Fig 7. The effect of the lower voltage bound V_low of oscillations on f_res and Z_max constrains the optimal models.

a. An example of the change in Z(f) measured in the biological PD neuron for V_low = -60mV (black line) and V_low = -70mV (grey line). Inset shows the bounds of voltage clamp inputs in the two cases. b. Shifting V_low from -60 mV to -70 mV lowers the value of f_res measured in the PD neuron significantly, without influencing Z_max (b. Experimental). f_res and Z_max values measured in a random subset of optimal model neurons corresponding to low or high ${\overline{g}}_{Ca}$ values produced the same f_res and Z_max values at V_low = -60mV (black dots), but distinct f_res and Z_max values at V_low = -70mV (low ${\overline{g}}_{Ca}$: red dots; high ${\overline{g}}_{Ca}$: cyan dots). A subset of optimal models could reproduce the experimental result in which f_res shifted to significantly lower values without affecting Z_max. (grey dots). (c) ${\overline{g}}_{Ca}-V_{1/2}^{C{a}_{\infty }^h}$ relationship separating out the different groups of models producing different responses to changes in V_low (colors correspond to b Model panel). Models depicted by grey dots are referred to as intermediate ${\overline{g}}_{Ca}$ models. (d1-e3) mean voltage-gated ionic currents I_Ca, I_H and I_Ca+I_H and I_total, shown as a function of voltage for V_low = -60 mV (d1-d3) and V_low = -70 mV (e1-e3). Numbers correspond to the location along the ${\overline{g}}_{Ca}-V_{1/2}^{C{a}_{\infty }^h}$ as shown in c. f. The intermediate ${\overline{g}}_{Ca}$ models (grey dots) show a distinct ${\overline{g}}_{Ca}-V_{1/2}^{C{a}_{\infty }^h}$ linear correlation. g. Intermediate ${\overline{g}}_{Ca}$ models (grey dots) show a distinct and tighter ${\tau }_m^{Ca}-{\tau }_h^{Ca}$ correlation compared to all optimal models (black dots). h. Intermediate ${\overline{g}}_{Ca}$ models (grey dots) show a strong ${\overline{g}}_{Ca}-{\overline{g}}_H$ linear correlation that is not observed for all optimal models (black dots).