One of the most accurate ways to spot cerebral palsy early is simply to watch how a baby moves. The General Movements Assessment (GMA) — a short observation of an infant’s spontaneous wriggling at around three to four months — can flag a high probability of cerebral palsy with remarkable accuracy. Its catch: it requires a specially trained, certified expert, and there aren’t enough of them.
Researchers think artificial intelligence could change that — by reading a baby’s movements from an ordinary smartphone video.
“AI has the potential to increase accessibility to early screening; health professionals will always need to be involved in the pathway to diagnosis.”
— Roadmap for automated GMA, eClinicalMedicine, 2025
Why automate the movement test
The absence of normal “fidgety” movements at three to four months has a sensitivity of about 95–98% for predicting cerebral palsy, making the GMA the most accurate early prediction tool available. But, as a 2025 roadmap in eClinicalMedicine notes, its use is “limited by the need for training and certification of assessors.” That bottleneck means many infants — especially in under-resourced areas — are never screened.
How AI could help
The approach is straightforward: a parent records a short video on a phone or tablet; software estimates the baby’s body movements (a step called “pose estimation”); and a machine-learning model flags movement patterns linked to cerebral palsy. An international consortium has published a roadmap to standardize the data, validate the models, and bring this to families worldwide — including in low- and middle-income countries. The authors are clear that AI is meant to widen access to screening, not replace clinicians: a concerning result still leads to a full assessment by a health professional. This pairs naturally with starting support early; see our guide to cerebral palsy diagnosis.
This article is for general education and is not medical advice. AI-based screening tools are still being developed and validated; talk with your child’s doctor about any concerns regarding development.
Sources
- “Towards universal early screening for cerebral palsy: a roadmap for automated General Movements Assessment.” eClinicalMedicine, 2025. ncbi.nlm.nih.gov
- “Automating General Movements Assessment with quantitative deep learning to facilitate early screening of cerebral palsy.” Nature Communications, 2023. nature.com
