FIELD: medicine.
SUBSTANCE: parameters of electroencephalogram (EEG), F-response from muscles of thenar eminence during stimulation of right median nerve; parameters of statistical and spectral variability analysis of RR intervals dynamic range (HRV) are recorded. The following parameters are calculated: power of EEG theta vibrations in O2 lead, average frequency of EEG theta vibrations in P3 lead, power of EEG alpha vibrations in P4 lead, power of EGG alpha vibrations in T4 lead, power of maximum F-response, voltage index according to data from HRV analysis, mean square deviation of RR intervals dynamic range, power of low-frequency component of HRV spectrum. The resulting parameters are analyzed using artificial neural network constituting a multilayer perceptron with 8 input layer neurons, 4 intermediate layer neurons and 1 output neuron, previously guided to predict dynamics of carbon dioxide level in the exhaled air in subjects to hyperventilation stress.
EFFECT: method enables to increase the accuracy of prediction, which is achieved by accounting the complex of studied neurophysiological parameters.
3 tbl
