Leveraging Machine Learning for Early Detection of Neurodegenerative Disorders Among the Elderly Using Diffusion Tensor Imaging (DTI)

Background and Aims: A promising tool for understanding age-related neurodegeneration is diffusion tensor imaging (DTI) of the brain's white matter. Researchers are using machine learning to improve the early detection of neurological disorders such as Parkinson's. The study examines the effectiveness of machine learning-based DTI applications in increasing diagnostic accuracy, identifying markers, and enabling early interventions in older people. Methods: In this review study, a review of English language articles from January ۲۰۱۹ to November ۲۰۲۴ with AND, OR operators and English keywords Having the full text of the article, the year of publication between ۲۰۱۹ and ۲۰۲۴, Clinical trial studies, and Human studies with Rayyan software were done through scientific databases including PubMed, Web of Science, Science Direct, Cochrane Library, APA PsycNET, and Google Scholar. A total of ۲۳,۵۵۹ articles with related keywords were extracted. Then, after checking and considering the entry and exit criteria, ۲۱ articles were selected and included in the study. Results: Machine learning models demonstrated varied performance across studies, with Fractional Anisotropy (FA) emerging as the most reliable DTI metric for identifying white matter alterations. For instance, support vector regression using FA achieved a mean absolute error of ۷.۷۴–۱۰.۵۴ years for brain age prediction. Support Vector Machine classifiers achieved an accuracy of up to ۸۷.۵% for differentiating patients with idiopathic Rapid Eye Movement Sleep Behavior Disorder from healthy controls. Similarly, structural connectome profiles combined with Machine Learning (ML) models enhanced classification accuracy for focal epilepsy but showed limited efficacy in predicting Ant-seizure Medication responsiveness. Early detection of neurodegenerative conditions like Parkinson’s Disease remains challenging, with diagnostic performance ranging from ۶۷% to ۸۷% depending on the input features and ML algorithms utilized. Conclusion: the combined use of DTI and ML shows promise in detecting and categorizing neurodegenerative disorders in older adults. FA is superior in detecting age-related and pathological changes. Integration of various imaging techniques improves accuracy but challenges remain. Future research should focus