Latest contribution from Prof. Mohamad Sawan's research group on drug addiction characterization and mechanisms

  Drug addiction is a complex brain disorder involving abnormalities in multiple neural circuits including the reward system, cognitive control network, and emotional regulation.

Electroencephalography (EEG), with its non-invasive nature and high temporal resolution, has become a crucial tool for investigating the neural mechanisms of addiction. Among various EEG analytical methods, EEG microstate analysis represents an emerging approach that segments continuous brain activity into transient, quasi-stable spatial patterns (microstates), each lasting 60-120 milliseconds. These microstates reflect the brain's dynamic functional organization and are considered neural substrates of cognition, emotion, and behavioral regulation.

Fig.1.EEG microstate topographies.

While EEG microstates show great potential in psychiatric research, previous neurocharacterizations of addiction remained broad, lacking frequency-band-specific precision.

A recent study published in Frontiers in Psychiatry by CenBRAIN Neurotech addresses this gap. Our team employed multi-band frequency analysis and multi-task paradigms to investigate microstate dynamics, combined with machine learning to successfully distinguish methamphetamine users (MUD) from healthy controls with 85.5% classification accuracy, establishing refined endophenotypes for addiction research.

Xurong Gao, a PhD student from our center, is the first author of the paper, and Chair Professor Mohamad Sawan and Dr. Yun-Hsuan Chen is the corresponding author This work was supported by Westlake University, the Zhejiang Key R&D Program, with additional collaboration from Zhejiang Gongchen Compulsory Isolated Detoxification Center and Zhejiang Liangzhu Compulsory Isolated Detoxification Center.

Research Highlights

• Pioneered single-frequency-band microstate analysis;

• Incorporated multiple cognitive paradigms;

• Revealed inter-task microstate statistical differences;

• Established precise endophenotypes with improved classification accuracy;

Abstract

This study employed multi-band EEG microstate analysis combined with machine learning to investigate neural endophenotypes of methamphetamine use disorder (MUD). Leveraging machine learning's inherent sensitivity to input features, we characterized and ranked microstate parameters to identify those with the highest potential as MUD-specific endophenotypes.

Fig.2.Illustration of experiment protocol.

MUD patients showed significant α-band (8-12Hz) microstate abnormalities, including decreased coverage of Class A microstates and increased transition rates. The machine learning model based on α-band features achieved 85.5% diagnostic accuracy for MUD. This study is the first to establish frequency-specific microstate endophenotypes for MUD, providing new biomarkers for clinical diagnosis and treatment evaluation. Unlike previous studies relying solely on resting-state tasks, we incorporated visual induction tasks to expand the identification scope of MUD endophenotypes.

Fig.3. Classification results under different tasks.

EEG microstate analysis provides new perspectives for understanding the neural mechanisms of drug addiction. With its high temporal resolution and network-specific advantages, this approach is expected to advance precision medicine in addiction diagnosis and treatment. As technology progresses, this method may become an important paradigm in brain disorder research.