Researchers have demonstrated that machine learning may help identify people at risk of developing Fragile X-related tremor/ataxia syndrome before obvious symptoms appear.
Why early prediction of Fragile X is important
Fragile X-associated tremor/ataxia syndrome (FXTAS) is a progressive neurodegenerative disease that affects some male Fragile X premutation carriers later in life. Currently, clinicians lack reliable tools to predict who will develop the disease or when symptoms will appear. Early identification of risk may improve monitoring, planning, and future preventive treatment strategies for fragile X carriers.
Machine learning methods in the Fragile-X cohort
This preliminary longitudinal analysis evaluated 103 male participants, including 72 Fragile X premutation carriers with a mean age at enrollment of 60.4 years and 31 healthy controls with a mean age at enrollment of 57.8 years. Across a total of 299 visits, researchers analyzed neuropsychological tests, motor assessments, brain MRI findings, and health indicators.
We compared multiple machine learning model and feature selection combinations to identify the best approach for two tasks: detecting existing FXTAS and predicting the emergence of future FXTAS among Fragile X carriers. The researchers developed a random forest binary classifier using selected variables such as age, psychiatric symptoms, executive function, motor measures, IQ, BMI, and structural brain measures. The data was randomly split into multiple training and testing sets, and the average classification performance metrics were evaluated across the splits. The number of completed follow-up conversion results is not reported separately.
Results highlight clinical and MRI risk signals
This model showed promising ability to identify current FXTAS cases and preemptively predict their subsequent emergence while maintaining a reasonable balance between precision and recall. In Fragile Structural brain MRI measurements have significantly increased predictive power beyond clinical variables alone.
Implications for care and future research
These findings suggest that machine learning may be a valuable tool for early Vulnerability X risk stratification, enabling proactive neurological monitoring and tailored lifestyle interventions before symptoms develop. The authors note important limitations and preliminary status, which means larger validation studies are needed.
reference
Gupta C et al. Using machine learning to identify risk markers for fragile X-associated tremor/ataxia syndrome: A preliminary analysis. Anne neurol. 2026;DOI:10.1002/ana.78217.
Featured image: Happy with Adobe Stock
