Back to top anchor

Domain-adapted AI for equitable brain-based cognitive assessment in ADHD

Year:
2026
Duration:
24 months
Approved budget:
$150,000.00
Researchers:
Dr Jack Scott
,
Dr Narun Pat
,
Associate Professor Byron Rangiwai
,
Dr Dione Healey
,
Professor Paul Glue
,
Associate Professor Jeremiah Deng
Host:
University of Otago
Health issue:
Mental health (and sleep disorders)
Proposal type:
Explorer Grant
Lay summary
ADHD assessment in children currently relies on behavioural tests and parent reports, which cannot easily distinguish whether a child has genuine cognitive limitations or simply performed poorly due to anxiety, fatigue, or cultural discomfort with testing. Artificial intelligence (AI) may be able to resolve this, but tools trained on overseas predominantly European populations risk performing poorly for Māori children if deployed in New Zealand without local adaptation. We aim to develop a brain-based assessment tool using AI and electroencephalography (EEG), which measures brain activity through sensors placed on the scalp. By training our AI on international data and then adapting it using specialist AI techniques on census-matched New Zealand samples with equitable Māori representation, we can create AI-powered cognitive assessment tools responsive to our New Zealand context, without requiring thousands of participants. If successful, this cost-effective tool could provide objective brain-based assessment to support clinical decision making, deployable outside of main centres.