The chaotic behaviour of resting human respiration.

The study of non-linear mathematics has led to the concept of an attractor that can confine to within certain boundaries the path traced out in multidimensional space when a variable is plotted against itself delayed by various lag times. The type of attractor present can be determined by monitoring the exponential rate (called a Lyapunov exponent) at which almost identical values within a time series become dissimilar. Lyapunov exponents estimated from data collected during the normal resting breathing of eight adults were all positive, end-tidal PCO2 had the lowest Lyapunov exponent (0.06 +/- 0.01 bits/s; Mean +/- SE), breath and expiratory duration had the highest exponents (0.23 bits/s). The results indicate that resting respiration is chaotic. This chaotic behaviour may allow fast and flexible responses to sudden changes, allow if necessary complete readjustment of the control parameters, and may be involved in the phase locking of respiration to external stimuli.