Identification of ventilatory dynamics using pseudorandom binary sequences

A controversy exists concerning how best to describe the dynamics of the ventilatory response to exercise in man. The results of studies which used step, ramp and sinusoidal forcings of work have implied that there is a single slow, presumably humoral, component. However, application of short duration pulses of very heavy work gave results consistent with a rapid, presumably neural, component in addition to the slower dynamics. There are two factors which may be responsible for the conflict: (1) differences in the frequency content of the input signals and (2) the large work loads used in the pulse study may have induced an artifact. In the present study five subjects exercised on a bicycle ergometer while the work load was varied as a pseudorandom binary sequence (PRBS). The first order Wiener kernels were calculated by cross-correlation techniques. All subjects clearly demonstrated a rapid component. When the frequency content of the PRBS was lowered such that it excited a range similar to the sinusoidal studies, a rapid component was not evident in the cross- correlation. This suggests that previous studies failed to identify a rapid component because of the low range of frequencies over which the respiratory system was stimulated, Work loads used in this study were of moderate intensity. Evidence is presented which verifies that the PRBS technique is a valid method of testing the respiratory system. The advantages of using this technique are also discussed.