Fig 4. Overestimate of the variability of the SPR by linearization, disregarding the local cGMP depletion or suppressing spatial resolution of the ROS.

Time dynamics of $t\to \text{CV}\left[{\text{E}}_{\text{int}}^{*}\left(t\right)\right]$ and t → CV(I_int(t)), as defined in (Eqs 15 and 5). Computations are for WT (blue traces); NLD (non local depletion) (dashed red traces), i.e., with the coefficient of [E*]_σ in the right hand side of Eq 9 kept at its dark value [cGMP]_dark; LIN (linearized) (dotted cyan), where the nonlinear fluxes of Eqs 10–12 were linearized; LIN+NLD (dotted purple), where both effects were enforced. Panels A: FSR Model; Panels B: TWS Model; Panels C: GWS Model. Upper Panels (A1,B1,C1): Mouse with D_cG = 120 μm²/s; Middle Panels (A2,B2,C2): Mouse with D_cG = 330 μm²/s; Lower Panels (A3,B3,C3): Salamander with D_cG = 160 μm²/s.