Alzheimer’s Disease (AD) is the most severe type of brain disorder found mainly in people 60 years of age or over.The latest developments on Multimodal Neuroimaging(MN) data have allowed the identification of the disease in life that was a breakthrough in neurosciences. Early diagnosis of AD is essential for the progress of more prevailing treatments.Recently, substantial focus has been paid to applying Deep Learning(DL) to early identification and automatic diagnosis of AD as rapid progress in neuroimaging techniques has generated large-scale MN data.
To save the overall time and energy of the doctors and increase their effectiveness in saving the lives of the sufferers, we suggest a classification of the AD using deep neural networks such as CNN,ANN and MLP.This research is aimed to provide an overview and critical assessment for the early identification of the AD to help the physicians in providing appropriate treatment to the patients based on the four different classes which we have considered in our dataset.
We found out that CNN was the most optimum neural network for this classification task as it outperformed other neural networks by a vast margin, with a test accuracy of 0.91. To add to that, we also discovered that ANN and MLP performed quite similarly and gave us an accuracy score of around 0.77 and 0.75