Skip to main content

ORIGINAL RESEARCH article

Front. Neurosci.
Sec. Translational Neuroscience
Volume 18 - 2024 | doi: 10.3389/fnins.2024.1524513

Partial Directed Coherence Analysis of Resting-State EEG Signals for Alcohol Use Disorder Detection Using Machine Learning

Provisionally accepted
Ainul Khairiyah Mohd Nazri Ainul Khairiyah Mohd Nazri 1Norashikin Yahya Norashikin Yahya 2*Danish Mahmood Khan Danish Mahmood Khan 3Noor’Izni Zafirah Mohd Radzi Noor’Izni Zafirah Mohd Radzi 2Nasreen Badruddin Nasreen Badruddin 2Abdul Halim Abdul Latiff Abdul Halim Abdul Latiff 2Mohammed Abdulaal Mohammed Abdulaal 4
  • 1 PETRONAS Gas Berhad, Kerteh, Malaysia
  • 2 University of Technology Petronas, Tronoh, Perak, Malaysia
  • 3 Sunway University, Bandar Sunway, Malaysia
  • 4 King Abdulaziz University, Jeddah, Makkah, Saudi Arabia

The final, formatted version of the article will be published soon.

    Excessive alcohol consumption poses significant risks to physical and psychiatric well-being, lifestyle, and societal interactions. Chronic alcohol abuse induces alterations in brain structure, leading to the onset of alcohol use disorder (AUD), a condition with diverse detrimental effects.Early diagnosis and appropriate intervention are crucial for effective management and recovery.However, current diagnostic practices, primarily reliant on the subjective questionnaires, may benefit from supplementary objective measures. This study proposes a novel approach that utilizes electroencephalogram (EEG) classification, specifically focusing on effective connectivity (EC) derived from resting-state EEG signals in conjunction with support vector machine (SVM) algorithms for AUD detection. Estimation of EC is achieved through the partial directed coherence (PDC) technique, employing an EEG dataset comprising 35 individuals with AUD and 35 healthy controls (HCs). The primary objective is twofold: first, to evaluate the efficacy of connectivity features in distinguishing between AUD and HC using EEG signals, and subsequently, to develop and assess an EEG classification technique employing EC matrices and SVM for AUD identification. Results demonstrate promising performance, with the proposed methodology achieving a peak accuracy of 94.5% and an area under the curve of 0.988, notably utilizing frequency bands 29, 36, 45, 46, and 52. Furthermore, using feature reduction, classification based on brain rhythm analysis using SVM yielded an accuracy of 96.37 ± 0.45%, notably employing reduced PDC adjacency matrices from the gamma band. These findings highlight the 1 Ainul et al.potential of our developed classification algorithm as a valuable tool for AUD detection, enhancing diagnostic precision, and informing tailored treatment strategies.

    Keywords: Alcohol use disorder (AUD), effective connectivity, Electroencephalography (EEG), Partial directed coherence (PDC), support vector machine (SVM)

    Received: 08 Nov 2024; Accepted: 23 Dec 2024.

    Copyright: © 2024 Mohd Nazri, Yahya, Khan, Mohd Radzi, Badruddin, Abdul Latiff and Abdulaal. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Norashikin Yahya, University of Technology Petronas, Tronoh, 31750, Perak, Malaysia

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.