Attention deficit hyperactivity disorder (ADHD) detection for IoT based EEG signal

Area of Science:

• Biomedical Engineering
• Neuroscience
• Computer Science

Background:

• Attention Deficit Hyperactivity Disorder (ADHD) is a common childhood behavioral disorder.
• Current ADHD diagnosis relies on subjective assessments, lacking objective markers.
• Early identification is crucial to mitigate ADHD's impact on life outcomes.

Purpose of the Study:

• To develop an innovative Internet of Things (IoT) based ADHD detection system.
• To utilize electroencephalogram (EEG) signals for objective ADHD identification.
• To enhance diagnostic accuracy and support clinical decision-making.

Main Methods:

• EEG signal processing with min-max normalization.
• Extraction of advanced features: improved fuzzy features, fractal dimension, wavelet transform, and non-linear features.
• Development of a hybrid PUDMO algorithm for optimal feature selection and classifier weight tuning.
• Implementation of a hybrid detection system integrating IDBN and LSTM classifiers.

Main Results:

• The proposed PUDMO algorithm achieved a high accuracy of 0.9649.
• Significantly outperformed existing methods (e.g., SLO, SOA, SMA, BRO, DE, POA, DMOA).
• Demonstrated the effectiveness of the hybrid classifier system in ADHD detection.

Conclusions:

• The IoT-based EEG analysis offers a promising objective approach for ADHD detection.
• The hybrid PUDMO algorithm enhances feature selection and classifier performance.
• This technology can assist clinicians in making more informed and objective ADHD diagnoses.