Analysis of hemodynamic features to identify the activation of different brain regions of children during mental tasks of different difficulty levels

This is the authors' abstract. We don't add key points for this paper.

The study of brain activation in children for mental tasks significantly contributes to several areas of cognitive research and practical applications. This study identifies the key brain regions of children while performing mental tasks with different difficulty levels. In this study, the hemodynamic response is recorded using functional near-infrared spectroscopy (fNIRS) of 17 children while performing the mental tasks of three different levels of complexities, low, medium and high. A total of 1080 hemodynamic features are extracted from the preprocessed signal data. In order to rank the features, a mutual information-based feature ranking method is applied to select significant features. Thereafter, the k-nearest neighbor (KNN), support vector machine (SVM) and Ensemble classifiers are used to find the most significant set of features to classify the different levels of mental tasks as low, medium and high. The ensemble classifier based on bagging demonstrated superior performance compared to KNN and SVM, achieving high accuracy in differentiating the three levels of task difficulty. The findings highlight that changes in task complexity are highly associated with activation of the dorsolateral prefrontal cortex (dlPFC) and ventrolateral prefrontal cortex (vlPFC) of the left hemisphere. The regions are significantly involved in executive functions such as working memory, attention control and decision-making, indicating their important role in adapting to increasing cognitive demands. The study provides more insight into the brain activation of children, which may help in early diagnosis of developmental disorders, support mental health research and designing personalized instructional materials.