نهمین کنگره نقشه برداری مغز

نمایش مقاله

Title:

Machine Learning Applications in Resting-State EEG for Pediatric Cognitive Subtyping: A Systematic Review of Executive Functions and Emotion Regulation

Subject:

Artificial Intelligence and Brain Mapping

Presenter Author:
Name:	Mahnaz Vahab

Author:
Name 1:	Mahnaz Vahab - Corresponding - Presenter
Affiliation 1:	Department of Psychology, Islamic Azad University, Tehran West Branch, Tehran, Iran

Name 2:
Affiliation 2:

Name 3:
Affiliation 3:

Name 4:
Affiliation 4:

Name 5:
Affiliation 5:

Name 6:
Affiliation 6:

Name 7:
Affiliation 7:

Name 8:
Affiliation 8:

Name 9:
Affiliation 9:

Name 10:
Affiliation 10:

Abstract

Original Research

introduction :

Resting-state electroencephalography (EEG) is increasingly used in pediatric cognitive neuroscience as a non-invasive method for capturing individual variability in brain function. At the same time, advances in machine learning (ML) have provided new opportunities for data-driven identification of neurocognitive profiles that extend beyond conventional diagnostic frameworks. The present systematic review summarizes current research applying ML techniques to resting-state EEG data for cognitive subtyping in children, with an emphasis on executive functioning and emotion regulation.

Methods:

A systematic search of the literature was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. PubMed and Scopus were searched for original peer-reviewed articles published between 2015 and 2025. Studies were eligible if they examined resting-state EEG in pediatric samples (children or adolescents) and employed machine learning methods to investigate executive functions, self-regulation, emotion regulation, or closely related cognitive domains. PubMed searches combined Medical Subject Headings (MeSH) with free-text terms, whereas Scopus searches targeted titles, abstracts, and keywords. Studies involving task-based EEG, adult-only samples, non-machine learning analyses, animal models, review articles, or conference proceedings were excluded.

Results:

The database search identified 1,130 records, from which 460 duplicates were removed prior to screening. Across the reviewed studies, machine learning models applied to resting-state EEG showed moderate to high effectiveness in distinguishing cognitive subgroups associated with inhibitory control, working memory, cognitive flexibility, and emotion regulation. Reported classification accuracies most commonly fell within the 65%–85% range, while deep learning models trained on larger datasets reached performance levels close to 90%. Spectral features—particularly activity within the theta and alpha frequency bands—along with functional connectivity metrics and non-linear EEG features, were most frequently identified as informative. Approximately 70% of studies incorporated spectral or connectivity-based measures in their predictive or clustering frameworks. Nonetheless, considerable variability was observed across studies with respect to EEG recording protocols, feature extraction approaches, machine learning techniques, and validation strategies.

Conclusion:

Overall, the literature supports the potential of integrating resting-state EEG with machine learning approaches as a robust framework for cognitive subtyping in pediatric populations. This approach allows for a more nuanced characterization of executive and emotion regulation profiles beyond categorical diagnostic labels. Progress in this field will depend on improved methodological consistency, clearer reporting standards, and greater attention to the interpretability of machine learning models to enhance clinical relevance and applicability within developmental neuroscience and brain mapping research.

Keywords:

Resting-state EEG-Machine learning-Cognitive subtyping-Executive functions-Emotion regulation-Brain mapping