Visual neuroprostheses / one experimental retina
Retina Laboratory
Place an implant. Follow the signal. Explore the resulting vision.
Drag to orbit · scroll to zoom · shift-drag to place
Stratified, procedural cells informed by retinal histology. Density is reduced for inspection. The shared retinal chart registers every device and response; the conductive solvers retain their documented planar domains.
Implant parts share one surface: a flat PRIMA plate or a curved electrical array. Close-up views crop the implant at the tissue window. Gap is referenced at placement; tissue displacement around an implant is not simulated. Hide Device to inspect obscured cells.
One scene. Four stages.
Original scene
Common visual field for every technology.
Device input
Scene sampled at the active device geometry.
Retinal response
Calculated activity on the common retinal population.
Predicted percept
Select a model and run an experiment.
Coupled output is an illustrative fixed-scale activity decoder. Published electrical and PRIMA appearance models use different brightness definitions and are not interchangeable measures of human vision.
Input, percept & legacy reference options
Legacy files and presets run their original engines inside this workspace. Switch a scene control to resume the common visual input. Current input: shared.
Follow the response in time.
Ganglion-cell spike raster
Tier 2 voltage and spike traces are from an assumed leaky integrate-and-fire retinal network.
Compare mechanisms, not scores.
Same scene, retinal chart, tissue parameters, response model and observation window. Device pulse schedules and model assumptions remain explicit. No overall device ranking is calculated.
Keep the experiment.
Settings, units, seed, equations and calculated arrays travel together. Imported results are recomputed.
Complete experiment settings
All retained engine parameters are editable here, including legacy eye, sampling and image settings. Applying settings clears calculated results.