Harnessing Machine Learning for Early Detection of Alzheimer’s Disease

Alzheimer’s Disease (AD) is a degenerative neurological condition that results in the deterioration of cognitive function. Early non-invasive detection is crucial for implementing timely interventions and slowing disease progression. In this research, we aimed to analyse the predictivity for classifying AD stages by integrating neuroimaging data—diffusion tensor imaging (DTI) and structural MRI (sMRI)—with clinical biomarkers. In this paper Fractional Anisotropy (FA) scans, derived from DTI, denoting water molecule diffusion in brain tissue was assessed, particularly in identifying early microstructural changes during the mild cognitive impairment (MCI) stage of AD. To progressively monitor the structural pathology of AD in patients, OASIS sMRI dataset was used. Convolutional neural networks (CNNs) like ResNet 3D and LeNET were applied to analyse FA and sMRI data, achieving a significant accuracy score. In this proposed work, correlation and permutation analysis was also applied on the blood biomarker data to extract relevant features. Random Forest classifier was then used on these biomarkers. The highly correlated biomarkers were later used to obtain a significant impact in identifying the various stages of Alzheimer’s progression.