidence from cochlear implants and sensory substitution devices, we identify key determinants of successful adaptation: timely intervention, intuitive device usability, sustained rehabilitation, and long-term psychosocial support. Understanding the mechanisms of adaptation is essential for improving visual prostheses and ultimately enhancing quality of life for individuals with restored sight.Keywords: Artificial vision, Visual plasticity, Visual rehabilitation, Sensory substitution, Perceptual learning, Skill transfer Introduction While there have been, and continue to be, substantial advances in the design and development of visual prosthesis devices to restore sight, patient outcomes are not yet universally positive, and there is a lack of general understanding to explain why. Here, we review the literature of artificial vision specifically to examine the adaptability of post-implant patients to their new visual modality and the factors that affect both successful and unsuccessful results. We will look for conditions leading to the recipient of a sight-restoring intervention engaging their device as easily and readily as the sighted would reach for corrective lenses, versus those leading to an endpoint where the patient no longer uses their device regularly, and returns to daily life without visual input. These broad-stroked criteria are difficult to assess with objective, quantifiable and rigorous metrics that accurately reflect successful outcomes. For example, the standard measurement of restored acuity with a traditional clinical test may reflect adequate performance, perhaps even above thresholds for legal blindness, but if the patient does not find advantage to the device, a personal and subjective evaluation that is challenging to quantify, the rewards have not justified the risks of treatment. Fortunately, the multi-decadal, continuing development of the cochlear auditory prosthesis to address deafness provides an established model of functional restoration in a different sensory system. And thus, we include here a selective summary of relevant work from the cochlear implant field in order to gain additional insight on the process of recovery of sensory function, that we will combine with examples from sensory substitution devices for vision, surgical restoration of vision, and simulations of artificial vision. These branches will serve as comparative controls for progress achievable with current visual prostheses. Ultimately, the evidence we gather will support a new framework for understanding adaptation to prosthetic vision, as introduced below, that will guide and focus developmental efforts where task performance has not translated into sustained daily use and long-term device retention. A new framework for understanding adaptation for artificial vision The diverse outcomes observed across visual prosthesis devices and users, from immediate perceptual challenges to long-term psychosocial responses, can be understood within a unified framework that captures the hierarchically dependent nature of adaptation. We propose adaptation to artificial vision as a progression through five stages with each stage shaped by a distinct set of influences (Fig. 1) and quantitative evaluations (Table 1). While some individuals successfully traverse all stages and achieve functional acceptance of their device, others disengage at intermediate stages, leading to device abandonment. This model captures disparate findings from clinical performance metrics, subjective reports, and case studies, helping to identify where and why failures in adaptation may occur. Understanding these stages can guide the design of future devices and rehabilitation programs by highlighting critical areas where developmental effort may improve outcomes. Stages of visual prosthesis adaptation and their influences. This figure outlines five hierarchical stages involved in adapting to a visual prosthesis. Each stage is shown with a primary influencing factor. Device rejection can happen at any stage, but acceptance happens only once Psychosocial Adjustment (Stage 5) has been achieved. Table 1An operational definition of the Five Stages of Adaptation.Each of the five stages in the proposed framework is defined by observable behavioral evidence and explicit exclusion rules, allowing reported outcomes from diverse clinical, psychophysical, and qualitative studies to be evaluated using a common classification framework.StageFunctionalMeaningObservableInclusion CriteriaExplicitExclusion CriteriaMinimumReporting RequiredTypical Task1 Device ActivationThe system produces sensory percepts.Detection or categorical localization of stimulation; phosphenes time-locked to stimulation.Subjective statements of sensation without controlled confirmation.Any controlled detection task or documented reproducible percept elicited by stimulation.nAFC phosphene detection paradigms; threshold estimation.2 Perceptual CalibrationThe percept carries interpretable stimulus information.Discrimination of spatial, temporal, or motion properties requiring finer mapping in laboratory tasks.Simple detection, brightness judgment, or awareness without discrimination.Discrimination or mapping task demonstrating stimulus-dependent percept variation.Spatial localization; orientation, direction-of-motion, or size discrimination.3 Cognitive IntegrationArtificial vision is used to perform goal-directed tasks.Object identification, visuo-motor integration, interpretation of visual patterns.Only forced-choice tasks reported; performance relying solely on scanning strategy without interpretation.A task where vision serves as an information source rather than a detection cue.Letter, shape, object recognition; word reading; closed-set identification; hand-eye coordination.4 Behavioral GeneralizationArtificial vision transfers beyond structured testing.Use in real-world activities without experimenter control.Performance restricted to laboratory or training environments.Use of vision in natural settings or unsupervised environments.Door finding; line following; obstacle avoidance; daily living tasks.5 Psychosocial AdjustmentArtificial vision becomes voluntarily adopted.Sustained everyday use, preference over non-visual strategies, or lifestyle integration.Occasional, novelty-driven, therapist-prompted, or demonstration use.Long-term voluntary use outside training context.Device logs; longitudinal retention; quality-of-life instruments; structured interviews. An operational definition of the Five Stages of Adaptation. Each of the five stages in the proposed framework is defined by observable behavioral evidence and explicit exclusion rules, allowing reported outcomes from diverse clinical, psychophysical, and qualitative studies to be evaluated using a common classification framework. Operational definition of stages To ensure the framework functions as an analytical tool, each stage is defined by behavioral markers reported in the literature, with stage assignment determined by reported outcomes. Each stage is introduced below with additional detail provided in Table 1, defined using criteria that can be employed to classify outcomes directly from future reported results as they become available. Stage 1 (S1), Device Activation, is defined by reliable elicitation of percepts following stimulation. Evidence includes detection of flashes, phosphenes, or categorical light localization (e.g., quadrant-based) significantly above chance during controlled testing. Stage 1 (S1), Device Activation, is defined by reliable elicitation of percepts following stimulation. Evidence includes detection of flashes, phosphenes, or categorical light localization (e.g., quadrant-based) significantly above chance during controlled testing. Stage 2 (S2), Perceptual Calibration, is defined by consistent mapping between percept and stimulus properties. Evidence includes discrimination of position, motion direction, orientation, or shape under controlled laboratory conditions. Stage 2 (S2), Perceptual Calibration, is defined by consistent mapping between percept and stimulus properties. Evidence includes discrimination of position, motion direction, orientation, or shape under controlled laboratory conditions. Stage 3 (S3), Cognitive Integration, is defined by goal-directed task use of vision requiring interpretation rather than detection. Evidence includes object identification, navigation decisions, or task completion using vision as an active information source. Stage 3 (S3), Cognitive Integration, is defined by goal-directed task use of vision requiring interpretation rather than detection. Evidence includes object identification, navigation decisions, or task completion using vision as an active information source. Stage 4 (S4), Behavioral Generalization, is defined by transfer of visual use outside controlled testing. Evidence includes real-world mobility, activities of daily living, or spontaneous use of vision in unconstrained environments. Stage 4 (S4), Behavioral Generalization, is defined by transfer of visual use outside controlled testing. Evidence includes real-world mobility, activities of daily living, or spontaneous use of vision in unconstrained environments. Stage 5 (S5), Psychosocial Adjustment, is defined by sustained voluntary adoption. Evidence includes continued everyday use, preference for device-mediated vision over non-visual strategies, or integration into lifestyle and identity. Stage 5 (S5), Psychosocial Adjustment, is defined by sustained voluntary adoption. Evidence includes continued everyday use, preference for device-mediated vision over non-visual strategies, or integration into lifestyle and identity. The stages are ordered by hierarchical dependence rather than by temporal sequence: each requires the capabilities of those below it, but progression need not be monotonic, uniformly paced, or irreversible. A stage assignment therefore describes the highest stage supported by the available evidence rather than a fixed position occupied at all times. We will use these stages to assess the underlying capabilities of a given device by looking at population data, but they can also be used to gauge the progression of an individual patient. For the devices we will review, a stage is considered supported when the framework criteria are met by a majority of recipients across independent tasks. Progression of individual cases, including exceptional high-performing or highly motivated ones, will be discussed descriptively to illustrate variability and adaptation trajectories and to separate fundamental technological capability from user-specific variability. Evidence for a device will be considered sufficient for a given Stage only when reported population performance is statistically significant, or there is statistically significant performance for a majority of subjects, as compared to baseline or chance. It should be noted that statistically significant performance does not necessarily imply performance levels that are meaningful or useful, as will be discussed throughout. In general, if a device provides performance at a given stage only by statistical significance against blindness, and not through performance comparable to normal vision, the device often does not qualify at the next stage. For safety-critical tasks (such as in Stage 4), a higher reliability threshold is required. Performance must demonstrate consistently high success rates (at least 99% correct or the equivalent) without external assistance, reflecting the lower tolerance for error inherent in such behaviors. Stage assignment is referenced to comparator-based performance rather than to absolute values of any single metric. The metrics at each stage are often reported in incommensurable units across devices, and their magnitudes are bounded by device factors such as electrode pitch and pattern coverage. A threshold placed on such metrics would therefore emphasize device hardware rather than recipient adaptation, but a recipient who identifies letters through a coarse array has performed the same act of interpretation as one who identifies letters through a fine array. That said, where an absolute criterion is needed because of potential impact, as with safety-critical behavior at Stage 4, it is included. When multiple tasks within a given stage are reported for a device, the criterion must be met in a majority of independent assessments. This breadth requirement prevents stage assignment based on isolated success in a single narrow or highly trained paradigm. If only a single task is reported with qualifying performance, then stage classification is considered to be tentative, limited, or partial. Scope and selection of evidence This review is intended as an integrative synthesis rather than a systematic review. The goal is to examine the extent to which restoration of visual function through visual prosthesis devices is plausible by evaluating reports describing behavioral outcomes following patterned visual input to the nervous system. Because the proposed framework concerns adaptation rather than device engineering, evidence was drawn primarily from long-term clinical prosthesis studies in human participants, supplemented by cases of non-prosthetic sight restoration that produce comparable perceptual challenges. Cochlear implants are considered as a comparative model of sensory restoration, and sensory substitution devices examined in controlled laboratory settings or extended training paradigms are included where they inform mechanisms of adaptation. Studies were selected based on their relevance to understanding adaptation to artificial vision and the availability of post-implantation behavioral, experiential, or psychosocial data. For each project described below, all reports available through indexed searches were reviewed. Publications on fully developed devices were selected over developmental, proof-of-concept, or exploratory reports. The assembled evidence is organized into a classification that describes the highest demonstrated stage within the framework. Evidence using digital or optical magnification was not included. Quantitative psychophysical measurements and structured functional tasks were considered primary evidence. Observational functional reports were treated as supporting evidence. Psychosocial and qualitative sources were included when they provided first-person accounts, structured interviews, or validated questionnaires addressing user experience, expectations, identity, quality of life, or long-term device use. Detailed first-person patient descriptions were included along with primary scientific reports. Qualitative interviews and patient narratives were used to characterize patterns of use, expectations, and abandonment, but were not treated as prevalence estimates. The availability of systematic qualitative assessment remains limited across artificial vision technologies, with many (but not all) reports based on small samples, single-case studies, or retrospective accounts. Psychosocial outcomes are therefore discussed to identify recurring themes and adaptation challenges rather than to estimate prevalence or generalizability. Selection of devices Within the extent of visual prosthesis projects, we reviewed the ones with results from human trials that included evidence for adaptation at the various stages described in the framework, where devices are currently under development, as well as key historical projects. Projects of historical importance that are reviewed include the Argus II, Alpha IMS/AMS, and Dobelle prosthesis. We did not include Brindley and Lewin's seminal experiments (Brindley and Lewin, 1968); although foundational to the field, they were exploratory in nature, rather than the testing of a device ready for clinical deployment. Contemporary devices that are reviewed include the BVT suprachoroidal prosthesis, IMIE 256, CORTIVIS, PRIMA, and Orion. Restoration of vision Basic operation of visual prostheses Visual prostheses generate artificial sight by electrically stimulating intact portions of the visual pathway to elicit visual percepts known as phosphenes (Brindley and Lewin, 1968). These devices translate input from a camera or sensor array into patterned electrical signals delivered via implanted electrodes. The resulting percepts are spatially coarse, often (but not always) spatially separated, and influenced by multiple biological and technical variables, including electrode-tissue interface dynamics, eye movements, and the nonlinearities of neural coding. Despite these limitations, some users can learn to interpret phosphene patterns as meaningful visual information, for example, Argus II retinal prosthesis recipients have reported recognizing large objects, doors, and people during navigation tasks (da Cruz et al., 2016; Luo et al., 2015), potentially making artificial vision functionally useful in daily tasks under specific conditions (e.g., high-contrast environments with structured training). Efforts in the field of visual prosthetics have targeted four possible locations for electrical stimulation along the early visual pathway (The Lasker/IRRF Initiative for Innovation in Vision Science, 2014): the retina, the optic nerve, the lateral geniculate nucleus (LGN) of the thalamus, and the primary visual cortex. Table 2 provides an overview of major visual prosthetic devices, categorized by active/inactive status and clinical stage, providing a snapshot of the field's current landscape. Table 2Visual prosthetic devices summary.This table provides an overview of the translational readiness for many of the prominent visual prosthesis devices along with the name of the company or research group responsible. Devices are grouped according to current activity and clinical stage. Sensor location indicates where the imaging sensor is placed, as part of the stimulating implant (such as in an epiretinal design), or an externally worn apparatus. Gaze contingency refers to either natural gaze steering enabled by the imaging sensor being intrinsically controlled by eye movements, or gaze compensation achieved through an eye-tracking system. For reported acuity, n/a indicates the data are not available. Features and limitations are guiding characteristics of each device. For each device, a representative reference is given from the project's literature.DeviceCompany orResearch GroupStimulationTargetActiveClinicalStatusSensorLocationGazeContingentBestAcuityFeatures and LimitationsKey ReferenceOPTO-EPIRETEpi-Ret GmbH, Germanyepiretinalnoanimalintegratedinherentn/aintegrated sensor chips; dense center; wide FOV; surgical optimization needed Schaffrath et al. (2021) BRIPBionic Eye Technologies, USAsubretinalnoanimalexternalnon/ahermetic package; no tack fixation; lower infection risk Kelly and Rizzo, 2017 V-DONENIDEK Ltd, Japanoptic nervenofeasibilityexternalnon/a3 wire electrodes; focal phosphenes; low thresholds; variable phosphene size Nishida et al. (2015) OptiVipUCLouvain, Belgiumoptic nervenofeasibilityexternalnon/a4-contact spiral cuff; broad phosphenes; simple patterns; limited selectivity Brelén et al. (2005) Brelén et al. (2010) EPIRET3Epi-Ret GmbH, Germanyepiretinalnoclinical trialexternalnon/a25 electrodes; protruding contacts; low thresholds; fully intraocular Menzel-Severing et al. (2012) Argus IISecond Sight Medical Products, USAepiretinalnoapprovedexternalno20/126260 electrodes; limited FOV; high cognitive load; the most widely implanted device Güven et al. (2023) IRIS IIPixium Vision, Franceepiretinalnoapprovedexternalnon/a150 electrodes; modular design; flexible array Muqit et al. (2019) Alpha IMS Alpha AMSRetina Implant AG, Germanysubretinalnoapprovedintegratedinherent20/54620/5001500/1600 stimulation sites; limited FOV; no camera/gogglesStingl et al. (2015); Edwards et al. (2018)POLYRETINAEPFL, SwitzerlandETH, Switzerlandepiretinalyesanimalintegratedinherentn/a10,498 photovoltaic pixels; ~43° FOV; curved flexible implant Vagni et al. (2022) Thalamic Visual ProsthesisMass General Hospital, USAthalamicyesanimalexternalyesn/a500+ phosphenes; full field coverage Killian et al. (2016) VISNEComplutense Univ of Madrid, Spainthalamicyesanimalexternalnon/ato be determined Panetsos et al. (2011) GennarisMonash Vision Group, Australiacorticalyesanimalexternalnon/a43-electrode wireless tiles; no transcranial wires; scalable architecture; limited resolution Rosenfeld et al. (2020) TBDPhosphoenix BV, NetherlandsTBDyesanimalexternalyesn/ato be determined Chen et al. (2020) STSJapan Artificial Vision Project NIDEK Ltd, Japansuprachoroidalyesfeasibilityexternalnon/a49 channels; low retinal trauma; farther from neurons; stable placement; lower resolution Endo et al. (2018) ICVPIllinois Institute of Technology, USAcorticalyesfeasibilityexternalnon/aunexpectedly low resolution; disadvantageous phosphene pattern; limited FOV Barry et al. (2023) PRIMAScience Corp, USAsubretinalyesclinical trialintegratedinherent20/438high resolution; integrates with surviving peripheral vision; digital zoomMuqit et al. (2024); Holz et al. (2025)IMIE 256IntelliMicro Medical Co, China Golden Eye Bionic, USAepiretinalyesclinical trialexternalno20/800256 electrodes; high-density array; improved safety/efficacy; better localization; limited FOV Xu et al. (2021) OrionCortigent Inc, USAcorticalyesclinical trialexternalnon/a60 channel; based on Argus II architecture; placed on interhemispheric surface; limited FOV Levy et al. (2024) CORTIVISBiomedical Technologies, SL, Spaincorticalyesclinical trialexternalyesn/aUtah array placed in V1 surface; limited FOV Fernández et al. (2021) BVT/Phoenix 99Bionic Vision Australia Bionic Vision Technologiessuprachoroidalyesclinical trialexternalno20/839744-channel 2nd generation; 99-channel 3rd generation under development; limited FOV Petoe et al. (2024) Visual prosthetic devices summary. This table provides an overview of the translational readiness for many of the prominent visual prosthesis devices along with the name of the company or research group responsible. Devices are grouped according to current activity and clinical stage. Sensor location indicates where the imaging sensor is placed, as part of the stimulating implant (such as in an epiretinal design), or an externally worn apparatus. Gaze contingency refers to either natural gaze steering enabled by the imaging sensor being intrinsically controlled by eye movements, or gaze compensation achieved through an eye-tracking system. For reported acuity, n/a indicates the data are not available. Features and limitations are guiding characteristics of each device. For each device, a representative reference is given from the project's literature. Visual prostheses: behavioral performance and device use patterns A handful of visual prosthesis devices are being or have been tested in long-term clinical trials (Table 2). Behavioral studies consistently show that functional performance with prosthetic vision remains well below normal vision, especially in unconstrained environments. Device use is typically restricted to short daily sessions, with most users operating the system for 1–4 h in a day, but not every day (Rojahn, 2013; Nuzbrokh et al., 2021). Capabilities are limited to basic visual tasks such as localization or navigation, and these require sustained cognitive effort. The effortful experience often leads to fatigue and frustration, factors that may contribute to reduced long-term engagement (Stingl et al., 2013; Erickson-Davis and Korzybska, 2021; Kapetanovic et al., 2020; Strickland and Harris, 2022; Maimon et al., 2022). These themes will be revisited as we review individual devices below. Historical devices Argus II The Argus II retinal prosthesis (Humayun et al., 2012) is among the most extensively implanted visual prostheses, with nearly 400 recipients worldwide (Nadolskis et al., 2024), and it has been the subject of the largest body of literature on functional outcomes. While Second Sight Medical Products (Sylmar, California, USA; reorganized in 2022 as Cortigent) ceased development and production of the Argus II in early 2020, follow-up of implanted patients continues (Güven et al., 2023). The system uses an epiretinal electrode array driven by a camera mounted on external goggles, and is steered by head scanning rather than eye motion. Laboratory and clinical evaluations of the Argus II demonstrate reliable progression through the early stages of adaptation, while progress into later stages is uneven. We will use the Argus II as the first application of the five-stage framework; reviews of other devices will be less detailed, but summary tables of the results for each will be given to aid the reader. Stage 1, Device Activation. Stage 1 is established when electrical stimulation reliably produces a consciously experienced visual percept. In the Argus II pivotal trial, all implanted subjects perceived light when their systems were activated, and stimulation thresholds were measurable on at least one electrode in every case (Humayun et al., 2012). Earlier feasibility work likewise demonstrated discrete phosphenes elicited by epiretinal stimulation (Ahuja et al., 2011). Qualitative interviews describe these initial percepts as flashes, sparks, or small points of light that were clearly artificial and required attention to notice (Erickson-Davis and Korzybska, 2021). Since a majority of subjects were able to detect phosphenes, Stage 1 is established. Stage 2, Perceptual Calibration. Stage 2 is reached when recipients learn to associate phosphene patterns with simple external structure. The tests for Argus II used spatial localization and motion discrimination. In the square localization task, subjects identified the position of a bright square presented on a touch screen. Population mean performance was significantly better with the device engaged than with it off (6.1 ± 0.1 cm vs 13.7 ± 0.4 cm, paired t-test, p < 0.01), thus this is a stage-qualifying task. At annual assessments of grating acuity, which required orientation discrimination of static square-wave gratings, performance improved beyond the inclusion threshold of logMAR 2.9 in 33–48% of subjects across the same period (da Cruz et al., 2016); as only a minority passed the threshold, the test is not stage-qualifying, despite the best reported acuity being logMAR 1.8 in one subject (Humayun et al., 2012). Motion perception was evaluated with a drifting bar. Group mean error was lower with the system on than off at every annual visit through five years (e.g., 61° vs 83° at year 3; n = 16–28 across years), though the proportion of individually significant recipients declined from 55.6% at year 3 to 50.0% at year 5 (da Cruz et al., 2016). The task is therefore stage-qualifying through year 3 and marginal by year 5. However, accurate spatial localization in Argus II users depends on regular calibration of the camera position, because perceived visual direction drifts substantially over time (Barry and Dagnelie, 2016). Together, these findings demonstrate that a map of phosphene locations in visual space can be established (Dorn et al., 2013) through which a majority of recipients can perceive spatial features of the environment, and thus Stage 2 is confirmed. Stage 3, Cognitive Integration. Stage 3 is reached when recipients use prosthetic vision to accomplish explicit goals that require interpretation of visual patterns beyond simple detection or discrimination. Letter identification is a prime example. In a forced-choice paradigm, mean percent correct across letters grouped by visual complexity improved from approximately 12–18% with the system off to 52–72% with the system on (p < 0.001 across groups, n = 21 subjects; da Cruz et al., 2013). Because population performance improved significantly, letter identification is a stage-qualifying task, although the 48–69 s average response time per letter does not approach the rapidity of equivalent standard clinical assessments. Four subjects selected for high-performance on the letter task could reliably complete 4-letter word identification (da Cruz et al., 2013); thus word reading is not stage-qualifying because the evidence does not establish the capability across the population. Goal-directed motor tasks further test integration of visual input into action. In a high-contrast reach-to-grasp paradigm (n = 6 subjects), the percentage of successful grasps increased from a median of 0% with the system off to a median of 69% with the system on (Friedman test p = 0.04 at first visit, p = 0.02 at second visit; Kotecha et al., 2014; Luo et al., 2015). Statistical significance was not found between normal prosthesis operation and with a scrambled phosphene map (p = 0.5 at first visit, p = 0.08 at second visit), suggesting that the task was actually testing object detection and not identification, so was being performed by the subjects as a Stage 2 assessment. In a second motor task, testing fine hand-movement tracing, mean trajectory error was reduced by 38–63% with the system on relative to residual vision (p ≪ 0.01; Barry and Dagnelie, 2012), although completion times increased by more than a factor of two, indicating greater cognitive effort. As error reduction was demonstrated at the population level, the tracing task is stage-qualifying. These findings demonstrate that a majority of recipients can use prosthetic percepts to guide purposeful actions and symbolic interpretation, with the caveat that performance of these tasks is effortful and substantially slower than implicit expectations set by normal vision. Accordingly, Stage 3 of adaptation is confirmed, but with the concerns noted. Stage 4, Behavioral Generalization. Stage 4 is reached when artificial vision transfers from structured testing to use in real-world or simulated real-world environments, preferably without direct experimenter supervision, and with accordingly stricter performance criteria for safety-related tasks. With Argus II, some recipients demonstrated measurable benefit in semi-structured daily-life tasks with the device on compared to off, however the absolute benefit was often limited and did not reach framework criteria for Stage 4. In a two-outcome task to discriminate the walking direction of an experimenter (Dagnelie et al., 2017), 66% (18/27) indicated the direction of motion above chance with 25/27 improving with the system on versus off (Renshaw and Pezaris, 2026); thus, this test is a stage-qualifying result. In a door location task, mean success rates across a cohort of 30 patients to walk 20 m to a door-sized black cloth hung on a wall ranged from 51% to 66% between 1 and 5 years (da Cruz et al., 2016); this is a safety-related task, so was not stage-qualifying due to the more stringent staging requirements of 99% or higher correct. The same cohort had a mean success rate for following a high-contrast line on the floor several meters long from 64% to 73% over the same period (da Cruz et al., 2016); this task is also safety-related, so is also not stage-qualifying. Sorting common objects (socks with three different colors) yielded 35% to 70% correct for the population against 33% chance (Dagnelie et al., 2017), which is stage-qualifying. In outdoor assessments, 67% of participants were able to track a sidewalk without deviating more than 1 m from its edge (Dagnelie et al., 2017); as a safety-related task, this result is not stage-qualifying. These results indicate that some, but not all, recipients were able to apply prosthetic vision to mobility and object-related tasks beyond laboratory discrimination paradigms. However, performance even on simplified tasks remained variable, frequently required high-contrast conditions or structured layouts, and often remained within ranges that continued to necessitate navigational aids or assistance. Behavioral Generalization was therefore achieved in selected contexts and in subsets of recipients, but consistent, independent real-world use across environments was not broadly established, and Stage 4 is not supported. Stage 5, Psychosocial Adjustment. Stage 5 is reached when artificial vision is voluntarily adopted, sustained outside structured training, and integrated into daily life in preference to non-visual strategies. Patient-reported outcomes for Argus II recipients indicate perceived benefit but variable long-term reliance. On the FLORA instrument, independent observers evaluated most participants as performing tasks more easily with the prosthesis engaged in domains such as orientation, mobility, and social interaction (Geruschat et al., 2015), however patient self-assessment of impact declined from 80% positive at 1 year to 65% at 3 years (Ho et al., 2015). On the VisQoL questionnaire (Misajon et al., 2005; Peacock et al., 2008), no significant difference was found for the mean utility scores over 24 months; for patients who indicated blindness was having a substantially negative QoL impact in specific domains (9 to 20 of 30 patients), scores related to injury avoidance, coping with life demands, and role fulfillment improved significantly, but those related to independence and joining in daily activities did not (Duncan et al., 2017). Direct reports of device use further reveal marked heterogeneity. Across psychophysical and rehabilitation-focused studies (Yücel et al., 2022; Saltzmann and Stiles, 2025; Sommerhalder and Pérez Fornos, 2017) and case reports (Güven et al., 2018, 2023; Nuzbrokh et al., 2021), use ranged from once per month to approximately 2–5 h per day at 3–4 days per week. Upper bounds of 8–10 h per day were reported during the first two post-implant years in highly motivated recipients (Güven et al., 2023), often alongside occupational therapy or white cane assistance, but such extended daily wear was uncommon. Even within small cohorts, variability was substantial, with some recipients engaging for multiple hours regularly and others only sporadically. Individual cases describe intensive self-training and measurable gains (Demchinsky et al., 2019; Nadal and Iglesias, 2018), but with use frequently selective, supervised, or limited to high-contrast contexts. Longer-term follow-up places these usage patterns in context. Shirian et al. (2025) reported that none of 10 patients followed 8–9 years post-implant were still using the device, with a mean duration of use prior to cessation of 3.34 ± 2.65 years (range 6 weeks to 7 years). Discontinuation was attributed less to isolated technical failure than to limited visual quality, restricted functional utility, and the sustained cognitive effort required to interpret the perceptual signal. In a separate study, patients with longer-duration implants (above 80 months) consistently reported a declining utility (Saltzmann and Stiles, 2025). Qualitative interviews similarly describe a gap between anticipated and actual benefit (Nadolskis et al., 2024). Erickson-Davis and Korzybska (2021) recount that a researcher reported none of 12 recipients from a prior study were still using the device two years later, and that a rehabilitation specialist suggested that many recipients discontinue use shortly after structured training concludes. Recipients described using the system for orientation, social monitoring, or aesthetic experiences such as fireworks or sporting events, while others found it distracting, slow, or physically uncomfortable and reverted to established non-visual mobility strategies for routine navigation (Erickson-Davis and Korzybska, 2021; Rojahn, 2013; Strickland and Harris, 2022). Across studies, adoption was therefore variable, context-dependent, and often constrained by perceptual effort and fatigue, contributing, over time, to abandonment (Shirian et al., 2025; Erickson-Davis and Korzybska, 2021; Stiles et al., 2021), and Stage 5 is not supported. Framework summary. The Argus II literature shows a consistent pattern across study types and follow-up durations (Table 3). Recipients reliably experienced electrically evoked percepts and learned to use them for basic localization and discrimination tasks, demonstrating Device Activation and Perceptual Calibration (Stages 1–2). Many also learned to interpret simple forms and letters under structured conditions, indicating Cognitive Integration (Stage 3). However, transfer of these abilities into independent navigation and everyday activities was inconsistent across subjects and environments, and sustained autonomous reliance on the device was uncommon despite training and prolonged implantation. Within the proposed framework, Argus II therefore supports progression through Stages 1–3, with limited advancement into Stage 4 and infrequent attainment of Stage 5. Table 3Argus II tasks and framework staging.This table combines psychophysical, functional, questionnaire, and qualitative outcomes reported for Argus II and maps them onto framework stages and support. Across that evidence, Argus II supports progression through S1-S3, with selected S4-related gains but no broad support for safety-grade behavioral generalization or sustained voluntary adoption. FLORA improvement and aided-task performance should be interpreted separately from long-term everyday retention.TaskMetricOutcomeCitationStageDo resultssupport stageSquare localizationMean localization error; % subjects significantly better ON vs OFF21 subjects, 81% significantly better on vs off at year 5; per-subject test criterion p < 0.05 da Cruz et al. (2016) S2YesDirection of motion of moving barMean angular error; % subjects significantly better ON vs OFF20 subjects, 50.0% significantly better on vs off at year 5; per-subject test criterion p < 0.05 da Cruz et al. (2016) S2Partial (Yes through Y3; Marginal at Y5)Grating visual acuity/grating orientationlogMAR; % scoring ≤2.9 logMAR38% scored ≤2.9 logMAR at year 5 with system on (33–48%), 0% off; best recorded acuity 1.8 logMAR da Cruz et al. (2016) S2PartialLetter identification, closed-set% correctGroup A letters: 72.3±24.6% on vs 17.7± 12.9% off, p < 0.001; Group B: 55.0±27.4% vs 11.8±10.7%, p < 0.001; Group C: 51.7±28.9% vs 15.3±7.4%, p < 0.001 da Cruz et al. (2013) S3YesLetter-size reductionSmallest letter size identifiedBest MLS 0.9 cm (1.7 deg) at 30 cm; 6 subjects completed test da Cruz et al. (2013) S3YesWord reading, open-choiceNumber correct/10 words4/21 subj selected for good prior performance read unrehearsed words with standard mode on: 2-letter 5–10/10, 3-letter 5–9/10, 4-letter 3–10/10; off 0–2/10; no p value reported da Cruz et al. (2013) S3NoReach-and-grasp of a high-contrast cuboid objectSuccessful grasps (%, median, interquartile range), Visit 1, 2V1: off, 0 [0–50]%; on, 69 [67–95]%; Scr = 59 [42–95]%; p = 0.04.V2: Off, 0 [0–25]%; On, 69 [53–100]%; Scr = 28 [13–63]%; p = 0.02.Off v On: V1, p = 0.046; V2, p = 0.042Scr v On: V1, p = 0.5; V2: p = 0.08 Kotecha et al. (2014) S2/S3PartialReach-and-grasp of a high-contrast cuboid object, w/finger LEDSuccessful grasp (%)71.3±27.1%, prosthesis on, finger marker on77.5±24.5%, prosthesis on, finger marker off0.0% both off Luo et al. (2015) S3YesPath tracing/fine hand-movement guidance, right-angle pathsTracing error; trace time; binomial accuracyError reduced 63%, time increased 156%, both p < 0.001; correct-orientation choice 95% with system ON, p < 0.001 Barry and Dagnelie (2012) S3YesPath tracing/fine hand-movement guidance, mixed-angle single-turnTracing error; trace timeError reduced 53%, time increased 184%, both p < 0.001 Barry and Dagnelie (2012) S3YesPath tracing/fine hand-movement guidance, two-turn pathsTracing error; trace timeError reduced 38%, time increased 252%, both p < 0.001 Barry and Dagnelie (2012) S3PartialShape matching (same/different, 2AFC)Percent correct (mean, std, n = 11)Visual-visual 65±16%, p = 0.012; Visual-tactile 57±23%, p = 0.3 Saltzmann and Stiles (2025) S3PartialDoor finding% successYear 3: 54.2±6.2% on vs 19.0±4.3% off Ho et al. (2015) S4No (below 99%)Line following% successYear 3: 67.9±6.5% ON vs 14.3±3.8% OFF Ho et al. (2015) S4No (below 99%)Navigation: path reproduction, triangle completionPrecision and accuracy at endpoint, improvementPR: prec. 0/4 sig.; acc. 0/4 sig.TC: prec. 2/4 sig.; acc. 0/4 sig. Garcia et al. (2015) S4NoPerson-finding with thermal camera% successful trials99% with thermal filter vs 55% unfiltered, p < 0.01 Sadeghi et al. (2024) S4Only with thermal imagingObstacle detection/avoidance with distance-filter adjunct% obstacles detected/avoided88% with distance filter vs 10% unfiltered, p < 0.01 Sadeghi et al. (2024) S4No (below 99%)FLORA observer-rated functional task batteryObserver-rated task improvement (lower is better) w/device on; Wilcoxon p values24/35 tasks improved with system on; domain changes: Orientation −1.36, p < 0.0001; Mobility −0.82, p = 0.0005; Daily life −0.58, p < 0.0001; Interaction −0.79, p < 0.0001; all moderate to difficult Geruschat et al. (2016) S4Yes; mobility-related difficultFLORA observer-rated functional task batteryObserver-rated task improvement (lower is better) w/device on; Wilcoxon p value16/35 (46%) tasks significantly (p < 0.05) improved; all moderate to difficult Delyfer et al. (2021) S4No; mobility-related difficultFLORA self-reported impactPercent Positive or Mild-Positive assessmentYear 1: 80%; Year 3: 65% Ho et al. (2015) S5Yes, but decliningVisQoL quality-of-life questionnaireDomain-specific mean survey score, 12–36 mo; composite utility (CU) scoreInjury, n = 15, 3.8–2.8, p = 0.036; Life, n = 20, 4.4–3.7, p = 0.007; Roles, n = 16, 4.6–3.8, p = 0.001; Assistance, n.s.; Activities, n.s.; CU, n.s. Duncan et al. (2017) S5PartialLong-term self-evaluation of home-useFrequencyn = 11: 5 daily, 2 several times per week, 2 weekly, 2 monthly Saltzmann and Stiles (2025) S5No Argus II tasks and framework staging. This table combines psychophysical, functional, questionnaire, and qualitative outcomes reported for Argus II and maps them onto framework stages and support. Across that evidence, Argus II supports progression through S1-S3, with selected S4-related gains but no broad support for safety-grade behavioral generalization or sustained voluntary adoption. FLORA improvement and aided-task performance should be interpreted separately from long-term everyday retention. Alpha IMS, AMS The Alpha IMS and its successor, the Alpha AMS, were subretinal prostheses, developed by Retina Implant AG (Zrenner et al., 2011; Table 4). The devices used an approach in which photodiode-based stimulation elements were implanted beneath the retina to target bipolar cells, operating as man-made photoreceptors to functionally replace diseased natural ones (Stingl et al., 2015). The Alpha IMS incorporated a 1500-electrode array with integrated photodetection, requiring external power and control (Stingl et al., 2013), and relying on surviving retinal circuitry to convey stimulation patterns to the visual brain, but not requiring a separate camera. The Alpha AMS addressed stability and longevity concerns by replacing the IMS's analog amplifiers with digital control, and by incorporating more robust materials (Stingl et al., 2017). Although Retina Implant AG (Reutlingen, Germany) ceased operations in 2019, reports have been published evaluating device efficiency and offering important information on functional outcomes that allow us to understand how well the device was integrated into the lives of recipients. Table 4Alpha AMS task results and framework staging.Studies of Alpha IMS and Alpha AMS are summarized here in relation to the adaptation framework, drawing on psychophysical, functional, questionnaire, and qualitative findings. The evidence is strongest for S1-S2, with limited S3 and partial S4-related support, and the more convincing later results come from Alpha AMS rather than the earlier IMS reports. Single-subject retraining results (Kapetanovic et al., 2020) are included for context but should not determine device-level stage assignment.TaskMetricOutcomeCitationStageDo resultssupport stageLight detection (BaLM flash/no-flash)2-AFC pass/faildevice on: 5/5 functional implants passeddevice off: 0/6 passed Edwards et al. (2018) S1YesTemporal resolution (BaLM 1 vs 2 flashes)2-AFC pass/failon: 5/5 functional implants passedoff: 0/6 passed Edwards et al. (2018) S1YesLight localization (BaLM quadrant localization)4-AFC pass/failon: 5/6 passedoff: 0/6 passed Edwards et al. (2018) S1YesMotion direction (BaLM moving dot pattern)4-AFC pass/failon: 0/6 passedoff: 0/6 passed Edwards et al. (2018) S2NoGrating orientation discrimination (BaGA)Highest spatial frequency correctly detectedon: max 3.3 cpd in 1 participant once; others, 1.0, 0.33, 0.33, 0.1 cpdoff: at chance Edwards et al. (2018) S2YesGrayscale contrast discriminationCorrect responses across all contrast pairson: 61%, 72%, 28%, 65%, 70%off: 23%, 0%, 0%, 0%, 0% Edwards et al. (2018) S2YesTabletop geometric-shape recognitionNo. correctly named (max 4)on: 2/6 correctly named 3–4 shapesoff: 0/6 Edwards et al. (2018) S3Yes, limitedTableware-item recognitionNo. correctly named (max 4)on: 2/6 correctly named 3–4 itemsoff: 0/6 Edwards et al. (2018) S3Yes, limitedClock-face recognitionNo. correct out of 12 at months 3, 6, 9 (medians)P1: 12 on vs 2 offothers: 3.5 vs 1.0, 0 vs 0, 0 vs 0, 0.5 vs 0, 5 vs 1 Edwards et al. (2018) S3Yes, limitedReading short words/letter recognitionVisual acuity0.04 decimal, 1.39 logMAR, 20/500 after intensive training (study of 1 subject) Kapetanovic et al. (2020) S3Yes, limitedComplex-shape recognition (e.g., restroom symbols)Visual acuity0.0048 decimal, 2.3 logMAR, 20/4000 (study of 1 subject) Kapetanovic et al. (2020) S3Yes, limitedHand-eye coordination with tabletop shapesPlacement errorRequired correction by 3–4 cm (study of 1 subject) Kapetanovic et al. (2020) S3Yes, limitedSelf-assessed mobility in familiar/unfamiliar areas (uncontrolled test)5-point Likert scale of difficulty (extreme to none)3/6 improved by > 3 levels for walking in familiar areas1 participant improved from extreme to none in unfamiliar areas, and from moderate to none in familiar areas Edwards et al. (2018) S4Partial Alpha AMS task results and framework staging. Studies of Alpha IMS and Alpha AMS are summarized here in relation to the adaptation framework, drawing on psychophysical, functional, questionnaire, and qualitative findings. The evidence is strongest for S1-S2, with limited S3 and partial S4-related support, and the more convincing later results come from Alpha AMS rather than the earlier IMS reports. Single-subject retraining results (Kapetanovic et al., 2020) are included for context but should not determine device-level stage assignment. Early reports with the Alpha IMS had positive outcomes. In a pilot study, recipients reported that with 4–6 h of daily practice over several weeks they could control eye position, improve visuomotor coordination, and perform basic tasks such as pointing to objects and recognizing simple shapes (Zrenner et al., 2011; Stingl et al., 2013). 82% (9/11) of subjects reported light perception mediated by the implant, and 64% (7/11) were able to localize a light source significantly above chance. Five subjects could accurately report the motion of dot patterns up to 35° /sec. Grating acuity was measured in six subjects with a maximum of 3.3 cpd. Geometric shapes and tableware objects presented in high-contrast settings were counted, identified, and located all significantly better with the device on than off across the population (each test, p < 0.05). Subjective feedback from recipients included acclimatization to the flickering percepts presented by the implants and successful steering of gaze and attention (Zrenner et al., 2011), as well as the ability to recognize objects and structures in non-laboratory interior and exterior settings (Stingl et al., 2013). These results are consistent with success into S3, with anecdotal support for S4. However, longitudinal studies indicated that these gains were not reliably sustained. In a cohort of 29 patients followed for 1–6 months, visual function as measured by BaLM tests showed inconsistent outcomes. Although 50–55% of participants initially performed better with the device on versus off, scores fell back to 40% by month three, with no significant difference by month six (Stingl et al., 2015). Tasks such as grayscale perception and object recognition initially improved but reverted to system-off levels after several months, although there was often an increase in system-off performance due to a hypothesized up-regulation of growth factors (Schatz et al., 2011) that are also seen in other systems (Chow et al., 2004; Pardue et al., 2005). The maximum visual acuity (20/546 Snellen, 1.44 logMAR) was one of the highest scores reported among clinically tested devices, but the BaLM grating acuity task showed inconsistent results. The percentage of patients successfully passing the different tests varied considerably through the six months of study, with no significant improvement found for the system on versus off at the six-month mark due in part to improved system-off scores (Stingl et al., 2015), but overall results suggest difficulty in consolidating perception for Behavioral Generalization (S4). Of the 29 participants, 25 reached primary endpoints related to mobility and activities of daily living, and 21 reached secondary goals including improvements in light perception (S1) and visual acuity (S3), yet only 13 reported vision adequate for everyday use (S4) (Stingl et al., 2015). Technical constraints, including underestimated spreading of electrical fields from the 1500 photodiodes and disease-driven retinal remodeling, may have reduced the effective resolution, affecting the capability for fluent perception (Shivdasani et al., 2014). The revised architecture, Alpha AMS, introduced refinements aimed at greater stability and durability. In a one-year study (Edwards et al., 2018), five of six subjects passed BaLM light perception and localization tests (S1, S2); the sixth subject's implant was damaged during implantation, rendering it non-functional. The same five were significantly better at a six-level visual contrast test with the device on than off (across contrasts, 62 ± 13% versus 5.7 ± 5.1%, mean and SD, p < 0.001, paired t-test; S1), and could count high-contrast tabletop objects better with the device on versus off (88 ± 16% versus 7.5 ± 17%, mean and SD, p < 0.001, paired t-test; S2). All five passed the BaLM grating acuity test at 0.1 cpd (S2), and two at 0.33 cpd. Two subjects achieved object identification by month three (S3, but not qualifying), and three subjects were significantly better at reading the time from an analog clock (S3). None passed BaLM motion detection tests. No quantitative tests were performed that would be considered for S4. The Turano-based Independent Mobility Questionnaire evaluating the difficulty of performing various activities showed substantial improvement in confidence with walking in familiar areas and avoiding obstacles (Edwards et al., 2018). Subjective experiences included the ability to locate balls on a billiard table, see windows and doors, kitchen appliances, high-contrast laundry, buildings, a mountain against the sun, a car, and Christmas lights. Despite none of the subjects achieving independent mobility outside their home, the broad-based questionnaire improvements indicate a high level of adaptation and satisfaction, suggesting Psychosocial Adjustment (Stage 5) is possible. In an attempt to improve outcomes from these earlier patients, an intensive rehabilitation plan was explored for one patient three years after implantation (Kapetanovic et al., 2020). This patient had reported a benefit at 12-months, but their motivation had subsequently dropped, finding that the use of the device became an impediment, especially in familiar environments. The second round of training included screen-reading, complex shape recognition, eye-hand coordination, path following, obstacle recognition and outdoor distance estimation. Daily 2-h sessions for five days allowed the subject to successfully recognize small words and shapes (1.39 and 2.2 logMAR respectively), to identify obstacles and estimate distances in daytime and nighttime conditions, and had re-established eagerness for the implant (Kapetanovic et al., 2020). These qualitative results suggest that the patient had progressed into S4, and thus that post-implantation training has an important role to sustain or restore user engagement during plateaus in perceived benefit. The Dobelle prosthesis The first camera-based visual prosthesis implanted in a human subject was the Dobelle cortical prosthesis (Dobelle et al., 1979; see also Brindley and Lewin, 1968 for seminal proof-of-concept experiments, and Tassicker, 1956 with a design that anticipated contemporary direct imaging approaches such as currently used by PRIMA). The system used a glasses-mounted camera whose output was converted into electrical stimulation that was delivered to an array of surface electrodes on the occipital cortex. Many of the scientific publications on the Dobelle prosthesis are either qualitative, or case reports, although there is one retrospective study that will be included below. Early reports described that recipients perceived multiple phosphenes (Dobelle et al., 1974) and could, after training, perform visually guided behaviors such as navigating around obstacles or locating objects within a room (Dobelle, 2000). One participant, known as Jerry, reportedly learned to operate the device within a day and demonstrated task-directed mobility. However, functional use remained limited. Jerry used the system only a few hours per day in a structured routine, and another recipient, Jens Naumann, described restricted use during laboratory sessions and eventual abandonment after persistent frustration and device failure (Naumann, 2012). Even in these early demonstrations, performance depended on deliberate scanning strategies and sustained cognitive effort rather than spontaneous perception. A later retrospective cohort study interviewing thirteen of the sixteen participants from the second phase of the Dobelle program, which used camera-based implants, provides a broader account of recipient experience (Lane et al., 2015). Participants consistently reported perception of phosphenes varying in number, color, and location, confirming reliable elicitation of visual percepts. Some individuals used these percepts for orientation, mobility, and object detection, yet functional ability did not correlate with phosphene count and many participants were unable to organize percepts into meaningful patterns. Functional interpretation therefore varied widely across individuals and across time within the same individual. Reported performance depended heavily on concentration and strategy, and perception often failed to generalize beyond specific trained situations. The same cohort described strong emotional responses including anxiety, frustration, resentment, and a sense of loss when percepts disappeared or failed to meet expectations (Lane et al., 2015). This idea of fear of experiencing loss of vision for a second time was echoed by potential prosthesis recipients in a different study (Karadima et al., 2023). Dobelle participants frequently reported a mismatch between anticipated visual restoration and the actual performance. Engagement was often motivated by hope or participation in pioneering research rather than sustained functional benefit. These reports parallel earlier individual accounts describing enthusiasm after implantation followed by reduced use when everyday benefit remained limited (Dobelle, 2000; Naumann, 2012). These qualitative reports suggest that the Dobelle system reliably generated percepts and occasionally supported task-directed visual behaviors, but interpretation of the percepts remained inconsistent and did not generalize into independent everyday reliance. Percept generation and intermittent goal-directed use were demonstrated without evidence of stable behavioral generalization or durable adoption. In combination, these observations suggest an assignment of S3 that is tentative due to lack of reported methodological detail (Table 5). Table 5Dobelle cortical prosthesis tasks and framework staging.Reported Dobelle outcomes span psychophysical, functional, and qualitative domains and are summarized here with respect to framework staging. Taken together, the literature supports S1-S3, including reproducible percept generation and some goal-directed visual use. Because much of the record is qualitative, case-based, or retrospective, the table does not support strong claims about broad behavioral generalization or durable everyday adoption.TaskMetricOutcomeCitationStageDo resultssupport stagePhosphene detection thresholdamplitude threshold46 tested sites: 0.6–2.8 mA, mean 1.3 mA32 tested sites: 2.1–8.1 mA, mean 4.5 mA Dobelle et al. (1974) S1YesTwo-point phosphene discriminationspatial resolution threshold (cortical surface)Resolution breaks down between 2 and 3 mm Dobelle et al. (1974) S2YesCortical braille letter identificationcorrect responses19/26 correct, 73%, self-paced; 85% correct, fixed-time 0.5 s presentation Dobelle et al. (1976) S3YesCortical braille sentence readingcharacter rate; completion time30 characters/min; example sentences completed in 54 s and 52 s Dobelle et al. (1976) S3YesLetter recognition (Snellen, HOTV, Lea figures)visual acuity20/1200 Dobelle (2000) S3YesTumbling-E orientation discriminationvisual acuity20/1200 Dobelle (2000) S3YesLandolt-ring discriminationvisual acuity20/1200 Dobelle (2000) S3Yes Dobelle cortical prosthesis tasks and framework staging. Reported Dobelle outcomes span psychophysical, functional, and qualitative domains and are summarized here with respect to framework staging. Taken together, the literature supports S1-S3, including reproducible percept generation and some goal-directed visual use. Because much of the record is qualitative, case-based, or retrospective, the table does not support strong claims about broad behavioral generalization or durable everyday adoption. Devices currently under development Suprachoroidal visual prosthesis (BVT/Phoenix 99) The BVT suprachoroidal visual prosthetic device (Ayton et al., 2014) uses an electrode array placed in the suprachoroidal space between the sclera and choroid of the eye. It functions much like an epiretinal prosthesis in that the imaging camera is decoupled from the array of stimulating electrodes, and, like the Argus II, lacks gaze contingency, so subjects must be trained to hold their eyes fixed in the orbit and scan with head movements alone. Studies assessing psychophysical performance, mobility and daily living outcomes with respect to device usage suggest a lack of integration into normal activities (Petoe et al., 2021, 2024; Karapanos et al., 2021; Ayton et al., 2014). In a pilot clinical trial of the first-generation, 33-electrode implant (Ayton et al., 2014), three patients were followed for 12 months and evaluated with the BaLM (Bach et al., 2010) and Landolt C module of FrACT tests (Bach, 1996). Reliable phosphenes were elicited from all three subjects (Sinclair et al., 2016). Although participants generally benefited from the use of the system, the improvements were modest: BaLM light localization with the device on was significantly above chance and above performance with the device off (P1, 98% vs 28%; P2, 71% vs 25%; P3, 67% vs 25%), yet the mean Landolt C visual acuity (P1, logMAR 2.62, Snellen 20/8397) was below theoretical prediction for the electrode array (logMAR 2.33, Snellen 20/4242) and identifying optotype orientation in the task required approximately 1 min of time (Ayton et al., 2014). Subsequent studies with the second-generation, 44-electrode BVT device assessed performance on quantitative laboratory tasks as well as quality-of-life instruments (Petoe et al., 2021, 2024; Karapanos et al., 2021; Titchener et al., 2020). Screen-based square localization tasks showed significantly better performance with the device on versus off for all subjects (p < 0.001; 9.5–11.4° with the device on versus 22.0–36.0° with the device off, and 13.8–15.7° with the electrode map scrambled). Two of four subjects performed a 4-way motion discrimination task significantly better with the device on versus off at all tested speeds, up to 30° per second. No subject was able to consistently meet the passing criterion in a grating detection task. All four subjects were significantly better at locating high-contrast objects on a table (p < 0.01), but only two were significantly better at identifying those objects (both p < 0.05). These test results show stage S1 for the population, and stage S2 for two of the four individuals. Three of four subjects were significantly better at a laboratory-based door location task with the device on than off (p < 0.001), but none approached the safety-related S4 thresholds. The same three were significantly better at a grasping task that required visuo-motor coordination (p < 0.001), but all had significantly higher accuracy for unsuccessful misses (p < 0.01). Gradual improvements on the FLORA instrument were seen. Participants described that finding doorways and locating familiar objects became progressively easier with the device on, shifting from difficult to moderate, while tasks remained mostly difficult with the system off (Petoe et al., 2021). By the study's end, mobility was moderate on average (Petoe et al., 2024). Participants reported experiences such as detecting obstacles, sorting objects, estimating distances, and counting people, with most of the quantitative improvements during the period 4–10 months after implantation (Petoe et al., 2021; Karapanos et al., 2021). These reports suggest intermittent progression into Cognitive Integration (Stage 3) and attempts at Behavioral Generalization (Stage 4). Device use outside the lab, however, was limited. According to the device logs, the frequency of use was every two to three days for 2.3 h each time for two subjects (Petoe et al., 2021), later decreasing to once every two weeks, especially when COVID-affected periods eliminated social life and mobility outside (Petoe et al., 2024), and for six 30-min occasions in total for one subject (Petoe et al., 2021). Recipients described that after initial exploration to determine where and when their device would be useful, they used it in specific circumstances, often only for 2–3 h at a time (Petoe et al., 2024). These usage patterns of brief engagement indicate that most patients did not advance to generalized use, and thus evidence does not yet exist for S4 or S5. A third-generation, 99-electrode device is under active development as the Phoenix 99 (Eggenberger et al., 2021). With publications from clinical testing of the new design, or extended long-term testing of the 44-contact version, the stage classification (Table 6) may change. Table 6BVT suprachoroidal prosthesis tasks and framework staging.This table compiles psychophysical, functional, and qualitative evidence reported for the first-generation 33-channel BVT, and the second-generation 44-channel BVT device to show adaptation progress across stages. The results support S1-S3, but mobility-related and real-world tasks fall short of the stricter criterion for S4, and at-home use does not support S5. Laboratory improvement therefore should not be taken as evidence of reliable independent daily use.TaskMetricOutcome (Best/Representative quantitative resultsfrom latest report; P = subject)CitationStageDo resultssupport stageBaLM light localization% correct, 4AFCP1–P3: 98%, 71%, 67% correct with device on all p < 0.0001 vs chance Ayton et al. (2014) S1YesLandolt C optotype recognitionlogMARon: mean 2.62 logMAR, range 2.35–3.02off: below software floor 3.24 logMAR Ayton et al. (2014) S2YesSquare localizationPointing error (deg)on, 9.5–11 deg; off, 22–36 deg; all p < 0.001 scrambled, 13.8–15.7 deg Petoe et al. (2024) S2YesMotion discrimination% correct, 4AFC, chance 25%7 deg/s: on, P1 67%, P2 79%, P3 42% vs off, 21%, 25%, 29%; all p < 0.00230 deg/s: P1 79%; P2 79%; P3 single study-end attempt 54% Petoe et al. (2024) S2YesSpatial discrimination (Basic Grating Acuity)% correct, 2AFC, pass criterion 75%0.033 cpd: P1, 63–75%; P2, 75–100%, P3, 29–75%0.01 cpd: P3, 50–75% at study end Petoe et al. (2024) S2PartialTabletop searchObject localization %, object ID %, touch success, fingertip-to-object distancelocalization: on, 60–90%, all p < 0.001ID improved: P1, P2 only, ≤45%F/T: on, 3.7–10 cm; off, 20–37 cm; all p < 0.001 Petoe et al. (2024) S3YesModified door taskTouch success, fingertip-to-target distance, completion timeF/T: on, 0.3–17 cm; off, 62–107 cm; all p < 0.001Touch success better w/on: p < 0.001 for P1, P2, P4; p = 0.02 for P3 Petoe et al. (2024) S4No, not near 99% reliabilityObstacle avoidanceDetection, collision rate, zero-collision trials, walking speeddetection better on: all p < 0.001fewer contacts/collisions: P1, P2, p < 0.001; P3, p = 0.008; P4, p = 0.001zero-collision trials improved w/on P1–P3 slower w/on, all p < 0.001 Petoe et al. (2024) S4No, not near 99% reliabilityAt-home device useFrequency, durationP1–P3: every 2–3 days initially, later about once per fortnight; typical use 2–3 h/sessionP4: rare use, <30 min/session Petoe et al. (2024) S5No BVT suprachoroidal prosthesis tasks and framework staging. This table compiles psychophysical, functional, and qualitative evidence reported for the first-generation 33-channel BVT, and the second-generation 44-channel BVT device to show adaptation progress across stages. The results support S1-S3, but mobility-related and real-world tasks fall short of the stricter criterion for S4, and at-home use does not support S5. Laboratory improvement therefore should not be taken as evidence of reliable independent daily use. IMIE 256 epiretinal array In five subjects with end-stage retinitis pigmentosa, the 256-channel IMIE epiretinal prosthesis produced reliable phosphene perception and measurable task performance over a 90-day rehabilitation program (Xu et al., 2021). Training consisted of repeated high-contrast laboratory exercises paired with structured mobility activities, while acuity tasks were held back from the training regimen to limit practice effects. With the system on, all subjects achieved the single tested grating acuity (20/800) and Tumbling-E acuity (20/1200), detected motion direction, localized targets, and performed orientation-and-mobility tasks such as door finding and line following with high accuracy, whereas device-off performance was absent or near chance. The limited numbers of trials dilutes on-vs-off statistical power, making the comparisons not significant despite the positive reported performances. Because testing focused on these practiced tasks and did not vary aspects like stimulus difficulty or contextual setting, it remains unclear whether performance reflected general visual capability or learned task routines. Subjects had unlimited time on each task, with typical response times in the 1–3 min range at the end of training. These results suggest that intensive, structured rehabilitation enabled recipients to move through Device Activation and Perceptual Calibration (Stages 1 and 2) and into Cognitive Integration (Stage 3), with some evidence of early Behavioral Generalization (Stage 4) in simulated daily-living activities (Table 7). Although the subject pool was small (n = 5), the consistency of improvement underscores the role of training in facilitating adaptation. Follow-up, long-term study remains necessary to determine if these gains can be sustained and expanded into complex real-world contexts, translating ultimately into Psychosocial Adjustment (Stage 5). Table 7IMIE 256 task and framework staging.This table summarizes single-subject psychophysical and functional evidence from IMIE 256 and relates each task to stage support. The reported findings support S1-S2 and suggest provisional S3-S4 performance after intensive rehabilitation. For the Direction of motion, Square localization, and Tumbling E tasks, all subjects were significantly above chance for the device-on condition at the 0.05 level, and all subjects were not significantly above chance for the device-off condition. The limited number of trials in the device-off condition, however, weakened statistical power, rendering the comparisons between device-on and device-off conditions to be not significant (p typically 0.2 or higher), despite high device-on success rates. The Line and Door tasks are considered provisional despite similar high success rates because of the low trial counts and lack of statistical assessment.TaskMetricOutcome (P1 through P5)CitationStageDo resultssupport stagePhosphene detectionMean stimulation threshold on example electrodesP2, 183–256 μA; P3, 44–59 μA; P4, 202–392 μA; P5, 150–242 μA; P1 not reported Xu et al. (2021) S1YesGrating visual acuity4AFC correct/total-trials at 0.025 (20/800, 1.60 logMAR); mean completion timeon: 1/2, 4/5, 2/2, 4/4, 3/3off: 0/1, 0/2, 0/2, 0/1, 0/1time w/on, 28 s, 27 s, 32 s, 18 s, 22 stime w/off, 9 s, 10 s, 8 s, 10 s, 12 s Xu et al. (2021) S2YesDirection of motion8AFC correct/total-trials; mean completion timeon: 3/4, 4/4, 3/3, 4/4, 4/4off: 0/1 for alltime on: 72 s, 100 s, 95 s, 70 s, 56 s Xu et al. (2021) S2Provisional (limited trial count)Square localizationcorrect/total-trials; mean completion timeon: 4/5, 4/4, 3/3, 4/4, 2/2off: 0/1, 0/1, 0/1, 0/1, 0/2; chance 2.6%time w/on: 45, 100, 207, 99, 103time w/off: 15, 20, 39, 22, 17 Xu et al. (2021) S2Provisional (limited trial count)Tumbling E visual acuity4AFC correct/total-trials at 0.017 (20/1200, 1.78 logMAR); mean completion timeon: 3/4, 3/5, 4/4, 2/3, 4/4off: 0/1 for alltime w/on: 95 s, 98 s, 75 s, 120 s, 69 stime w/off: 28 s, 12 s, 14 s, 20 s, 24 s Xu et al. (2021) S3Provisional (limited trial count)Line taskcorrect/total-trials; mean completion timeon: 2/2, 4/4, 3/3, 2/2off: 0/1, 0/1, 0/1, 0/1time w/on: 82 s, 75 s, 72 s, 66 stime w/off: 16 s, 22 s, 29 s, 27 sP2 data not available Xu et al. (2021) S4Provisional (no statistical assessment)Door taskcorrect/total-trials; mean completion timeon: 2/2, 2/2, 1/1, 1/1off: 0/1, 0/1, 0/1, 0/1time w/on: 90 s, 130 s, 188 s, 105 stime w/off: 20 s, 32 s, 24 s, 20 sP2 data not available Xu et al. (2021) S4Provisional (no statistical assessment) IMIE 256 task and framework staging. This table summarizes single-subject psychophysical and functional evidence from IMIE 256 and relates each task to stage support. The reported findings support S1-S2 and suggest provisional S3-S4 performance after intensive rehabilitation. For the Direction of motion, Square localization, and Tumbling E tasks, all subjects were significantly above chance for the device-on condition at the 0.05 level, and all subjects were not significantly above chance for the device-off condition. The limited number of trials in the device-off condition, however, weakened statistical power, rendering the comparisons between device-on and device-off conditions to be not significant (p typically 0.2 or higher), despite high device-on success rates. The Line and Door tasks are considered provisional despite similar high success rates because of the low trial counts and lack of statistical assessment. CORTIVIS cortical prosthesis The CORTIVIS cortical prosthesis uses a Utah intracortical microelectrode array implanted in primary visual cortex to deliver visual information. Initial published results described one participant followed over approximately six months (Fernández et al., 2021), and subsequent reports from the same clinical program indicate additional implanted participants, with at least four participants tested by 2026 (Grani et al., 2025; Lozano et al., 2026; Alfaro et al., 2026). During an initial two-month period of daily 10–15 min training sessions, the first participant learned to reliably localize electrically evoked phosphenes and distinguish them from spontaneous percepts, establishing Device Activation and enabling Perceptual Calibration (S1–S2). Over the subsequent four months, multi-electrode and sequential stimulation patterns were introduced, allowing discrimination of oriented forms, recognition of object boundaries, and identification of specific letters, including distinctions between lower- and upper-case presentations across repeated trials, indicating continued progression into Cognitive Integration (S3). Later camera-driven stimulation produced recognizable silhouettes and shapes, but performance depended on deliberate interpretation under supervised testing rather than spontaneous visual experience. A subsequent CORTIVIS case study further showed that perceived phosphene location remained gaze-dependent, and that video-based tracking of a cosmetic eye could support a gaze-contingent screen-search task in an enucleated participant, adding to the importance of sensorimotor coupling and control design in visual prostheses (Waclawczyk et al., 2025). Within the framework, these results demonstrate reliable percept generation and spatial mapping followed by learned pattern interpretation, while leaving open whether behaviors would generalize beyond guided experimental contexts (Table 8). No independent everyday use or autonomous navigation was reported, so Behavioral Generalization (Stage 4) was not established and Psychosocial Adjustment (S5) cannot yet be assessed. Subsequent methodological work from the same program has also reported recordings from three blind human volunteers with implanted Utah arrays, indicating that the patient cohort has expanded, although the scope of the report did not include stage-defining behavioral outcomes (Lozano et al., 2026). Table 8CORTIVIS tasks and framework staging.The table compiles single-subject and small-cohort evidence from psychophysical and constrained functional testing to indicate stage support for the CORTIVIS cortical prosthesis. Published results support Device Activation and Perceptual Calibration strongly, and Cognitive Integration more narrowly through trained discrimination and gaze-contingent task performance. The newer CORTIVIS reports broaden support for S1-S2 and add limited additional S3 evidence, but the evidence remains highly structured, involves very few implanted participants, and does not yet establish Behavioral Generalization or Psychosocial Adjustment.TaskMetricOutcomeCitationStageDo resultssupport stageEvoked vs sham phosphene detectionbutton-press detection; correct detections, false positives947/1000 stim detected, 23/500 sham false positives Fernández et al. (2021) S1YesPhosphene discriminationpercent correct>95% correct Fernández et al. (2021) S1YesSingle-electrode phosphene thresholdthreshold current66.8 ± 36.5 μA, range 7–128 μA; 88/96 electrodes evoked phosphenes Fernández et al. (2021) S1YesMulti-electrode phosphene thresholdthreshold current4-electrode threshold 48.0 ± 22.6 μA vs 66.8 ± 36.5 μA single-electrode, 28% lower, p < 0.001 Fernández et al. (2021) S1YesBehavioral phosphene threshold from psychometric curvescharge per phase at 50% detectionS1 3.45 ± 1.27 nCS2 2.42 ± 0.99 nC Grani et al. (2025) S1YesNeural prediction of phosphene thresholdcorrelation between behavioral and neural thresholdS1 R = 0.71 (n = 49, p < 0.001) S2 R = 0.71 (n = 28, p < 0.001) Grani et al. (2025) S1YesPhosphene location mapping across sessionspositional variation18 mapping sessions; location scatter <0.5 deg between sessions Fernández et al. (2021) S2YesBrightness/size vs number of stimulated electrodessubjective brightness and sizeboth increased with electrode number, p < 0.001 Fernández et al. (2021) S2YesSimultaneous 2-electrode two-point discriminationtrials with 2 distinct percepts24% two discrete phosphenes, 76% single phosphene; merged percept size 1.9 ± 0.1 vs 0.8 ± 0.1, p < 0.001 Fernández et al. (2021) S2YesDelayed 2-electrode discriminationdistinct-phosphene reports>250 ms delay at 400 μm separation: 90%; 565 μm or greater: 100% Fernández et al. (2021) S2YesHorizontal vs vertical pattern discrimination2AFC, percent correct81.4% overall across 22 sessions/4 d; later sessions 100% Fernández et al. (2021) S2YesLarge white-square left/right localization2AFC, percent correct100% across 70 trials Fernández et al. (2021) S2YesSmall white-square 4-location pointingaccuracy; reaction time100% across 150 trials spanning 5 days; RT 8.6 ± 0.6 s to 5.4 ± 0.3 s Fernández et al. (2021) S2YesBrightness scalingneural-brightness correlationneural-brightness correlation R = 0.871, p < 0.001 Grani et al. (2025) S2YesSequential phosphene separationintertrain delay threshold for perceiving two distinct phosphenesS1 0.30 ± 0.07 sS2 0.33 ± 0.04 s Grani et al. (2025) S2YesNeural prediction of separate-phosphene thresholdcorrelation between behavioral and neural thresholdS1 R = 0.74 (n = 11, p < 0.01)S2 R = 0.72 (n = 12, p < 0.01) Grani et al. (2025) S2YesLetter discrimination (i/L, i/O, L/C, C/V, L/V)2AFC, percent correct>70% overall Fernández et al. (2021) S3YesUpper/lowercase O vs o learning task2AFC, percent correct90–95% after 70 trials; 100% in latest sessions; 20 sessions, 400 trials Fernández et al. (2021) S3YesGaze-contingent screen searchtarget localization accuracy; mean reaction time100%, 4 target locations; sessions 1, 2: 11.9, 10.5 s, n.s. Waclawczyk et al. (2025) S3Partial CORTIVIS tasks and framework staging. The table compiles single-subject and small-cohort evidence from psychophysical and constrained functional testing to indicate stage support for the CORTIVIS cortical prosthesis. Published results support Device Activation and Perceptual Calibration strongly, and Cognitive Integration more narrowly through trained discrimination and gaze-contingent task performance. The newer CORTIVIS reports broaden support for S1-S2 and add limited additional S3 evidence, but the evidence remains highly structured, involves very few implanted participants, and does not yet establish Behavioral Generalization or Psychosocial Adjustment. PRIMA subretinal device The PRIMA subretinal device (Palanker et al., 2020) integrates small photovoltaic sites with stimulating electrodes to create dense pixelated patterns of electrical stimulation in the retina. It has demonstrated sustained functionality over 4 years in feasibility studies (Muqit et al., 2024) and native prosthetic acuity closely matching the implant's theoretical resolution in a pivotal trial (1.32 ± 0.16 logMAR at 12 months; Holz et al., 2025; Table 9 includes additional reports with zoom use). In early feasibility cohorts, all implanted subjects showed prosthetic light perception in the central scotoma and achieved prosthetic acuity closely matching the 100 μm pixel pitch (about logMAR 1.39, Snellen 20/500) without magnification (Palanker et al., 2022). With training, participants recognized letters and sequences of letters and transferred this ability to reading words on packages and signs (Muqit et al., 2024) including at home (Holz et al., 2025). The transparent glasses allowed simultaneous use of central prosthetic vision and residual peripheral natural vision in the same eye (Palanker et al., 2022). In a multicenter trial, most participants demonstrated improved acuity relative to baseline while peripheral natural acuity remained unchanged (Holz et al., 2025). Enhanced performance could be obtained with electronic magnification, but non-zoomed acuity reflected the spatial sampling limit on the retinal surface of the implant. Table 9PRIMA tasks and framework staging.Outcomes are summarized across psychophysical, functional, and questionnaire measures. Using non-zoomed PRIMA results to allow cross-device comparisons, the table supports S1-S2; with zoom, it supports S3 as indicated by an asterisk; no structured task battery establishes S4 and no questionnaire evidence supports S5.TaskMetricOutcomeCitationStageDo resultssupport stageCentral perception (microperimetry stimulation)Sensitivity improvement (dB)mean 23.4 dB; 30/32 participants demonstrated central perception, of whom 24/30 were reliable against false positivesHolz et al. (2025) (appendix)S1YesBar orientation discrimination% correct88–96% correct across subjects Palanker et al. (2020) S2YesMinimum bar-width discriminationMinimum resolvable bar width (pixels)1-pixel bars discriminated at 50–100% correct Palanker et al. (2020) S2YesImplant resolution testvisual acuity1.32 ± 0.16 logMAR (20/417)Holz et al. (2025) (appendix)S2YesLandolt C/ETDRS prosthetic acuityvisual acuity1.39 logMAR (20/500) Muqit et al. (2024) S2YesBest-corrected visual acuity improvement (zoom use allowed)≥0.2 logMAR improvement26/32 participants (81%) improved ≥0.2 logMAR at 12 mo Holz et al. (2025) S3*YesMean visual acuity change (zoom use allowed)logMAR change0.49 logMAR improvement (24.5 letters) Holz et al. (2025) S3*YesImpact of Vision Impairment (IVI) questionnaireRasch-transformed score, overall (baseline, 6 mo, 12 mo)−0.88, −0.83, −0.45; repeated-measures p = 0.35 (n.s.)Holz et al. (2025) (appendix)S5NoIVI emotional subscaleRasch score (baseline, 12 mo)−0.64 to +0.41, p = 0.12 (n.s.)Holz et al. (2025) (appendix)S5NoIVI mobility subscaleRasch score (baseline, 12 mo)−0.62 to −0.43, p = 0.70 (n.s.)Holz et al. (2025) (appendix)S5NoIVI reading subscaleRasch score (baseline, 12 mo)−2.09 to −1.47, p = 0.30 (n.s.)Holz et al. (2025) (appendix)S5No PRIMA tasks and framework staging. Outcomes are summarized across psychophysical, functional, and questionnaire measures. Using non-zoomed PRIMA results to allow cross-device comparisons, the table supports S1-S2; with zoom, it supports S3 as indicated by an asterisk; no structured task battery establishes S4 and no questionnaire evidence supports S5. Within the framework, we will consider only non-zoomed performance in order to have consistent comparisons with other projects (Table 9). PRIMA reliably achieves Device Activation and Perceptual Calibration (S1–S2) through stable central percepts and reproducible spatial form perception. Recognition of letters and words demonstrates Cognitive Integration (S3). Although some transfer to real-world printed material was reported, the absence of statistically significant improvement in quality-of-life measures (Holz et al., 2025) suggests that consistent Behavioral Generalization (S4) and Psychosocial Adjustment (S5) have not yet been established. Orion visual cortical prosthesis The Orion Visual Cortical Prosthesis (Levy et al., 2024) adapts the Argus II architecture for direct cortical stimulation, using surface electrodes implanted over the interhemispheric portion of visual cortex. In an early feasibility trial, four of six participants provided detailed narratives after 2.8 to 3.5 years of device use that was limited to 2 h per day by safety protocol (Levy et al., 2024). All participants reported perceiving phosphenes yet emphasized that the experience fundamentally differed from natural sight, requiring months of training with research staff and low-vision therapists to interpret patterns (Levy et al., 2024; McDonald, 2019). One subject compared it to viewing a radar screen: useful for detecting contrast but lacking depth and detail. Others described the process of interpretation as akin to learning a new language (Levy et al., 2024). Similar descriptions appear in an autobiographical account of Orion use, where the percept is described as sparse flashes that must be interpreted through deliberate scanning and training rather than immediate visual recognition (McDonald, 2019). Dynamic current-steering experiments further showed that sequential stimulation could produce recognizable shapes and traced letters rather than isolated flashes (Beauchamp et al., 2020), and interim multi-subject testing demonstrated improved localization and motion-direction discrimination with the system on compared with off (Dorn, 2020). The absence of gaze contingency likely increased the difficulty of achieving stable Perceptual Calibration (S2), as its absence interferes with normal scene building (Caspi et al., 2018; Paraskevoudi and Pezaris, 2021; de Ruyter van Steveninck et al., 2024). Despite these challenges, some participants reached Cognitive Integration (S3), using Orion for investigator-directed tasks such as sorting laundry by brightness, identifying doorways, or following the contrast between sidewalk and grass. However, functionality did not consistently translate to benefit; one recipient could detect sidewalk edges with the device, but concluded the device's utility as “marginal,” as it did not eliminate reliance on a cane (Levy et al., 2024), an important consideration for device acceptance (Karadima et al., 2023). Across reports, the device enabled certain tasks, but the inability to integrate artificial vision into real-world contexts highlighted the distinction between usability and usefulness (Table 10). Barriers to further adaptation included the cumbersome design of external components, difficulties with charging and operation, and aesthetic concerns, with some describing the headgear as unappealing and drawing unwanted attention (Levy et al., 2024), similar to the concerns expressed by Argus II patients (Erickson-Davis and Korzybska, 2021). Psychosocial Adjustment (S5) was partial: while some derived personal meaning and satisfaction from participation, and requested additional training, others found the benefits too limited to justify daily use (Levy et al., 2024; McDonald, 2019). Table 10Orion tasks and framework staging.This table summarizes psychophysical, functional, and qualitative findings from Orion to indicate the stages currently supported by the literature. The available evidence supports S1-S3 and suggests limited psychosocial benefit in some recipients without establishing broad behavioral generalization or routine everyday adoption. FLORA-related support remains limited because only a general score is reported, without detail on specific tasks or self-assessments.TaskMetricOutcomeCitationStageDo resultssupport stagePhosphene detectionn/N electrodes producing percepts8/8 electrodes Beauchamp et al. (2020) S1YesSingle-electrode phosphene localization (tactile fixation)2-D RMS localization error (deg)P1: 19.0 ± 4.7° to 4.0 ± 2.6° , p < 0.001P2: 12.0 ± 2.7° to 4.1 ± 2.0° , p < 0.001 Oswalt et al. (2021) S1YesMulti-electrode phosphene spatial mapping2-D RMS localization error (deg)P1: 3.9 ± 2.0° to 0.99 ± 0.95° aligned;P2: 3.2 ± 1.1° to 0.84 ± 0.41° , p < 0.001 Oswalt et al. (2021) S2YesBrightness discrimination (same electrode, 2IFC)probability lower amplitude judged brighter; AzExample P2: phase-2 0.62 vs phase-1 0.12; significant comparisons p < 0.05 Armenta-Salas et al. (2022) S2YesBrightness discrimination (different electrodes, 2IFC)reference electrode judged brighter; AzExample P2 low–med: phase-1 0.32 vs phase-2 0.67; multiple comparisons p < 0.05 Armenta-Salas et al. (2022) S2YesThree-electrode brightness orderingcorrelation between perceived brightness and serial positionP2 r = 0.33, p < 0.001P3 r = 0.30, p < 0.001 Armenta-Salas et al. (2022) S2YesSquare localizationsubjects significantly better, device on vs off4/5 subjects Dorn (2020) S2YesMotion direction discriminationsubjects significantly better, device on vs off5/5 subjects Dorn (2020) S2YesGrating visual acuityvisual acuity2.9 logMAR or better in 2/5 subjects Dorn (2020) S2NoLetter recognition via dynamic current steeringpercent correct26/30 correct (87%) Beauchamp et al., 2020 S3YesFunctional Low-Vision Observer Rated Assessment (FLORA benefit rating)positive or mildly positive benefit reported5/5 subjects Dorn (2020) S5Limited Orion tasks and framework staging. This table summarizes psychophysical, functional, and qualitative findings from Orion to indicate the stages currently supported by the literature. The available evidence supports S1-S3 and suggests limited psychosocial benefit in some recipients without establishing broad behavioral generalization or routine everyday adoption. FLORA-related support remains limited because only a general score is reported, without detail on specific tasks or self-assessments. Summary of stages The eight visual prosthesis devices reviewed here have not yet achieved substantial acceptance levels corresponding to S5 classification (Table 11). Of these projects, five are ongoing, and, with additional development and reporting, may advance into higher stages. Potential limitations to advancement are presented below (see Discussion). Table 11Highest stage for each of the eight reviewed devices.Asterisks indicate active projects that are anticipated to publish additional findings that are likely to affect the stage assignments. Entries marked as limited, partial, or provisional indicate assignments that are tentative due to limited measurements; hyphens indicate no relevant data have been published.DeviceS1S2S3S4S5Argus IIyesyesyesnonoAlpha IMS/AMSyesyeslimitedpartial-Dobelle Implantyesyesyes--BVT/Phoenix 99*yesyesyesnonoIMIE 256*yesyesprovisionalprovisional-CORTIVIS*yesyesyes--PRIMA*yesyesprovisional-noOrion*yesyesyes-limited Highest stage for each of the eight reviewed devices. Asterisks indicate active projects that are anticipated to publish additional findings that are likely to affect the stage assignments. Entries marked as limited, partial, or provisional indicate assignments that are tentative due to limited measurements; hyphens indicate no relevant data have been published. Cochlear implants as a model of completed adaptation The cochlear implant provides the clearest example of a sensory prosthesis that routinely progresses from percept generation to full acceptance including daily use. As such, the cochlear implant literature offers a comparative benchmark for evaluating visual prostheses. Both blindness and deafness produce cross-modal cortical reorganization, including recruitment of occipital cortex during auditory processing in blind individuals and visual cortex engagement during auditory tasks in deaf individuals (Fernández and Merabet, 2011). Restoration of auditory input through electrical stimulation therefore provides a useful analogue for studying how the brain incorporates patterned artificial sensory signals into everyday behavior (Hallum et al., 2007). Stage 1, device activation Electrically evoked auditory percepts emerge immediately after device activation. Recipients typically report tonal or buzzing percepts corresponding to stimulation delivered along the cochlear electrode array. Cortical responses to electrical stimulation can be measured within weeks of implantation, indicating reactivation of auditory pathways (Giraud and Lee, 2007). Behavioral improvement follows rapidly. In longitudinal studies using magnetoencephalography and speech perception testing, increases in word recognition scores during the first post-implant year correlate with measurable reorganization of auditory cortical responses (Pantev et al., 2006). Reliable perception of electrically evoked sound therefore emerges soon after implantation (S1). Stage 2, Perceptual Calibration Perceptual calibration continues for several years as the auditory system resolves mismatches between electrode stimulation patterns and the cochlea's natural tonotopic organization. Because electrode arrays stimulate only a subset of cochlear locations, frequency allocations often differ substantially from the natural cochlear map (Di Maro et al., 2022; Venail et al., 2015). Longitudinal pitch-matching experiments demonstrate that perceived pitch associated with a given electrode can shift by more than two octaves over a period of 3–5 years as the brain recalibrates its interpretation of electrical stimulation (Reiss et al., 2007). This progressive remapping reflects neural adaptation to altered spectral input and establishes stable perceptual associations between stimulation patterns and environmental sounds (S2). Stage 3, Cognitive Integration Once calibration has progressed sufficiently, recipients begin to interpret stimulation patterns during goal-directed tasks. Speech perception provides the most widely studied example. Word recognition scores typically improve rapidly during the first year after implantation, with the largest gains occurring during the initial months of device use (Blamey et al., 2013; Holden et al., 2013). Adult recipients implanted after post-lingual deafness frequently reach word recognition levels between 50% and 80% correct depending on age and duration of deafness (Ruffin et al., 2007). Improvements continue with prolonged device use. In a cohort of 864 unilateral implant recipients, mean speech recognition scores increased steadily for approximately two years after implantation and continued to improve in some individuals for up to ten years (Krueger et al., 2008). These results indicate that artificial auditory input becomes integrated into cognitive processes supporting language comprehension (S3). Stage 4, behavioral generalization Speech perception improvements translate into measurable functional benefits in everyday listening environments. Meta-analyses of cochlear implant outcomes report significant improvements in hearing-related quality of life measures following implantation, including increased social participation and reduced listening effort (Crowson et al., 2017). Early implantation further improves these outcomes, as individuals implanted soon after hearing loss achieve higher speech recognition scores and greater improvements in daily communication compared with those implanted after prolonged deafness (Necula et al., 2013; Loy et al., 2010). These findings demonstrate reliable transfer of artificial auditory perception to everyday listening situations (S4). Stage 5, Psychosocial Adjustment Population-level data confirm that cochlear implant hearing can become integrated into normal social and occupational activities. Long-term follow-up studies of early pediatric cohorts show that many recipients complete mainstream education and obtain regular employment, reporting life satisfaction comparable to hearing peers (Geers et al., 2011; Spencer et al., 2012). Children implanted early frequently develop spoken language proficiency and academic achievement similar to normally hearing classmates (Nicholas and Geers, 2007; Clark et al., 2012). These outcomes demonstrate that artificial auditory perception can support routine participation in education, employment, and social life. Psychosocial incorporation emerges as recipients begin to treat the implant as an ordinary sensory modality rather than an assistive technology requiring conscious interpretation. Improvements in hearing-related quality of life are observed even among individuals with moderate speech recognition performance, indicating that psychosocial benefit does not require perfect auditory function (Straatman et al., 2014). Surveys of adult cochlear implant recipients report that more than 90% would recommend implantation to others with severe hearing loss (Lassaletta et al., 2006; Rembar et al., 2009). Longitudinal studies show durable, multi-year improvements in hearing-related quality of life measures following cochlear implantation (McRackan et al., 2018; Ma et al., 2023). These findings indicate successful incorporation of the implant into everyday life and personal identity (S5). Technical advances Advances in implant design and rehabilitation have increased the reliability of this progression. The LOCHI study demonstrated that children implanted earlier achieve significantly better speech, language, and functional outcomes than later-implanted peers, with advantages persisting five years after implantation (Ching et al., 2018). Improvements in electrode design, signal processing, and clinical rehabilitation have reduced variability in outcomes relative to earlier implant generations (Blamey et al., 2013). The cochlear implant field therefore demonstrates that artificial sensory stimulation can progress through the full sequence from percept generation to durable everyday use. Perceptual learning without implantation Sensory substitution devices (SSDs) provide a complementary perspective on adaptation. These systems translate visual information into auditory or tactile stimulation, allowing users to interpret spatial information without direct activation of visual pathways. Because the sensory input is delivered through intact sensory channels, these devices isolate the process of perceptual learning from biological constraints associated with implanted electrodes. Building on that complementary example are cases of surgical restoration of vision which allow examination of adaptation following the reintroduction of natural retinal input after prolonged visual deprivation. The visual signal re-enters the normal sensory pathway, but having been absent for years or decades, is functionally a novel modality. These cases therefore provide an intermediate comparison between sensory substitution and neural prostheses: the input is visual, but its successful use still depends on post-restoration perceptual and cognitive adaptation. Sensory substitution devices for vision Users of sensory substitution devices initially encounter unfamiliar sensory codes that lack perceptual meaning. Systems such as The vOICe convert images into auditory soundscapes in which vertical position corresponds to pitch and brightness corresponds to loudness (Ward and Meijer, 2010). Other systems encode spatial information through tactile stimulation delivered to the skin (Danilov and Tyler, 2005), tongue (Bach-y-Rita et al., 1998), or musical encoding schemes such as Eye-Music (Abboud et al., 2014). More recent SSDs sonify color, distance, and temperature, extending the range of cross-modal mapping (Hamilton-Fletcher et al., 2022). Early experiences with these systems typically involve fragmented auditory or tactile patterns requiring deliberate interpretation (S1). Guided training allows users to associate these signals with spatial features of the environment. In one study using The vOICe system, blind participants achieved object recognition accuracy exceeding 80% after approximately 70 h of training distributed over several weeks (Amedi et al., 2007). Functional imaging studies demonstrate that training with sensory substitution devices activates occipital cortical regions normally associated with visual processing (Striem-Amit et al., 2012). These results demonstrate that the brain can learn to interpret novel sensory codes as spatial information (S2–S3). Learning trajectories vary substantially among individuals. Some participants achieve functional object recognition after days to weeks of training (Abboud et al., 2014; Amedi et al., 2007), whereas others require months of practice or fail to reach comparable performance even after extended exposure (Maidenbaum et al., 2014). Reviews of sensory substitution research emphasize that functional interpretation typically emerges only after prolonged training, often exceeding 50–100 h of practice (Kupers and Ptito, 2014). Despite these laboratory successes, everyday use remains limited. Many sensory substitution users discontinue device use after initial experimentation or restrict use to occasional demonstrations (Elli et al., 2014). Even expert users report that interpreting soundscapes requires deliberate scanning strategies and sustained attention (Ward and Meijer, 2010), and that environmental noise often interferes (Buchs et al., 2019). The cognitive effort required to interpret these signals therefore limits routine daily adoption. Surgical restoration of vision Restoration of optical vision following long-term blindness provides a parallel example in which perceptual capability can, but does not necessarily, translate into everyday visual behavior (see Supplementary Table 1). In some cases, especially when treatment occurs in childhood, rapid gains are observed. Children treated as part of Project Prakash (Sinha, 2016; Sinha et al., 2013; Sinha and Held, 2012) demonstrate quick progress on tests of visual function (Gandhi et al., 2017; Ostrovsky et al., 2006, 2009) with durable quality-of-life improvements (Kalia et al., 2017). In contrast, several well-documented adult cases describe the difficulty of translating restored sensory input into routine visual behavior. Patient MM, for example, who regained patterned visual input after more than forty years of blindness caused by corneal damage, showed profound deficits in object and face recognition. Despite restoration of visual acuity to 1.3 cycles per degree (roughly 20/460), MM relied on deliberate inspection of local features to overcome a lack of automatic coherent object perception (Fine et al., 2003). Dickens’ unnamed patient regained sight after penetrating keratoplasty and achieved visual acuity sufficient for face recognition and reading with magnification, yet continued to rely on auditory cues to identify familiar individuals and reported difficulty interpreting complex visual scenes and managing social interactions (Dickens et al., 2021). Patient SB, who regained sight after more than four decades of blindness, struggled to interpret visual scenes, perceiving fragmented glimpses rather than coherent scenes (Gregory and Wallace, 1963). Sacks’ patient Virgil, who experienced disorientation and confusion after cataract surgery restored partial visual input following decades of blindness, frequently reverted to tactile strategies during tasks he initially attempted visually (Sacks, 1995). The variability in outcome is thought to be related to the combination of age at treatment and duration of blindness (Lewis and Maurer, 2005; May et al., 2022). Restored sensation versus restored perception These observations indicate that restoring input enables but does not guarantee the restoration of perception, especially in adults. Sensory information may be available, but effortless, automatic interpretation to support action may not be achieved, even with training. Across both sensory substitution and surgical restoration, adaptation therefore appears to hinge on the perceptual learning that underpins the acquisition of visual fluency. This distinction between sensation and perception applies to functional restoration with visual prostheses as well. Discussion The opening premise of this review is that, despite substantial advances in device engineering, outcomes with visual prostheses remain limited, and an understanding of the causes of those limits is incomplete (da Cruz et al., 2016; Stingl et al., 2015; Petoe et al., 2024; Levy et al., 2024). The staged framework proposed here is meant to provide perspective by organizing those outcomes through the lens of broad behavioral achievement rather than through numbers of phosphenes or isolated psychophysical success. That distinction is necessary because artificial vision can create reproducible percepts, measurable discriminations, and even statistically significant task gains without yielding generalized, behaviorally useful vision (da Cruz et al., 2016; Dagnelie et al., 2017; Edwards et al., 2018; Petoe et al., 2024). Beyond whether a device is successful at restoring function, the relevant question is whether that function becomes a trusted source of information for everyday action. Viewed in that way, the literature shows a pattern across stages. Most devices reach Device Activation (S1) and Perceptual Calibration (S2) (da Cruz et al., 2016; Edwards et al., 2018; Petoe et al., 2024; Fernández et al., 2021). A smaller set of recipients proceed to Cognitive Integration (S3) under constrained conditions (da Cruz et al., 2013; Zrenner et al., 2011; Xu et al., 2021; Fernández et al., 2021). Far fewer show convincing Behavioral Generalization (S4), and sustained voluntary adoption through Psychosocial Adjustment (S5) remains rare at the population level (Dagnelie et al., 2017; Geruschat et al., 2016; Petoe et al., 2024; Levy et al., 2024). The framework therefore serves both descriptive and analytic purposes: it provides a common language to unify heterogeneous outcomes, and it identifies the principal bottleneck in present-day artificial vision. With devices demonstrating lower-stage successes, the central problem is now a failure to transform task-specific training into generalized visual behavior. The same framework also helps separate device technological capability from exceptional patient cases. Single-subject successes, highly trained users, and narrow task mastery can suggest broader functional restoration than the underlying evidence supports (Kapetanovic et al., 2020; Fernández et al., 2021; Dobelle, 2000). The framework, by emphasizing population-level performance, breadth across tasks, and high thresholds for safety-dependent use, resists that overgeneralization. Under the proposed stage criteria, the reviewed devices show functional promise, but only limited evidence for routine vision-like use. However, as artificial vision is a highly active field, these assessments are not set in stone, and new developments are expected to advance the stagings given here. Empirical pattern across devices Stage 1–2: consistent achievement Across retinal and cortical prostheses, the most reproducible findings are device activation, phosphene detection, coarse localization, and simple discrimination (da Cruz et al., 2016; Edwards et al., 2018; Petoe et al., 2024; Fernández et al., 2021). These early stages represent the most dependable achievement of current systems. Stage 3: structured perceptual interpretation Stage 3 is less common, but clearly present. Some recipients identify letters, discriminate simple forms, interpret constrained stimuli, or use device-mediated input for goal-directed action. Argus II provides the clearest evidence: closed-set letter identification improved markedly with the device on, and fine hand-movement tracing showed reduced error, although performance remained slow and effortful (da Cruz et al., 2013; Barry and Dagnelie, 2012). Similar patterns appear in subretinal systems, Alpha IMS/AMS, PRIMA, and IMIE, where some subjects read letters and words or perform controlled recognition tasks, but within narrow task structures and with substantial intersubject variability (Zrenner et al., 2011; Edwards et al., 2018; Muqit et al., 2024; Xu et al., 2021). Comparison with other sensory interventions clarifies the significance of that pattern. Cochlear implants commonly progress beyond early ambiguity into everyday listening benefit, quality-of-life improvement, and durable incorporation into education, work, and social life (Blamey et al., 2013; McRackan et al., 2018; Spencer et al., 2012). Sensory substitution devices likewise show that the brain can learn non-canonical mappings and extract meaningful spatial information after prolonged training, although often with substantial attentional demand and limited everyday adoption (Amedi et al., 2007; Ward and Meijer, 2010; Maidenbaum et al., 2014; Chebat et al., 2018). Surgical restoration after extended blindness provides the complementary and cautionary lesson that restored retinal input does not guarantee immediate coherent visual understanding (Fine et al., 2003; Gregory and Wallace, 1963; Šikl et al., 2013; Dickens et al., 2021). Stage 4–5: rare or absent Transfer to generalized behavior with visual prostheses remains uncommon. Independent scene interpretation, open-ended object search, unscripted navigation, and sustained everyday reliance are rarely documented at the population level. Some recipients show benefit in simplified mobility tasks or selected daily activities, and some report subjective value, but broad Stage 4 support is uncommon, as is Stage 5 (Dagnelie et al., 2017; Geruschat et al., 2016; Edwards et al., 2018; Petoe et al., 2024). Patient-reported benefit is seen, but does not preclude heterogeneous daily reliance, declining observed use, fatigue, and abandonment (Geruschat et al., 2016; Erickson-Davis and Korzybska, 2021; Güven et al., 2023; Petoe et al., 2024). Empirically, most devices lie at or below the boundary between structured interpretation (S3) and generalized behavioral use (S4). Adaptation trajectories are also heterogeneous across individuals and frequently can be non-monotonic. Gains present at one interval may be absent at another, as when Alpha IMS performance advantages had disappeared by six months (Stingl et al., 2015) and Argus II self-reported benefit declined from 80% to 65% between the first and third post-implant years (Ho et al., 2015). Interpretation can vary across time within the same individual (Lane et al., 2015), and disengagement can be reversed by renewed training (Kapetanovic et al., 2020). Variability in cortical plasticity following prolonged deprivation likely contributes: occipital cortex is recruited cross-modally in the long blind (Fernández and Merabet, 2011), yet remains available to trained non-canonical codes (Striem-Amit et al., 2012), and outcomes after surgical restoration depend on age at treatment and duration of blindness (Lewis and Maurer, 2005; May et al., 2022). Trajectory shape, rather than endpoint alone, is therefore informative, which motivates the recommendation below for longitudinal monitoring of device use. The point of failure: generalization Lack of skill transfer The most consistent point of failure is the transition from trained task performance to transferable visual behavior. Skills acquired during training on one task do not consistently extend to other tasks or novel situations (da Cruz et al., 2013; Fernández et al., 2021; Petoe et al., 2024). Instead, performance commonly remains tied to the details under which it was acquired, including stimulus class, contrast regime, spatial arrangement, response options, or scanning routine. A subject may learn to identify a limited letter set, discriminate a geometric form, trace a high-contrast line, detect a doorway, or localize a salient object, yet those abilities falter under more complex environments (Dagnelie et al., 2017; Stronks and Dagnelie, 2014). Response time carries particular importance. Many successful Stage 3 tasks require tens of seconds per trial, and sometimes multiple minutes (da Cruz et al., 2013; Ayton et al., 2014; Xu et al., 2021). That time scale is incompatible with ordinary visual behavior, which depends on rapid evaluation of scene structure to support continuous movement planning and updating. Slow performance suggests sequential problem solving through partial evidence and deliberate inference rather than fluid visual perception. That dissociation may help explain how statistically significant task gains do not necessarily imply real-world utility. The broader comparative literature supports the same interpretation. Cochlear implant users also begin with low-fidelity, unfamiliar signals, but speech perception typically becomes faster, more accurate, and less effortful, allowing conversation in ordinary settings (Blamey et al., 2013; Ma et al., 2023). Surgical visual restoration after long deprivation often yields normal visual acuity without normal visual understanding, showing that the presence of signal alone is necessary but not sufficient for higher-order perception (Fine et al., 2003; Gregory and Wallace, 1963; Dickens et al., 2021). Sensory substitution devices can support meaningful interpretation, but users often report high attentional cost and selective use (Ward and Meijer, 2010; Maidenbaum et al., 2014; Buchs et al., 2019). Possible reasons for failure of generalization Two potential explanations recur across the literature: task-specific learning and high cognitive load. First, recipients may solve laboratory tasks through memorized stimulus-response associations rather than through formation of broadly useful visual abilities. A trained letter set, limited number of objects, or familiar tracing pattern is well suited to template matching. Such strategies can be effective for a given task, but have little utility outside the trained context, consistent with broader accounts of perceptual learning in which improvement can remain closely tied to trained stimulus conditions and task demands (Ahissar and Hochstein, 2004; Sagi, 2011). Second, interpretation often appears effortful rather than automatic. Users describe deliberate decoding, concentration, and fatigue; behaviorally, long response times and sensitivity to task difficulty point to high inferential demand (Erickson-Davis and Korzybska, 2021; Levy et al., 2024; McDonald, 2019; da Cruz et al., 2013; Ayton et al., 2014; Xu et al., 2021). Learning under those conditions is slow, and the resulting competence is narrow. Contributing factors to cognitive load Several signal-level factors likely produce cognitive load. The first is low information content. Beyond the typically coarse resolution of current visual prostheses, nominal electrode count can overstate useable spatial information because current spread, receptive-field overlap, and overlapping activation volumes can produce crosstalk or phosphene-phosphene interaction, while fading and disease-driven circuit remodeling further reduce effective channel independence (Shivdasani et al., 2014; Sinclair et al., 2016; Bosking et al., 2022; Pérez Fornos et al., 2012; Moleirinho et al., 2021; Jones et al., 2012). The resulting signal may not support the spatial resolution required for scene segmentation and object disambiguation. The second is encoding mismatch. Electrically-evoked neural patterns differ substantially from natural retinal, thalamic, or cortical activity (Maturana et al., 2018; Millard et al., 2015). Even when downstream circuitry can extract some structure from the imposed code, the signal lacks many of the spatial and temporal regularities on which ordinary vision depends (Pérez Fornos et al., 2012; Jepson et al., 2014; Millard et al., 2015; Meital-Kfir and Pezaris, 2023). The brain may therefore learn correspondences to that code without converting it into a natural perceptual representation. A third factor is distorted spatial representation. Phosphene position, size, and shape vary across electrodes and, in some systems, across time (Sinclair et al., 2016; Bosking et al., 2022). That inconsistency weakens the reliability on which visuospatial learning depends. If similar inputs do not reliably produce similar sensations, each perceptual judgment requires separate interpretation rather than automatic readout. A narrow field of view compounds the problem by forcing serial sampling. Users must scan over time to accumulate information that natural peripheral vision would provide in parallel, with implications for mobility and safety (Ayton et al., 2014; da Cruz et al., 2016; Petoe et al., 2024). Restricted dynamic range adds a further constraint by compressing contrast and fine spatial structure. Decoupling from the oculomotor plant imposes yet another burden: eye and head movements may fail to generate the perceptual shifts expected by the visual system, especially in head-steered systems (Caspi et al., 2018; Paraskevoudi and Pezaris, 2019; Paraskevoudi and Pezaris, 2021; de Ruyter van Steveninck et al., 2024; Waclawczyk et al., 2025). Under all those conditions, the user must learn an unfamiliar code and an unfamiliar way of sampling the world. Implications for device design These observations place information throughput at the center of device design. Numbers of electrodes matter only insofar as they increase the amount of information available for behavior. Effective spatial resolution, tissue selectivity, confinement of current spread, and reduced cross-channel interaction are important because they determine how much distinct structure the signal can carry. The engineering objective is thus a pattern of stimulation rich enough to support higher-order perception and visually guided action, rather than just many phosphenes. Field of view warrants equal emphasis. Many current systems have electrode arrays across a spatially limited portion of the visual field, creating a narrow sampling window that forces serial scanning and lacks critical peripheral representation to support orientation and obstacle avoidance. By expanding field of view, or more effectively integrating exploratory behavior across the visual scene, cognitive burden should be reduced, and the chances for generalization improved. Encoding methods remain similarly important. We know from cochlear implants that stimulation patterns need not replicate natural neural activity exactly to provide meaningful utility, but many encoding approaches used in visual prostheses to date have relied on relatively simple image transformations, including luminance-based stimulation, contrast enhancement, and edge detection (Liu et al., 2005; Parikh et al., 2010; Stronks and Dagnelie, 2014; Holiel et al., 2024). Encoding schemes more fully informed by biological coding principles may produce percepts that require less deliberate reconstruction and support faster interpretation (Freeman et al., 2011; Beyeler et al., 2017). The same argument supports automated calibration of phosphene patterns that becomes especially important with high electrode-count devices, because distorted or drifting perceptual maps can turn otherwise basic interpretation into a continuing perceptual burden. Sensorimotor coupling should be treated as a primary design feature. Integration between stimulation and eye or head movement will improve visuomotor calibration, reduce the need for compensatory strategies, and bring device use closer to natural visual function (Caspi et al., 2018; Paraskevoudi and Pezaris, 2021; Titchener et al., 2018; de Ruyter van Steveninck et al., 2024; Waclawczyk et al., 2025). Interfaces that mimic natural gaze control through eye-and-head steering will align percept generation with intended sampling of the visual scene, thereby supporting faster visuomotor calibration and more fluent use. Implications for rehabilitation Rehabilitation should be organized around progression between stages rather than optimization for isolated tasks. Training should begin with easier tasks, because easier discriminations are learned more readily and, by recruiting higher-level representations, are more likely to generalize (Ahissar and Hochstein, 1997, 2004). Easier tasks, with their faster learning in artificial vision (Killian et al., 2016; Rassia and Pezaris, 2018), should function as scaffolds rather than endpoints. If training remains exclusively on detection or narrow discrimination tasks, it may strengthen compensatory routines without promoting perception. For that reason, rehabilitation should emphasize perception over detection, and generalization over task repetition. Recipients should encounter varied exemplars, altered layouts, changing contrast regimes, new contexts, and tasks requiring transfer rather than memorized response patterns with an objective of learning visual behavior that supports the use of device-mediated input as an information source across multiple domains. Accordingly, rehabilitation outcomes should be assessed on tasks, exemplars, and environments held out from training to avoid confounding improvement on a trained task with the memorized stimulus-response strategies described above, and to ensure that results accurately reflect evidence of adaptation. Retention should likewise be probed after training concludes, as gains acquired under supervision may not persist in the long term (Kapetanovic et al., 2020). Self-reinforcing tasks may be especially important. Tasks yielding intrinsically meaningful results in orientation, manipulation, mobility, or communication are more likely to sustain practice than abstract psychophysical drills; reading connected text, rather than identifying individual letters or isolated words, is a prime example (Sommerhalder et al., 2004; Rassia et al., 2022). Extended rehabilitation, including structured at-home work, is therefore likely necessary (Amedi et al., 2007; Maidenbaum et al., 2014; Kapetanovic et al., 2020). Adaptation will be facilitated with longer training periods, reinforcement outside the clinic, and rehabilitation aimed at broad functional use rather than endpoint demonstrations alone. Device usage should be recorded quantitatively and followed longitudinally, with intervention triggered when it declines. Device abandonment, selective deployment, and reduced wear time are practical indicators that higher stages are not being reached (Güven et al., 2023; Petoe et al., 2024; Erickson-Davis and Korzybska, 2021; Levy et al., 2024) and that, anecdotally, can be rescued when identified (Kapetanovic et al., 2020). Quality-of-life instruments, structured interviews, and psychosocial support should accompany behavioral testing. Stage 5 depends on sustained voluntary adoption beyond laboratory success, which requires continued therapeutic involvement. Psychosocial care is especially important to manage unmet expectations, disappointment, stigma, internal conflict, and self-attribution of failure that can reduce use even when some functional benefit exists (Lane et al., 2015; Erickson-Davis and Korzybska, 2021; Karadima et al., 2023; Levy et al., 2024). Overall, the reviewed literature supports a consistent conclusion. Present visual prostheses usually succeed in generating percepts and sometimes succeed in producing structured perceptual interpretation, but they rarely support generalized visual behavior and only inconsistently support sustained adoption (da Cruz et al., 2016; Edwards et al., 2018; Petoe et al., 2024; Levy et al., 2024). The principal barrier lies between Stage 3 and Stage 4. That threshold likely reflects multi-factorial influences that limit progress by limiting the complexity of learned skills and impeding their generalization. Progress will therefore require coordinated advances in engineering and rehabilitation: better signals, better mappings, better user interfaces, and training designed explicitly for transfer of skill. Conclusion This review proposes that adaptation to artificial vision is best understood as a multi-stage pathway influenced by technology, training, context, and personal meaning. By formalizing these stages and comparing outcomes to auditory and non-visual modalities, we reveal critical barriers to achieving perceptual fluency and behavioral integration. The findings suggest that future success in artificial vision will require a shift from focusing solely on improving technical parameters, such as phosphene count, to enabling intuitively useable, meaningful vision. High-resolution stimulation alone may be insufficient: what matters most is what users can do with the perception they receive, how naturally it appears, and how easily fluency is developed, so that artificial vision becomes a habitual part of engagement with the world. Meaningful restoration of function will require going beyond just delivery of sensation, as has been the historical focus, to the recreation of visual perception. As the field advances, progress in hardware miniaturization, stimulation specificity, and perceptual encoding holds promise for overcoming current limitations. Just as cochlear implants evolved from crude auditory prostheses into highly effective communication tools, visual prostheses may eventually deliver rich, intuitive vision. Realizing this potential will require technical advances in concert with rehabilitation plans to support users through each stage of adaptation, from first percept to fluent use in daily life. Supplementary Material Acknowledgements This work was supported by: the NIMA Foundation; the National Institutes of Health EY029322; The William M. Wood Foundation, Bank of America, Trustee; the Fulbright Foundation in Greece; Peter and Yayi Pezaris; Donald Good; Michael Gersh. No sponsor had any role in the planning, preparation, or realization of this study. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.preteyeres.2026.101494. Footnotes Human and animal subjects declaration This report is a review of existing literature, and as such, no human or animal subjects were used for data collection. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the authors used ChatGPT for copy-editing. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. CRediT authorship contribution statement Katerina Eleonora K. Rassia: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. Lydia Liapi: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. John S. Pezaris: Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Software, Supervision, Writing – review & editing. Declaration of interests JSP serves on the Scientific Advisory Board of Phosphoenix BV. KEKR and LL declare no competing interests. References Associated Data Supplementary Materials