Fusion Model Using Resting Neurophysiological Data to Help Mass Screening of Methamphetamine Use Disorder.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Data Science

Background:

• Methamphetamine use disorder (MUD) prevalence has increased, necessitating efficient mass screening methods.
• Previous MUD detection relied on physiological responses during drug-cue reactivity, which is time-consuming.
• The utility of resting-state physiological data for MUD detection without cue induction remained unexplored.

Purpose of the Study:

• To develop and evaluate a clinically applicable intelligent system for MUD detection using resting-state physiological data.
• To assess the efficacy of fusing electroencephalogram (EEG), heart rate variability (HRV), and galvanic skin response (GSR) data for MUD screening.
• To determine if physiological abnormalities at rest can differentiate MUD patients from healthy controls.

Main Methods:

• A system fusing 5-channel EEG, HRV, and GSR data was developed for MUD detection.
• Machine learning algorithms were employed to analyze resting-state data from 46 MUD patients and 26 healthy controls.
• Classification performance of different data fusion models was systematically compared.

Main Results:

• The fusion of HRV and GSR features achieved the highest separation accuracy of 79%.
• Integrating EEG, HRV, and GSR data provided robust information, enhancing classification accuracy across various models.
• The system demonstrated effective MUD detection without the need for drug-cue reactivity induction.

Conclusions:

• Resting-state EEG, HRV, and GSR fusion offers a clinically applicable method for MUD detection.
• The developed intelligent system is efficient, accurate, and suitable for mass screening in clinical settings.
• This approach simplifies MUD screening by eliminating the need for complex experimental setups and drug-cue administration.