A multimodal approach for ADHD with coexisting ASD detection for children
Abstract Identifying attention-deficit/hyperactivity disorder (ADHD) with coexisting autism spectrum disorder (ASD) for children is a challenging issue due to their complexity and overlapping symptoms. This study investigated from handwriting and executive function viewpoints simultaneously and deve...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-05000-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Identifying attention-deficit/hyperactivity disorder (ADHD) with coexisting autism spectrum disorder (ASD) for children is a challenging issue due to their complexity and overlapping symptoms. This study investigated from handwriting and executive function viewpoints simultaneously and developed a novel multimodal approach for identifying ADHD with coexisting ASD by fusing pen tablet and fNIRs data. This study used pen tablet and fNIRs device to compare writing dynamics and brain activity between ADHD with coexisting ASD and typically developing (TD) children during handwriting patterns. Two handwriting tasks including Zigzag line (ZL) and periodic lines (PL) were adopted for data collection. Each task had two conditions: trace and predict. Various statistical features were derived from pen tablet and fNIRs data for each task. These features were then combined by fusing features derived from the trace and predict conditions to make two datasets (PL and ZL). The potentiality of these features was evaluated using Sequential Forward Floating Selection (SFFS)-based algorithm and support vector machine (SVM) was employed to evaluate the performance of ZL and PL tasks. Data were collected from 13 ADHD children with co-occurring ASD and 15 TD children to evaluate the proposed ZL and PL tasks. The experimental results demonstrated that the proposed SFFS-SVM model achieved a classification accuracy of 96.4% for PL task. This is an improvement of more than 2% classification accuracy compared to existing studies. This approach shows promising potential and assisting physicians and clinicians to provide an objective and accurate diagnosis of ADHD with coexisting ASD. This study proposes a novel approach that increase the detection rate and provides new insights for further research. |
|---|---|
| ISSN: | 2045-2322 |