A real, numerically simulated walk through rod phototransduction — the biochemical cascade that turns one captured photon into a measurable electrical signal. Every figure on this page is the direct output of a real simulation, not an illustration drawn to match a textbook number.
scroll ↓
At the dim-light threshold a human observer can just barely detect, the real simulated flash intensity needed is — photons at the cornea on average — and because photon arrival is a Poisson process, the actual count on any given trial varies around that mean.
Most of those photons never reach a rod's light-sensitive outer segment at all: pre-retinal tissue absorbs roughly half, and even a photon that arrives is only captured with real outer-segment cross-section probability.
A captured photon isomerizes rhodopsin with real quantum yield ≈0.67 (Dartnall 1972) — most captures succeed, some don't. A successful isomerization produces one active R* molecule, the real trigger for everything downstream.
R* is short-lived — this model uses a human/primate-appropriate ~100ms effective lifetime, deliberately faster than the classic toad-rod recordings the field's early quantitative models were built on.
While active, one R* catalytically activates many transducin (G*) molecules. Each G* activates a PDE6 molecule — one of the fastest enzymes known — which hydrolyzes cGMP. This is the real amplification step: one photon's effect multiplies through three enzymatic stages before it ever reaches the membrane.
This simulation models each stage as its own coupled differential equation (R*→G*→PDE*→cGMP), integrated with real 4th-order Runge-Kutta numerics — not a single black-box "response" number.
cGMP-gated channels in the outer segment membrane are open in darkness, carrying a steady "dark current." As the cascade hydrolyzes cGMP, cooperative (Hill-equation) channel gating closes a fraction of them, suppressing that current — the real measurable single-photon response.
This simulation's single-photon run: peak current suppression of —, recovering to 10% of peak by —.
300 real points from the cascade simulation's RK4 integration (downsampled by stride for the chart, not smoothed). Peak suppression and recovery time are read directly from this trace.
A real, checkable physiological property reproduced by the model: response scales roughly linearly with photon count at low numbers (1→2 photons is almost exactly 2×), then saturates as cGMP depletion becomes nonlinear.
The thermal-activation relationship behind spontaneous ("dark") rhodopsin isomerization — the same real mechanism that predicts colder-blooded species should show exponentially lower dark noise, consistent with Aho et al. 1988's cross-species data.