Supervising-type stage identification method by using mixed sleep electroencephalogram signals and electrooculogram signals

Abstract

The invention discloses a supervising-type stage identification method by using mixed sleep electroencephalogram signals and electrooculogram signals. The supervising-type stage identification method includes: subjecting multiple epochs of the mixed sleep electroencephalogram signals and electrooculogram signals to stage marking; filtering the epochs to remove interferences, and subjecting the the filtered epochs to characteristic parameter extraction so as to obtain characteristic parameters; subjecting the characteristic parameters to statistical processing, and establishing a decision-making tree; acquiring unmarked overnight sleep signals, and identifying sleep stages according to the decision-making tree to obtain an intermediate sleep stage result; revising the intermediate sleep stage result to obtain a sleep stage result. The supervising-type stage identification method by using the mixed sleep electroencephalogram signals and electrooculogram signals is capable of increasing sleep stage identification rate.

Description

Technical field [0001] The present invention relates to the technical field of sleep staging and interpretation, and in particular to a supervised sleep EEG and mixed signal stage interpretation method. Background technique [0002] Currently, in the paper "Virkkala J, Velin R, Himanen SL, et al. Automatic sleepstage classification using two facial electrodes[J]. Conf Proc IEEE Eng MedBiol Soc, 2008, 2008: 1643-1646" the sleep stage interpretation method is adopted FP1 and FP2 lead EEG signals of EEG, without using EOG signal and mandibular muscle electrical EMG signal. The algorithm uses the differential signal of the FP1 and FP2 lead signals to perform FFT transformation to obtain the total energy of each frequency band, and obtain the total energy of delta frequency band, total energy of theta frequency band, total energy of alpha frequency band, total energy of sigma frequency band, and total energy of beta frequency band. Energy, total energy of muscle electric frequency ba...

AI Technical Summary

Problems solved by technology

[0003] In the existing sleep stage interpretation method, the recording of the epoch of the sleep EEG mixed signal segment is limited by the noise collected by the hardware equipment, which affects the sleep signal discrimination results, and the characteristic parameters of some epochs are not obvious, so the independent use of characteristic parameters and decision trees Interpretation of this type of epoch will cause a lot of misinterpretation, and the dimension of the characteristic parameters is not rich enough, only limited to the EEG signal, which leads to a low recognition rate of sleep stages and a low utilization efficiency of sleep signals

Images