NNI Neurologist Uses Machine Learning to Predict How ALS Progresses — and Singapore's Data Is Different
Source: The Straits Times
An NNI neurologist is using machine learning to predict how ALS progresses in different people, with models validated on Singapore's own population. Local data tells a different story from the West: earlier onset, but patients living about twice as long with the disease.
ALS is one of the cruelest diseases to predict. Also known as Lou Gehrig's disease, it destroys the nerve cells controlling movement, and no two patients progress the same way — one might decline rapidly while another holds steady for years. In Singapore, between 250 and 400 people are living with the condition, and specialist centres like the National Neuroscience Institute diagnose 30 to 50 new cases every year. That unpredictability is exactly what NNI neurologist Crystal Yeo is attacking with machine learning.
Yeo's team built models that comb through large volumes of clinical data — blood test results, ALS monitoring scores measuring muscle strength and breathing, nerve conduction findings, even proteins and genes from damaged motor neurons grown out of patients' own stem cells — to identify the strongest predictors of how the disease will progress. Crucially, the models were validated not just on international data but on Singapore's population, so the predictions hold locally as well as globally. The findings were published in Muscle & Nerve, a peer-reviewed neurology journal, in 2025.
The Singapore data matters because it differs from the West in meaningful ways. Local research has found the disease tends to strike five to 10 years earlier here than in Caucasian populations, and patients in Singapore and Asia generally live about twice as long with the condition. The type of ALS also varies, sometimes moving between hands, feet and tongue. Yeo's broader point is that ALS is not really one disease but a group of them with different underlying biology, which is why the field is moving toward precision medicine: targeting therapies to the biology of each patient's form of the illness.
The same approach is being extended to treatment — AI that analyses stem-cell-derived motor neurons to predict which drugs or drug combinations are likely to work best for different patient groups. For patients and families, better prediction is not an academic exercise; it changes how lives get planned. Huang Jing Han, diagnosed in 2017 at 33, now runs a small business, answers customer enquiries via WhatsApp and uses an AI-powered eye-tracking app for technical analysis in day trading.
Why it matters for Singapore: This is precision medicine in action at a Singapore institution, using local patient data that turns out to be genuinely different from Western cohorts — a reminder that models trained abroad do not automatically fit here. As Asia's population ages and ALS becomes more prevalent, NNI's work positions Singapore as a place where AI is applied not just to finance and logistics, but to the hardest problems in healthcare.
