Automated detection of ADHD: Current trends and future perspective

Area of Science:

• Neuroscience
• Computer Science
• Medical Diagnostics

Background:

• Attention deficit hyperactivity disorder (ADHD) is a complex neurodevelopmental disorder often diagnosed late due to its heterogeneity and clinician shortages.
• Early diagnosis and treatment of ADHD are crucial for managing symptoms and improving neurodevelopmental outcomes.

Purpose of the Study:

• To review current literature on machine learning and deep learning applications for ADHD diagnosis.
• To identify and categorize diagnostic tools utilized in AI-based ADHD research.
• To pinpoint research gaps and suggest future directions for AI in ADHD diagnostics.

Main Methods:

• Systematic literature review of machine learning and deep learning studies focused on ADHD diagnosis.
• Categorization of studies based on diagnostic modalities: MRI, physiological signals, questionnaires, performance tests, and motion data.

Main Results:

• AI shows promise in diagnosing ADHD using diverse data sources including brain MRI, physiological signals, and motion data.
• Significant research gaps exist, particularly in the availability of public datasets for most modalities and the underutilization of wearable sensor data (ECG, PPG, motion).

Conclusions:

• AI offers a potential pathway to enhance the efficiency and timeliness of ADHD diagnosis.
• Future research should focus on creating accessible datasets and exploring novel data sources like wearable devices to develop robust AI-driven clinical decision support systems for ADHD.