RESEARCHERS have demonstrated that machine learning could help identify people at risk of developing Fragile X-associated tremor/ataxia syndrome, before overt symptoms appear.
Why Earlier Fragile X Prediction Matters
Fragile X-associated tremor/ataxia syndrome, or FXTAS, is a progressive neurodegenerative condition affecting some male fragile X premutation carriers later in life. Clinicians currently lack reliable tools to predict who will develop the disorder or when symptoms may emerge. Earlier risk identification could improve monitoring, planning, and future preventive treatment strategies for Fragile X carriers.
Machine Learning Methods in Fragile X Cohort
This preliminary longitudinal analysis evaluated 103 male participants, including 72 fragile X premutation carriers with a mean enrolment age of 60.4 years and 31 healthy controls with a mean enrolment age of 57.8 years. Across 299 total visits, researchers analysed neuropsychological testing, motor evaluations, brain MRI findings, and health metrics.
Multiple machine learning models and feature-selection combinations were compared to identify optimal approaches for two tasks: detecting existing FXTAS and predicting future emergence of FXTAS among Fragile X carriers. Using selected variables including age, psychological symptoms, executive function, motor measures, IQ, body mass index, and structural brain measurements, investigators developed random forest binary classifiers. Data were randomly split into multiple training and testing sets, with average classification performance metrics assessed across splits. Completion numbers for follow-up conversion outcomes were not separately reported.
Results Highlight Clinical and MRI Risk Signals
The models showed promising ability to identify current FXTAS cases and pre-emptively predict later emergence while maintaining a reasonable balance between precision and recall. Among Fragile X carriers, higher body mass index, executive function weaknesses, slower reaction time, reduced dexterity, and mental health changes were associated with greater future risk. Structural brain MRI measurements significantly improved predictive power beyond clinical variables alone.
Implications for Care and Future Research
These findings suggest machine learning may become a valuable tool for earlier Fragile X risk stratification, allowing proactive neurological surveillance and tailored lifestyle interventions before symptoms develop. The authors note important limitations and preliminary status, meaning larger validation studies are required.
Reference
Gupta C et al. Use of machine learning to identify markers of risk for Fragile X-associated tremor/ataxia syndrome: a preliminary analysis. Ann Neurol. 2026;DOI:10.1002/ana.78217.
Featured image: happie on Adobe Stock
