A Machine Learning Algorithm to Detect Alzheimer’s Disease Using MRI Images

Alzheimer's disease, a prevalent form of dementia, stands as a well-established neurodegenerative condition marked by the gradual decline of brain cells and neurons. Its challenging nature arises from the absence of surgical or medicinal remedies. Typically, individuals with AD face a maximum survival span of 3 to 4 years. Notably, during the initial stages of the disease, symptoms often remain undetectable. Professionals rely on the Mini-Mental State Examination (MMSE) and comprehensive familial history to diagnose dementia in its early phases, a task fraught with inherent complexities. Nevertheless, early detection and the deceleration of AD progression remain feasible objectives. Various neuroimaging methods, including Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), and Computerized Tomography (CT) scans, offer significant aid in promptly identifying Alzheimer's disease. This study employs MRI images to categorize AD stages using a Machine Learning approach, incorporating a proposed feature extraction technique. Addressing a substantial class imbalance within the MRI dataset, the SMOTE technique is utilized for dataset balancing before the implementation of ML algorithms. Notably, the proposed model outperforms contemporary models concerning precision, accuracy, recall, and the F1 score.