Differentiating ADHD and Autism Among Young People: A Machine Learning Solution

Could machine learning assist in exploring how brain connectivity patterns in young people with ADHD and autism are distinguished. Both conditions share overlapping traits, making diagnosis challenging.

In this study the researchers used brain imaging data and machine learning techniques to differentiate between the two conditions with up to 85% accuracy.

Why is this important?

ADHD and autism frequently co-occur, and many individuals receive both diagnoses. However, they also have distinct neurological underpinnings. By identifying brain connectivity differences, this study provides insights that could improve diagnostic accuracy and lead to better-tailored treatments.

How was the study conducted?

• Data Sources: Researchers used publicly available brain imaging datasets—Autism Brain Imaging Data Exchange (ABIDE) and ADHD-200 Consortium.
• Participants: 330 individuals (110 with ADHD, 110 with autism, 110 neurotypical controls) aged around 11.6 years.
• Method:
  • Functional Magnetic Resonance Imaging (fMRI) measured brain activity at rest.
  • Machine learning models analyzed brain connectivity patterns.
  • A technique called Linear Discriminant Analysis (LDA) helped distinguish ADHD from autism and control groups.

Key Findings

1. Machine learning was highly effective at distinguishing ADHD from autism, achieving an 85% accuracy rate.
2. Brain connectivity differences:
   • ADHD: Altered frontoparietal network (FPN) connectivity. This network is linked to attention, executive function, and impulse control.
   • Autism: More diverse brain connectivity differences, particularly involving language networks, the salience network (SN), and the default mode network (DMN).
3. ADHD vs. Autism: The frontoparietal network was a key differentiator. ADHD showed reduced connectivity in this area, whereas autism showed more variability in other networks.
4. Neurotypical vs. ADHD/Autism:
   • ADHD individuals had weaker connectivity in attention-regulating networks.
   • Autistic individuals displayed more widespread connectivity differences across various networks.

What does this mean?

• Potential for brain-based diagnostic tools: The study shows that fMRI and machine learning could help clinicians differentiate ADHD from autism more objectively.
• Distinct neural markers: The findings suggest that ADHD and autism are neurologically distinct conditions, not just variations of the same disorder.
• Future research directions: The study calls for larger datasets and further validation of machine learning models for real-world diagnostic applications.

Limitations

• The study only analyzed resting-state brain connectivity (not task-related activity).
• The dataset lacked adult participants and people with intellectual disabilities, limiting generalizability.
• More research is needed to see if these findings apply to individual clinical diagnoses rather than group comparisons.

Conclusion

This research provides strong evidence that ADHD and autism have distinct brain connectivity patterns.

By using machine learning, the study achieved high accuracy in differentiating the two, which could pave the way for more precise, neuroscience-based diagnostic tools in the future.

Sütçübaşı, B., Ballı, T., Roeyers, H., Wiersema, J. R., Çamkerten, S., Öztürk, O. C., Metin, B., & Sonuga-Barke, E. (2025). Differentiating Functional Connectivity Patterns in ADHD and Autism Among the Young People: A Machine Learning Solution. Journal of Attention Disorders, 29(6), 486-499. https://doi.org/10.1177/10870547251315230 (Original work published 2025)