Deep reinforcement learning-based propofol infusion control for anesthesia: A feasibility study with a 3000-subject...

Area of Science:

• Anesthesiology
• Artificial Intelligence
• Machine Learning

Background:

• Anesthesia management requires precise drug administration.
• Dynamic patient conditions and co-administered drugs complicate anesthesia control.
• Autonomous systems can potentially improve anesthetic stability.

Purpose of the Study:

• To develop and evaluate a deep reinforcement learning (DRL) system for autonomous propofol infusion.
• To create a simulated environment for training and testing the DRL system under various conditions.
• To assess the system's ability to maintain stable anesthesia.

Main Methods:

• A DRL model was designed for propofol infusion control.
• A patient simulation environment was developed using demographic data.
• The system was trained to manage bispectral index (BIS) and effect-site concentration.
• Evaluations were conducted using data from 3000 subjects.

Main Results:

• The DRL system demonstrated effective propofol infusion control.
• Stable anesthesia was achieved despite dynamic patient conditions and remifentanil administration.
• The system successfully managed bispectral index (BIS) and effect-site concentration.

Conclusions:

• Deep reinforcement learning offers a promising approach for autonomous anesthesia control.
• The proposed system can maintain anesthetic stability in simulated dynamic environments.
• This AI-driven method has the potential to enhance patient safety during anesthesia.