DEEP LEARNING FOR EARLY DETECTION OF ALZHEIMER'S DISEASE USING OPEN ACCESS NEUROIMAGING DATASETS

Abstract

Alzheimer’s disease (AD) is one of the most common neurodegenerative disorders, affecting millions of people worldwide and causing progressive decline in memory, cognitive abilities, and daily functioning. Prompt diagnosis of Alzheimer's disease is crucial for timely intervention, better patient management and slowing the progress of the disease. The traditional diagnostic techniques include clinical evaluations and manual examination of neuroimage information; they are time-consuming and may not be able to detect subtle changes in the brain at an early stage. The recent developments of deep learning have shown great promise in extracting complex patterns from medical images automatically and making the determination of Alzheimer's disease more accurate. This study focuses on the use of deep learning methods to detect and classify Alzheimer's disease in open-access neuroimaging databases early. To develop and test deep learning models, publicly available datasets, such as MRI-based brain imaging datasets, are used. The proposed approach includes image pre-processing of neuroimages, feature extraction, training of the model and performance evaluation of the model by using conventional metrics like accuracy, precision, recall, F1-score, and area under curve (AUC). The architectures of deep learning are investigated because they provide the ability to automatically extract meaningful spatial features from brain images without the need for a lot of manual feature engineering. The purpose of the study is to understand the ability of deep learning models to differentiate between various cognitive impairments, such as normal control, MCI and Alzheimer's Disease. Moreover, the research highlights the challenges of data variability, model interpretability and clinical implementation. The results show the feasibility of deep learning-based methods as a reliable decision-support tool to diagnose early Alzheimer's disease and offer perspective on future advances in AI-based neuroimaging interpretation.