Estimation of chemoreflex loop gain using pseudorandom binary CO2 stimulation.

We have developed a method for deriving estimates of the chemoreflex control loop gain (LG) from the ventilatory response to inhaled CO2, modulated between 0% and 5% in the form of a pseudorandom binary sequence. The corresponding changes in alveolar (and thus, arterial) CO2 result from two components: 1) the direct effect of breath-to-breath changes in inhaled CO2 and 2) the chemoreflex-mediated changes in ventilation. LG between 0.01 and 0.03 Hz, the frequency range pertinent to periodic breathing, was estimated by computationally delineating the first component from the overall ventilatory response. The method was tested against simulated and experimental data. In both cases, we found strong correlations between our predictions and LG magnitude estimates derived by other methods. However, LG phase estimates were considerably more variable when compared to model predictions based on small-signal analysis. We propose that our method, which uses data from a single test procedure lasting < 10 min, may be more useful than traditional tests of chemoresponsiveness for the quantitative assessment of respiratory control stability during changes in sleep-wake state.