A hybrid retinal implant using stem-cell-derived neurons and microelectrodes lowers activation thresholds and improves signal precision, offering a more natural interface for high-resolution vision restoration.
(Nanowerk Spotlight) Blindness from retinal degeneration disrupts the very start of the visual process, yet much of the downstream neural architecture stays intact. This asymmetry raises a compelling possibility: could vision be restored by bypassing the damaged cells and working with what remains?
Electronic retinal prostheses have taken up this challenge. These implants are designed to substitute for lost photoreceptors by electrically stimulating the surviving layers of the retina. Some patients fitted with these devices can detect points of light or large shapes. But the vision they experience is fragmentary, low in resolution, and difficult to interpret. In practical terms, most remain functionally blind.
The technical barriers are well known. Electrodes must be positioned close to their target cells to work effectively, yet physical gaps between the implant and retina, especially in epiretinal systems, cause the electric field to weaken and spread. This reduces spatial precision and creates signal overlap, or crosstalk, between adjacent electrodes.
At the same time, current implants do not engage the retina’s internal logic. Natural photoreceptors operate using continuous, graded signals and modulate neurotransmitter release accordingly. Electronic devices, by contrast, deliver short, uniform pulses that activate neurons indiscriminately, bypassing the retina’s specialized ON and OFF circuits that normally encode contrast and motion.
The device combines a densely packed array of microelectrodes with photoreceptor-like cells derived from human stem cells, creating a bioelectronic interface designed to improve specificity, reduce energy requirements, and restore more natural patterns of retinal activity.
Hybrid retina concept illustration. (Image: Reprinted from DOI:10.1002/adfm.202512621, CC BY) (click on image to enlarge)
The prosthesis is built around a three-dimensional micro-well structure fabricated using photolithography. Each microwell, measuring about 10 micrometers across, contains an electrode at its base and is designed to house a single neuron. The well geometry forces the cell into close contact with the electrode, minimizing the distance between them. This configuration creates a high-resistance seal around the cell, which confines the electric field and increases stimulation efficiency.
To test the effect of this geometry, the authors used computer simulations to model the electric field around the cell. When the seal was tight, the local field strength at the membrane increased by more than a thousand-fold compared to flat electrode surfaces. These results predicted that the charge required to activate a cell could be reduced to the level of just a few picocoulombs.
This was supported by in vitro experiments, where engineered neurons in sealed microwells responded to currents as small as 5 picocoulombs. In comparison, neurons in flat configurations required several orders of magnitude more charge to produce similar responses.
The team then evaluated how well the device could isolate stimulation to individual pixels. Crosstalk between adjacent electrodes is a major obstacle to improving resolution in retinal implants. By simulating two neighboring wells, the researchers found that a tightly sealed configuration produced a large contrast between the target cell and its neighbor.
The activation threshold for the target cell was over 400 times lower than that of the adjacent one. This suggests that the design could support pixel pitches as small as 10 micrometers, which would correspond to a theoretical visual acuity of approximately 20/40. While this figure is derived from modeling and not from visual behavior, it reflects a significant advance over existing systems, where acuity rarely exceeds 20/550.
Crucially, the biological component of the device does more than simply receive electrical input. The researchers used human embryonic stem cells to produce photoreceptor precursor cells, known as PRPs. These cells are not light-sensitive but are capable of releasing glutamate, the key neurotransmitter used by natural photoreceptors. Patch-clamp recordings and calcium imaging showed that the PRPs could be activated electrically and would then modulate their internal calcium levels, a known signal of synaptic activity.
Moreover, some nearby cells showed synchronous calcium changes, suggesting the presence of functional synapses. These responses were blocked when glutamate receptors were inhibited, confirming that the cells were chemically communicating.
The device was implanted into the subretinal space of a rat model of retinal degeneration. Fluorescently labeled PRPs seeded into the microwells remained viable for at least 30 days, with a survival rate of around 80 percent. Importantly, the cells stayed inside the wells and extended axon-like projections toward the host retina. Microscopy revealed the presence of ribeye, a presynaptic marker, at the tips of these projections, often located near bipolar cells labeled with PKC-alpha. This colocalization suggests that the transplanted cells may be forming synaptic contacts with the host network.
To confirm the physical interface between the cells and the microwells, the team performed high-resolution imaging using confocal microscopy and transmission electron microscopy. The results showed that the PRPs closely conformed to the walls of the wells, with minimal extracellular space. In many cases, actin filaments formed a ring at the contact surface, indicating mechanical engagement and adhesion. These features are known to improve electrical coupling by increasing seal resistance and focusing the current into a confined space.
Experimental tests using surrogate cells in prototype devices showed that cells in sealed microwells required far less charge to activate than cells on flat surfaces or those separated from electrodes by a collagen layer. The measured charge thresholds were consistent with those predicted in the simulation studies, reaching the low picocoulomb range for tightly sealed configurations.
The authors also examined the broader viability of the implant. The fabrication process is compatible with existing microelectronics and uses SU-8, a widely used photoresist, to create the microwell architecture. Each 1-millimeter-wide device contains more than 3000 wells. A flexible handle allows the device to be implanted subretinally and connected to external stimulation systems.
Although this work does not yet demonstrate functional vision restoration in animal behavior, it provides a strong foundation. By embedding glutamatergic neurons into the prosthesis, the device avoids many of the limitations of direct electrical stimulation. Instead of applying pulses to the retina and hoping for meaningful responses, the hybrid system uses living cells to interpret the signal and pass it on using natural chemical pathways. This may allow for finer control, more selective activation of retinal circuits, and stimulation patterns that are more aligned with real visual processing.
The implications extend beyond the eye. This work points to a general strategy for interfacing electronics with neural tissue: use engineered cells to translate physical input into biological signaling. For retinal prostheses, it offers a path to higher resolution, lower power consumption, and better integration with the host network.
The next step will be to pair this architecture with a stimulation system capable of delivering patterned input and to test its performance in more complex visual tasks.
For authors and communications departmentsclick to open
Lay summary
Prefilled posts
Nanowerk Newsletter
Get our Nanotechnology Spotlight updates to your inbox!
Thank you!
You have successfully joined our subscriber list.
Become a Spotlight guest author! Join our large and growing group of guest contributors. Have you just published a scientific paper or have other exciting developments to share with the nanotechnology community? Here is how to publish on nanowerk.com.
