Detection of ADHD From EOG Signals Using Approximate Entropy and Petrosain's Fractal Dimension

Area of Science:

• Neuroscience
• Biomedical Engineering
• Clinical Psychology

Background:

• Attention deficit hyperactivity disorder (ADHD) is associated with distinct eye movement patterns compared to neurotypical individuals.
• Electrooculogram (EOG) signals, reflecting eye movements, are hypothesized to differ between ADHD patients and healthy controls.

Purpose of the Study:

• To investigate and compare EOG signal characteristics in children diagnosed with ADHD versus a healthy control group.
• To evaluate the efficacy of machine learning classifiers in differentiating between ADHD and control groups based on EOG features.

Main Methods:

• Collected EOG signals from 30 children with ADHD and 30 healthy children during an attention-related task.
• Calculated approximate entropy (ApEn) and Petrosian's fractal dimension (Pet's FD) as key EOG signal features.
• Employed Support Vector Machine (SVM) and Neural Gas (NG) classifiers for group discrimination.

Main Results:

• Both ApEn and Pet's FD values were significantly lower in the ADHD group compared to the control group.
• The SVM classifier achieved higher accuracy (84.6%) in distinguishing between ADHD and control groups than the NG classifier (78.1%).

Conclusions:

• Reduced EOG complexity (ApEn, Pet's FD) in ADHD suggests slower, less variable eye movements, linked to attention deficits.
• Findings support the potential use of EOG analysis for ADHD assessment and the development of EOG-based biofeedback therapies.