Exploring Neural Network Approaches for Alzheimer's Disease Detection: an Analysis of RNN and CNN Performance

Alzheimer's disease is a degenerative neurological ailment marked by memory loss, cognitive decline, and behavioral disturbances. It is the most common type of dementia, affecting millions of people globally, primarily the elderly. The disease is characterized by the formation of plaques and tangles in the brain, which disrupt neuronal communication and eventually lead to cell death. In this study, we analyze the use of machine learning models in the context of Alzheimer's disease, which were CNN and RNN. The RNN is a deep learning model that has been trained to interpret and transform sequential data inputs into particular sequential data outputs. However, CNNs are deep learning models that excel at identifying spatial correlations in data, which makes them a good fit for image analysis tasks like analyzing MRI scans, which are frequently utilized in AD research. We used datasets from the OASIS, which offers publicly available brain imaging datasets for Alzheimer's disease research, and from Kaggle, a well-known platform for data science competitions and cooperation, to train and assess these models. Our experimental results demonstrate the promising performance of the models on both datasets. During our work, On the OASIS dataset, CNN achieved 100% accuracy, whereas RNN achieved 73.15%. On the Kaggle dataset, the CNN model achieved 97.33% accuracy, while the RNN model achieved 80.67%.