Machine learning read the shape of sick brain cells and picked out nine already-approved drugs that calmed them down
Researchers led by Flinders University in Australia published a study in Nature Communications on 20 August describing a drug screen for Sanfilippo syndrome, an inherited childhood dementia that progressively takes away memory, behaviour and physical ability and has no widely available treatment. The team grew brain cells from affected children; earlier work had shown those cells become abnormally excitable early in development and get stuck in overdrive. They then tested 63 medicines already approved by Australia's Therapeutic Goods Administration on the cells, using imaging and machine learning to score which compounds shifted diseased cells back towards a healthy state. Nine significantly improved cell function within two weeks: some reduced damage to the cells, others restored the signalling associated with learning and behaviour, and combinations worked better than single drugs. The paper — "Drug screen and machine learning predict neuroprotective agents in a preclinical human model of childhood dementia", by Zarina Greenberg and colleagues — is explicit that this is a laboratory model. None of the nine has been shown to be safe or effective for childhood dementia in patients.
Related stories
Given four hours and a GPU to improve the way AI is trained, the best agent scored 0.25 out of 1 — and most never tried2026-08-21The agent invented a second person to vouch for its code. A 24-year-old in Texas refused to believe either of them.
2026-08-21Pew put half a million web pages through a detector: a third of everything published since ChatGPT carries its marks
2026-08-21The FDA has cleared 1,357 AI medical devices. Three of them have been tested on whether patients live longer or better.
2026-08-20So many machine-written proofs arrived that mathematicians built a filing office — and one of its two checks is done by a language model
2026-08-20