Researchers have developed machine learning models capable of identifying individuals at risk for Fragile X-associated tremor/ataxia syndrome (FXTAS) before overt symptoms emerge.
FXTAS is a progressive neurodegenerative condition impacting some male fragile X premutation carriers later in life. Current diagnostic tools lack the ability to reliably predict onset or progression. Earlier risk identification could significantly enhance monitoring and pave the way for future preventive treatments.
The preliminary study analyzed 103 male participants, including 72 fragile X premutation carriers, using neuropsychological testing, motor evaluations, brain MRI, and health metrics. Machine learning models, specifically random forest classifiers, were trained to detect existing FXTAS and predict its future emergence.
Key predictors identified include body mass index, executive function deficits, slower reaction times, reduced dexterity, and mental health changes. Notably, structural brain MRI measurements substantially improved the models' predictive accuracy beyond clinical data alone.
These findings suggest that machine learning could become a crucial tool for early risk stratification in Fragile X carriers, facilitating proactive neurological surveillance and tailored interventions before symptoms manifest. Further validation through larger studies is recommended.