Research on artificial intelligence-based computer-assisted anesthesia intelligent monitoring and diagnostic methods in health care

In the field of health care, anesthesia is a crucial therapeutic measure, but it also carries certain risks. Insufficient or excessive anesthesia can lead to significant consequences for patients, such as intraoperative awareness and impaired spontaneous breathing. Therefore, monitoring the depth of anesthesia is one of the vital life-supporting measures during clinical surgery. Currently, commonly used clinical indicators such as blood pressure, heart rate, and respiratory rate are used to estimate the depth of anesthesia in patients. However, due to variations in patients' physical conditions and anesthesia medications, these indicators exhibit significant differences in their performance such that there is not reliable that analyzing these clinical indicators alone. Therefore, considering that electroencephalogram (EEG) reflects a high degree of brain activity, this paper proposes an intelligent detection for anesthesia based on the transformer framework and EEG signals. First, the original single-channel EEG is preprocessed to extract spectral and differential entropy features. Subsequently, the two types of features are fused and sent to the transformer encoder network to complete the anesthesia depth prediction. Finally, the validation of the proposed algorithm was completed on the sevoflurane anesthesia dataset from Waikato Hospital in Hamilton, New Zealand, and a high prediction probability of 85.32% was achieved.