Many natural phenomena behave in a chaotic way, and the classical linear analytical methods don't allow a reliable detection of interdependency among simultaneous measurements of different variables. Hence new analytical methods based on Lorenz system have been developed, in order to reveal non-linear interaction among those variables.
This is especially interesting in the
case of physiological monitoring, where signals are correlated with one
another by feedback mechanisms. Cardiorespiratory interaction -interdependency
between heart rate and respiration rhytms- is one example where traditional
linear tools have been used in order to gain information about its underlying
dynamics. In physiological conditions the power spectrum of the heart rate
contains a peek centered at the respiratory frequency. In pathological
conditions the classical tools, like the power spectrum, may show no signs
of interaction. As a consequence the use of a technique aimed to detect
hidden frequencies in times series may show an interdependance cardiac
and respiratory signals even in pathological conditions.