A multimodal approach for ADHD with coexisting ASD detection for children.

Area of Science:

• Neuroscience
• Developmental Psychology
• Biomedical Engineering

Background:

• Diagnosing co-occurring attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) in children presents significant challenges due to symptom overlap.
• Existing diagnostic methods often lack objectivity and can be influenced by subjective assessments.

Purpose of the Study:

• To develop and validate a novel multimodal approach for identifying children with ADHD coexisting with ASD.
• To fuse data from handwriting dynamics and brain activity to improve diagnostic accuracy.

Main Methods:

• A multimodal approach was developed, integrating pen tablet data (writing dynamics) and functional near-infrared spectroscopy (fNIRs) data (brain activity).
• Children with ADHD and ASD, alongside typically developing (TD) children, performed handwriting tasks (Zigzag line and periodic lines) under trace and predict conditions.
• Features from both data sources and conditions were fused, and a Sequential Forward Floating Selection (SFFS) algorithm with a support vector machine (SVM) classifier was used for evaluation.

Main Results:

• The proposed multimodal approach achieved a classification accuracy of 96.4% for the periodic lines (PL) task.
• This accuracy represents an improvement of over 2% compared to existing diagnostic studies.
• The method demonstrated significant potential in differentiating children with ADHD and coexisting ASD from TD children.

Conclusions:

• The developed multimodal approach offers a promising objective tool for diagnosing ADHD with coexisting ASD in children.
• This method can assist clinicians in providing more accurate and timely diagnoses.
• The study provides new insights into the neurodevelopmental characteristics of children with both ADHD and ASD.