AI finds three brain–behaviour profiles across childhood ADHD and anxiety

An analysis of Adolescent Brain Cognitive Development data finds three brain–behaviour profiles among children aged 9–11 with ADHD, anxiety or both. The work appeared in Nature Mental Health on 16 October.[1]

The broader research question is familiar in psychiatry. The US National Institute of Mental Health’s RDoC framework encourages scientists to study functions and mechanisms across diagnoses, because symptom combinations vary and conditions frequently overlap. The aim is to connect biology with behaviour at a finer level than the diagnostic label alone allows.[4]

What went into the profiles

The model combines resting-state functional MRI—measuring brain activity without a specific task—with five cognitive measures and eight parent-reported symptom scales. It learns two brain–behaviour dimensions, then groups children along them.[3][1]

The independent test is the more revealing step

In 224 independent, age-matched Healthy Brain Network participants, brain–behaviour correlations were r = 0.336 for cognitive control and r = 0.179 for emotion–attention regulation. The relationships transferred, but remained modest.[2]

That second cohort has a useful design difference. Healthy Brain Network recruited families from the New York area through community outreach, including parents concerned about a child’s mental health or learning. Its broad mix of difficulties was deliberately suited to studying patterns that span conventional diagnoses.[5]

This was observational research. Whether the profiles improve treatment selection requires prospective testing with clinical outcomes.[3]

Report an error or suggest a correction