Automatic sleep stage division method based on width neural network

Abstract

The invention discloses an automatic sleep stage division method based on a width neural network. The method comprises the following steps: screening physiological data according to different characteristics of multi-channel signals and retaining effective data; carrying out band-pass filtering by a band-pass filter based on characteristics of all the channels, and removing ocular artifacts in anelectroencephalogram signal; decomposing and reconstructing C3 and C4 channel electroencephalogram signals to five basic rhythms by utilizing wavelet packet transformation and extracting time-frequency spectrums of all reconstructed signals and other physiological signals as features by utilizing short-time Fourier transformation; and selecting proper parameters according to the features, constructing a width learning system, matching the extracted feature information in the feature models of all sleep stages, performing sleep staging on a to-be-processed signal according to matching results,and verifying the effectiveness of a measurement system by using the average accuracy of ten-fold cross validation as a final result.

Description

technical field [0001] The present invention relates to the fields of artificial intelligence and medical treatment, in particular to using a machine learning method to automatically classify human sleep stages. Background technique [0002] Sleep is an inherent physiological activity of animals, a natural product of the evolutionary process, and the same is true for humans. Studies have shown that in a person's life, about one-third of the time humans are in a sleep state. Sleep can not only eliminate the fatigue of human beings during daytime activities, but also can relax and repair people's brain and various organs of the body. However, as the pace of life in today's society is getting faster and faster, people's pressure is also increasing day by day, and the sleep time and quality are gradually declining. The reduction of sleep quality will not only affect the brain's thinking, but also lead to physiological dysfunction, so that the human body is in a substandard stat...

AI Technical Summary

Problems solved by technology

At present, sleep staging research based on deep learning algorithms is in the ascendant, but most of these studies are aimed at databases with small data volumes, and have not proved their effectiveness

When using a database with a large amount of data, its long running time has become a bottleneck restricting its use

Images