Empirical considerations regarding the use of ensemble-variance analysis of macroscopic currents

Two fundamental properties of an ion channel are its single channel current (i) and its open probability (Po), which are most directly defined by single channel recordings. However, under some circumstances either practical limitations (e.g., extreme voltages) or unique channel properties may preclude the use of single channel recordings for definition of such properties. As an alternative, the variance in macroscopic currents (non-stationary noise analysis or variance–mean analysis) has been exploited to define fundamental elementary properties of the underlying channels. Although some limitations of the variance–mean approach have been considered in previous work by others, here simulation methods were used to define the conditions under which variance analysis can be suitable for providing estimates of i and Po. Of particular interest is the extent to which reasonably reliable estimates of Po can be obtained, even under conditions of Po less than 0.5. Empirically, the analysis indicates that, with sufficient numbers of sweeps and with constraints on initial estimates of i, reasonably reliable estimates of Po can be made down to 0.2. The impact of a number of other factors on the utility of variance–mean analysis are also considered, including effects of inactivation, filtering, and the consequences of particular gating schemes.