Hybrid Deep Learning Model Integrating Long Short-Term Memory Networks and Convolutional Neural Nets for Predicting Anesthesia Outcomes in Surgical Procedures

Abstract

Anesthesiologists administer three types of medications during surgery to modulate anesthesia. These are analgesics to block pain, hypnotics to induce and sustain unconsciousness, and muscle relaxants to prevent muscle responses. Modern anesthesiologists use equipment to track unconsciousness in real-Time by attaching monitors to the patient's brow, which displays signals obtained from electroencephalogram activity, enabling determination of the patient's level of awareness. The amount of anesthetic injected is influenced by the patient's preoperative physical state, intraoperative vital signs, and postoperative recovery period. This necessitates quick thinking and specialized knowledge from anesthesiologists to safeguard patient vital signs throughout surgery. Given the unique intraoperative conditions and indicators of each patient, a practical system is proposed to forecast the degree of consciousness under anesthesia. As a result, the suggested method offers a reliable way to anticipate the patient's level of awareness while under anesthesia. The ensemble deep learning techniques using convolutional neural network and long short-Term memory achieved a 94% accuracy.