Objective diagnosis of ADHD using IMUs.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Medical Diagnostics

Background:

• Attention-Deficit/Hyperactivity Disorder (ADHD) diagnosis relies on subjective assessments.
• Objective diagnostic tools are needed to improve ADHD identification and management.
• Inertial sensors offer a potential method for quantifying movement patterns associated with ADHD.

Purpose of the Study:

• To evaluate the efficacy of miniature wireless inertial sensors for objective ADHD diagnosis.
• To develop a machine learning model for classifying ADHD patients based on motion data.
• To determine which types of activities yield the most accurate ADHD classification.

Main Methods:

• Wireless inertial sensors (accelerometers and gyroscopes) collected motion data from subjects in a clinical setting.
• A support vector machine (SVM) algorithm was employed for binary classification (ADHD vs. non-ADHD).
• Motion data was analyzed during both free time and a continuous performance test (CPT).

Main Results:

• The SVM model achieved a classification accuracy exceeding 95% for ADHD diagnosis.
• Motion data recorded during the CPT demonstrated superior classification performance compared to free time.
• Inertial sensor data effectively differentiated between individuals with and without ADHD.

Conclusions:

• Miniature wireless inertial sensors provide a highly accurate, objective method for ADHD diagnosis.
• Utilizing motion data from forced concentration tasks, like the CPT, enhances diagnostic accuracy.
• This technology represents a significant advancement in ADHD assessment tools.