Document Type : Original Research
Authors
1 Department of Electrical Engineering Khorasan Institute of Higher Education, Mashhad, Iran
2 Department of Electrical and Computer Engineering, Hakim Sabzevari University, Sabzevar, Iran
3 Psychiatry and Behavioral Sciences Research Center, Mashhad University of Medical Sciences, Mashhad, Iran
4 Biomedical Engineering Department, Semnan University, Semnan, Iran
Abstract
Background: Drug abuse causes substantial psychological and physical harm to individuals, highlighting the critical need for advanced diagnostic and treatment methodologies.
Objective: This study aimed to develop a highly accurate automatic detection system for substance abuse, specifically targeting Methamphetamine (Meth), Cannabis (Can), and Opioid (Op) users.
Material and Methods: This descriptive study developed a drug abuse detection system based on nonlinear Electroencephalogram (EEG) signal analysis combined with a Support Vector Machine (SVM) classifier. It also examined changes in EEG signal complexity associated with Meth, Can, and Op abuse by extracting determinism and complexity parameters using Recurrence Quantification Analysis (RQA).
Results: The observed decrease in EEG complexity in the Op and Meth groups suggests that these substances may reduce cognitive or behavioral complexity. Conversely, increased complexity in the Can group compared to the Healthy Control (HC) group may indicate enhanced complexity associated with cannabis use. The classification system achieved 88.77% accuracy, 87.69% sensitivity, and 96.30% specificity.
Conclusion: The designed automatic diagnostic assistance system, leveraging nonlinear brain data analysis, effectively differentiates Meth, Op, and Can users from HC individuals.
Keywords
)