EAD-DNN: Early Alzheimer's disease prediction using deep neural networks

Abstract

Early Alzheimer’s disease (EAD) diagnosis enables individuals to take preventative actions before irreversible brain damage occurs. Memory and thinking skills get worse in alzheimer disease, making it hard to do basic things. The abnormal buildup of amyloid and tau proteins in and around brain cells is thought to cause it. When amyloid builds up, it forms plaques around brain cells. Inside brain cells, tau tangles form when it accumulates. Healthy brain cells are damaged by the tangles and plagues, which causes them to shrink. The hippocampus, a part of the brain that aids in memory formation, appears to be the location of this damage. There are currently no methods that give the most accurate results and suggestions. With the methods we have now, alzheimer disease is not found early. So, we said that the Early Alzheimer’s disease - Deep Neural Network (EAD-DNN) method has found a way to predict alzheimer disease earlier. The Magnetic Resonance Imaging (MRI) dataset in the Comma Separated Value (CSV) format has been used by the EAD-DNN method. Convolutional Neural Network (CNN), the deep Residual Network (ResNet) has been used to train the MRI image dataset. This ResNet model can get more information from network levels with the help of Deep ResNet.The modified adam optimization has selected the best feature information from MRI scans of alzheimer patients and transferred it to another area while keeping the most important data. Using the EAD-DNN approach, a multi-class classification has been carried out. The extensive experiments show that the suggested method can achieve an accuracy rate of 98%.