Figure 4. An NCX partition of 25% in T-space and 75% in M-space provides the best fit to experimental data.

Peak Ca_m lags by 25 ms relative to peak I_NCX, reflecting the NCX partition. The fraction of NCX in T-space, β (right panels), estimated from best fit to experimental data (left panels). Experimentally recorded Ca_m (A, left) and I_NCX (B, left) evoked by caffeine pulse; V was held at −80 mV. Phase plots of I_NCX vs. Ca_i are constructed in C from the data in A and B. It is clear from panel C that I_NCX is a bi-valued function of Ca_m; that is, one value of Ca_m (e.g. c₁) is associated with two values on I_NCX (α₁ and α₂). In panel C, right, β (fraction of NCX in T-space) is varied from 0 to 0.5. The continuous black trace is for 25% of I_NCX (β= 0.25) in the T-space, which provides the best fit for the experimental data. Note that for β= 0 (grey dashed trace), I_NCX is a single-valued function of Ca_m. D, I_NCX (black) and Ca_m (grey), corresponding to the traces in panels A and B, are shown on an expanded time scale. Ca_m, normalized to its peak value, is superimposed on normalized and inverted I_NCX. Note that peak I_NCX precedes peak Ca_m by about 25 ms in both simulation and experiment.