DEEP LEARNING BASED PREDICTION OF ALZHEIMER’S DISEASE PROGRESSION USING MAGNETIC REASONANCE IMAGING AND CLINICAL DATA WITH U-NET IMAGE SEGMENTATION

ABSTRACT Accurate prediction of Alzheimer’s Disease (AD) progression remains a critical challenge in medical diagnosis and patient management. This study presents a deep learning-based framework that integrates Magnetic Resonance Imaging (MRI) with clinical data to predict the progression of Alzheimer’s Disease. A U-Net-based convolutional neural network was employed for precise segmentation of brain structures associated with AD pathology, particularly focusing on regions such as the hippocampus and ventricles. The segmented MRI features were then combined with clinical variables including cognitive scores, demographic information, and genetic markers—and input into a multi-layered neural network for disease stage classification and progression forecasting. The model was trained and validated on the ADNI dataset, achieving high accuracy in distinguishing between cognitively normal, mild cognitive impairment (MCI), and AD subjects. The proposed framework outperforms conventional machine learning baselines in both segmentation accuracy and progression prediction, indicating its potential for early diagnosis, monitoring, and personalized intervention planning in Alzheimer’s Disease. This approach highlights the synergy of medical imaging and deep learning in advancing neurodegenerative disease research.