METHOD OF ESTIMATING POWER DIFFERENCES OF OSCILLATORY COMPONENTS OF ELECTROENCEPHALOGRAM SIGNALS IN PSYCHOPHYSIOLOGICAL CONDITIONS BASED ON QUANTILE ANALYSIS Russian patent published in 2020 - IPC A61B5/476 

FIELD: medicine.

SUBSTANCE: invention relates to medical technologies, namely electroencephalography, and can be used to studying brain bioelectric activity, designing brain-computer interfaces, determining and analyzing diagnostic parameters for post-stroke neurorehabilitation and diagnosing neurodegenerative diseases. Disclosed is a method of estimating power differences of oscillatory components of electroencephalogram (EEG) signals in psychophysiological conditions, including measurement and analysis of electroencephalogram components with subsequent mathematical processing, consisting in recording electroencephalograph signals in the first and second states of the test person, extracting a one-dimensional target signal from a multidimensional array of recorded EEG signals, selecting an envelope of narrow-band oscillations of the target signal for two states using narrow-band filtration of the selected target signal with subsequent calculation of the absolute value of the time series and smoothing the resultant signal, method includes sampling Q quantiles q from 0 to 1 with pitch of 0.01 of the envelope narrow-band oscillations for the first state and the second state, for each threshold value h from a sample of quantile values for the first state, calculating the number N(h) of envelope points in the first state and the number M(h) of envelope points in the second state exceeding the threshold value h, further, the obtained ROC-curve, which is a relationship N(h)=f(M(h)), is analyzed, and the difference of two states is assessed by the area limited by the given ROC-curve: if the area within range of 0.4 to 0.6 shows no difference in power of oscillatory components of signals of the electroencephalogram between the first and second states, if the area is less than 0.4 and more than 0.6, observing the presence of significant differences in power of the oscillatory components of the signals of the electroencephalogram of the first and second states, then determining the cross-quantile coefficient from the ratio ERD(q)=(S(q)-S`(q))/S(q)×100 %, where S is the q-quantile value of the envelope distribution in the first state, S` is the q-quantile value of the envelope distribution in the second state, and the difference of two states is evaluated by the average value M of the synchronization/desynchronization coefficient values of the cross-quantile synchronization/desynchronization curve, which is determined from ratio ; if the average M value of the cross-quantile synchronization/desynchronization curve is within the range of -10 % to 10 %, there are no differences in power of the oscillation components of the electroencephalogram signals between the first and second states, if the average M of the cross-quantile synchronization/desynchronization curve is less than -10 % or more than 10 %, observing the presence of significant differences in power of the first and second state electroencephalogram signal oscillation components, wherein the negative value M indicates the presence of desynchronization in the second state compared to the first state, the positive values of M indicate the presence of synchronization in the second state compared to the first state.

EFFECT: invention provides higher effectiveness of differentiating psychophysiological conditions on the basis of analysis of cerebral oscillatory activity.

1 cl, 2 dwg