Cortical phosphene vision simulator

Interactive simulation of camera-based visual encoding for a hypothetical array of stimulation sites in human primary visual cortex (V1).

What is being simulated: stimulation sites sample a camera-derived activation map. Above a manual threshold, Graded coding maps intensity to requested current and a saturating brightness response; Binary coding switches on a fixed, site-specific brightness. Stimulation site density follows a V1 cortical-magnification approximation, and phosphene size is estimated from electrical current spread. The display uses general stimulation-site terminology; its current size and brightness models are based on electrical microstimulation and are not calibrated for optogenetic stimulation. Stable, mildly elliptical spots have adjustable central and peripheral edge sharpness. Processing models: Sobel emphasizes local luminance boundaries; Luminance represents image intensity; Luminance + Sobel mixes both. Sobel has fixed gain, without automatic normalization. A smaller field of view restricts the stimulation site array to the central camera region without zooming the image. This is a research visualization, not a patient-specific prediction.

QR code linking to the phosphene vision simulator

Ready.

Field of view i

Stimulation sites i

Processing model i

Stimulation site spacing i

Phosphene size i

Stimulation site dropout i

Position irregularity i

Activation threshold i

Boundary contribution i

Phosphene adaptation i

Adaptation time i

Recovery time i

Max attenuation i

Mirror camera i

Update debug views i

Stimulation coding i

Central sharpness i

Peripheral sharpness i

Processing 0 FPS

Working stimulation sites 0

1. Camera

2. Sobel edges

3. Encoded image

4. Stimulation site samples

Stimulation site view: white = active, gray outline = available but currently inactive, red = unavailable stimulation site.

Model and limitations

The model uses mildly elliptical generalized-Gaussian profiles, with no separate halo, and additive blending capped by display brightness. Sharpness is an adjustable phenomenological parameter: the central-to-peripheral gradient is not calibrated to a patient. Sharpness preserves the peak and half-maximum contour, but changes integrated light. Size uses a fixed reference-current spread approximation, not dynamically varying spread with requested current. A minimum displayed size of 0.55 processing pixels (Gaussian-equivalent scale) keeps tiny spots visible; this is a rendering compromise, not a physiological minimum. Camera horizontal field of view is assumed to be 70°, not measured. The simulator omits patient-specific retinotopy, cortical folding, gaze tracking, nonlinear stimulation site interactions, stimulation safety, and pulse-train timing. Adaptation and the graded current–brightness response are simplified models; binary coding is a comparison strategy.

Build: simple shaped / graded–binary / manual threshold / no automatic Sobel gain / original FOV — 8 September 2026.