An artificial intelligence model can predict cognitive decline in multiple sclerosis (MS) patients with 90% accuracy by integrating brain scans with clinical and demographic data. According to a study published in the European Journal of Neurology, researchers in Milan successfully identified key biological predictors of neurocognitive disorders, providing a potential framework for personalized monitoring in clinical settings.
How does the AI model predict cognitive outcomes?
The AI system functions by analyzing a multimodal dataset collected at the start of a patient’s evaluation. Researchers fed the model MRI brain images alongside clinical metrics, including disability scores, disease duration, and cognitive reserve—a measure of the brain’s ability to withstand damage. By identifying patterns across these variables, the model classified patients into stable or worsening groups with 90% accuracy during internal validation, according to the study “Explainable Artificial Intelligence to Predict Neurocognitive Disorder Progression in Multiple Sclerosis Using MRI and Clinical Data.”
Cognitive reserve acts as a buffer against MS-related damage. Patients with higher levels of education and intellectual engagement often show better cognitive performance despite physical evidence of brain lesions on MRI scans.
Which factors most influence cognitive worsening?
Using explainability tools, researchers pinpointed the specific features the AI prioritized when flagging patients at risk of decline. The most influential factors, in descending order of importance, were cortical gray matter volume, patient age, thalamus and hippocampus volume, MS-related lesion volume, and cognitive reserve. These findings align with existing clinical understanding, as the thalamus and hippocampus are primary regions linked to memory and learning, which are often affected by the immune system attacks characteristic of MS.
How does this compare to traditional diagnostic methods?
Current diagnostic frameworks categorize cognitive changes as mild or major neurocognitive disorders (NCDs) based on functional independence. However, these methods often lack predictive power regarding future progression. While traditional screening relies on periodic neuropsychological testing, the AI model offers a proactive approach. In the Milan study, 4% of participants met criteria for mild NCD and 11% for major NCD at baseline. By the follow-up period, 40% of those initially diagnosed with mild NCD had progressed to major NCD, suggesting that current clinical monitoring may miss the window for early intervention.
If you are managing MS, discuss the frequency of cognitive assessments with your neurologist. Early identification of mild NCD can help clinicians adjust disease-modifying therapies before major functional impacts occur.
What are the implications for personalized MS care?
The integration of explainable AI into clinical workflows could shift MS treatment from reactive to predictive. Traditional models often rely on a single data point, such as lesion load alone, which researchers noted provides an incomplete picture. By combining MRI data with demographic factors like age and education, this multimodal approach allows for a more nuanced assessment. The authors of the study concluded that the AI’s ability to highlight specific brain regions associated with decline underscores its potential for personalized patient monitoring.
Frequently Asked Questions
- Can this AI model be used in a doctor’s office today?
Not yet. While the model showed high accuracy in a research setting, it requires further large-scale validation before it can be implemented as a standard clinical diagnostic tool. - What is “cognitive reserve”?
It is the brain’s ability to maintain function despite the presence of physical damage, often bolstered by education, professional complexity, and lifelong learning. - Does the AI replace an MRI scan?
No. The AI actually requires MRI brain images as a primary input to generate its predictions.
Are you interested in how digital health tools are changing chronic disease management? Subscribe to our newsletter for the latest updates on medical technology or leave a comment below to share your experiences with cognitive health monitoring.
