Personalized features for attention detection in children with Attention Deficit Hyperactivity Disorder

Area of Science:

• Neuroscience
• Biomedical Engineering
• Psychiatry

Background:

• Electroencephalogram (EEG) signals are crucial for understanding brain activity and have applications in Attention Deficit Hyperactivity Disorder (ADHD) treatment.
• Frequency band powers and their ratios in EEG are established features for attention detection.
• Unanswered questions remain regarding the most discriminative EEG features for attention and their subject-specificity.

Purpose of the Study:

• To identify the most discriminative EEG features for distinguishing between attentive and non-attentive states in ADHD children.
• To investigate whether these discriminative EEG features are universal across subjects or subject-specific.

Main Methods:

• Utilized Mutual Information (MI) for subject-specific feature selection.
• Analyzed a large dataset comprising 120 children diagnosed with ADHD.
• Evaluated various EEG frequency band powers and their ratios as potential features.

Main Results:

• The theta beta ratio (TBR), relative beta power, and theta/(alpha+beta) ratio (TBAR) were found to be significant and informative for attention detection.
• Relative theta power, often used, showed limited discriminative information for a small percentage of subjects (3.26%).
• Demonstrated that while certain features are important on average, the optimal set of discriminative features varies significantly among individual subjects.

Conclusions:

• Relative beta power and TBAR are as significant as the commonly used TBR for attention detection in ADHD.
• Subject-specific analysis is essential for optimal feature selection in EEG-based attention monitoring.
• The findings contribute to a more nuanced understanding of EEG correlates of attention in ADHD, paving the way for personalized neurofeedback strategies.