.png){kind=icon}
.png){kind=icon}
Return
Psycho-Behavioral Disorders Classification Using Microstates Syntax and Natural Language Processing
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2026.3671788.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
EEG microstates represent global neural activity and are potential candidates for biomarkers for several mental disorders and brain diseases. The main analytical techniques for EEG microstates investigate their temporal features, such as duration, coverage, frequency of occurrence, and transition probabilities. Although these factors have proven to be essential metrics in some studies about mental disorders, the syntax of microstates, that is, their symbolic sequences, is mainly analyzed using information theory tools. Early studies have shown that these symbolic sequences have temporal characteristics, such as dependency, entropy, and periodicity, indicating that microstate syntax can be useful when treating problems related to mental disorders. In this paper, we propose a new methodology for detecting neural correlates of psycho-behavioral disorders. This new methodology uses microstate syntax and natural language processing tools to treat three problems of detecting psycho-behavioral disorders. The proposed method analyzes the microstate symbolic sequence as a text document and explores relevant word embedding features that differentiate patients into different groups. An SVM model trained with these features achieved accuracies of 98.56% for schizophrenia, 97.39% for depression, and 98.54% for dementia. These results demonstrate that the word embedding properties of symbolic microstate sequences are highly promising for the detection of psycho-behavioral disorders in binary classification settings.
Keywords:
EEG microstates
mental disorders
natural language processing
Journal
IF:
3.6
Papers:
9.8W
Citations:
29.4W