Three-channel sleep apnea and hypopnea syndrome recognition device

Abstract

The invention provides a three-channel sleep apnea and hypopnea syndrome recognition device based on LSTM-CNN, and the device is characterized in that the device comprises an electroencephalogram signal preprocessing module, an electroencephalogram signal feature extraction module, a sleep state recognition module used for carrying out binary classification on a sleep stage and a non-sleep stage, a thoracic and abdominal displacement signal and mouth and nose airflow signal preprocessing module, an LSTM-CNN neural network module used for inputting a mixed data segment formed by a preprocessed thoracic and abdominal displacement signal and a mouth and nose airflow signal into an LSTM-CNN neural network, conducting training by adopting a ten-fold cross validation method, judging the specific classification of the input segment and outputting a prediction result, and an SAHS state identification module.

Description

technical field [0001] The invention relates to SAHS recognition technology, and specifically designs a device for recognizing SAHS by using LSTM-CNN neural network through three-channel physiological monitoring information of C4 / A1 single-channel EEG, chest and abdomen displacement and mouth and nose airflow. Background technique [0002] Sleep apnea and hypopnea syndrome (Sleep Apnea Hypopnea Syndrome, SAHS) is a common attack during sleep, not easy to be noticed chronic sleep disorder disease. The main symptom of the disease is severe snoring accompanied by repeated shallow breathing or apnea due to obstructed breathing during night sleep. The disease can cause excessive daytime sleepiness, inattention, and reduce the quality of life of patients in mild cases; in severe cases, it can cause various complications such as high blood pressure, diabetes, and stroke, which seriously threaten the lives and health of patients. [0003] Polysomnography (PSG) is the gold standard ...

AI Technical Summary

Problems solved by technology

Relatively speaking, the snoring-based detection technology is susceptible to interference from external environmental factors, and the detection quality fluctuates greatly and is unstable.

Images