engleski

A new approach in time series analysis of psychophysiological data

The time series analysis is an appropriate technique for analysis of continous psychophisiological variables, which may appear in time as steady states, oscillations, deterministic chaos, and noise. The oscillations within certain cyclic processes can be linear, i.e. they are defined by a sinusoidal function, or non-linear, which significantly deviates from such a function. In the case of linear oscillations, exsisting mathematical tools (cosinor analysis, autocorrelations, spectral analysis) are appropriate for description and explanation of such rhythms. In the case of more complex rhythms, however, with tendency of nonlinear dynamics, these are not adequate. Theory of nonlinear dynamics (&#8220 ; chaos theory&#8221 ; , &#8220 ; deterministic chaos&#8221 ; , &#8220 ; science of complexity&#8221 ; ) seems appropriate for explanation of chaos in psychophisiological processes. Chaos has been formally defined as "stohastic behaviour in a deterministic system", i.e. a system which displays apparent random behaviour, but has an underlying pattern of lawfulness. The mathematical techniques dealing with such processes usually include description of chaotic systems, extraction of one-dimensional map of time series, geometrical portrait of the strange attractor generating chaotic behaviour, construction of a Poincare&#8217 ; s diagram, determination of fractal dimension of the attractor and its Liapunov exponent(s), and testing for nonlinear determinism. Over the last fifteen years there has been much interest in research of chaos in biological systems such as cardiovascular and nervous system, where the appropriate mathematical tools have been used for analysis of electrophysiological data. Some of the recent findings of such analyses indicate that time series of heart beat intervals may contain &#8220 ; hidden information&#8221 ; (multifractal structure of heart beat dynamics) which is neither visually apparent, nor extractable with conventional methods of analysis. More recent studies pointed out that mentioned mathematical techniques and the parameters they yield, may differentiate various mental load levels, as well as the stress effects, which may be reflected via specific chaotic dynamics of various biological subsystems.

time series analyisi; nonlinear analysis; chaos theory

nije evidentirano