Brain-like algorithm for sleep staging by applying prefrontal lobe single-channel electroencephalogram signals

Abstract

The invention discloses a brain-like algorithm for sleep staging by applying prefrontal lobe single-channel electroencephalogram signals, and the construction of the algorithm comprises the following steps: collecting Fp1-Fp2 electroencephalogram data, removing artifacts from the collected signals, filtering, carrying out segmentation processing and data labeling, extracting data features after segmentation processing, building an HDC model, carrying out model storage, and carrying out sleep staging, wherein the filtering operation on the collected signals comprises high-low pass filtering and 50Hz power frequency notch filtering on the electroencephalogram signals with artifacts removed. The algorithm has the advantages that the brain-like algorithm, namely the HDC algorithm, is used, the classification effect of a medium neural network can be achieved only through training of a small amount of data, even the model effect of classification can be achieved through one-time learning, and more importantly, the HDC algorithm is quite light and suitable for being applied to local hardware such as wearable equipment, direct learning and prediction can be realized; and a lot of resources can be saved in the aspects of data transmission and model storage.

Description

technical field [0001] The invention relates to a brain-inspired algorithm for sleep staging, in particular to a brain-inspired algorithm for sleep staging by using a single-channel EEG signal in the frontal lobe, and belongs to the technical field of sleep staging. Background technique [0002] Sleep monitoring, like heart rate, blood oxygen, blood pressure and other daily physiological indicators monitoring, is one of the very important daily monitoring items. Sleep monitoring is not only needed by people with sleep disorders, but also by many ordinary people and some patients with chronic diseases. Based on brain Electrical sleep staging will be much more accurate than methods based on ECG or pulse rate, body movement, etc. [0003] Currently, sleep monitoring methods based on EEG signals are mainly based on multiple channels, such as EEG, EOG, and EMG, or a single channel based on the position of the central axis of the head, such as Fpz-Oz, Pz-Cz and other single-chann...

AI Technical Summary

Problems solved by technology

[0003] Currently, sleep monitoring methods based on EEG signals are mainly based on multiple channels, such as EEG, EOG, and EMG, or a single channel based on the position of the central axis of the head, such as Fpz-Oz, Pz-Cz and other single-channel EEG signals, but there are few algorithms based on the single-channel Fp1-Fp2 of the prefrontal cortex, and most of the current calculation methods for sleep staging use machine learning or deep neural network algorithms , the amount of data required by this type of algorithm is often relatively large, and the trained model will be relatively large, which is not suitable for sleep stage learning and prediction on the local microcontroller or FPGA hardware

Images