Automatic sleep staging method based on LSTM (long short term memory) and using multiple physiological signals

Abstract

An automatic sleep staging method based on LSTM and using multiple physiological signals comprises the following steps: step 1, signal acquisition: acquiring an ECG (electrocardiogram) signal, a respiration signal and an acceleration signal of a testee; step 2, signal processing; step 3, classification feature extraction; step 4, model construction: inputting artificially extracted features in a first-layer LSTM model, taking an output probability of the first-layer LSTM model as new features, and inputting the new features into a second-layer LSTM model together with the artificially extracted features to construct classifiers for different classification tasks; and step 5, applying trained models to classification of sleep staging. Multiple physiological signals, including the ECG signal, the thoracic-abdominal respiration signal and a head acceleration signal but not including an EEG (electroencephalogram) signal, are adopted in the invention, and defects caused by application of EEG to sleep staging are overcome; and meanwhile, the LSTM models are adopted and suitable for big samples and big data, besides, the temporal correlation between sleep events is considered, and the accuracy and reliability of sleep staging are improved.

Description

technical field [0001] The invention belongs to the technical field of biomedical signal processing, and relates to signal processing of electrocardiogram, respiration, and acceleration, and in particular to an automatic sleep staging method based on a long-short-term memory model LSTM and utilizing multiple physiological signals. Background technique [0002] Sleep is one of the most important life activities of the human body, and research on sleep patterns and sleep structures can help people improve their sleep quality. Since the 1960s, sleep medicine has formed a standard system after decades of development, and now the most commonly used is the sleep staging standard formulated by the American Academy of Sleep Medicine (AASM) in 2007: Based on Polysomnography (PSG), according to the "AASM Interpretation Manual of Sleep and Related Events", the nocturnal sleep activity is divided into five different periods, namely, wakefulness (W), sleep stage I (N1), sleep stage II s...

AI Technical Summary

Problems solved by technology

The neural network method based on LSTM has good classification ability for time series data, but it has not been well used in the field of sleep staging research

Images