Elahe Jozpoor - Medical Informatics, Shahid Beheshti University of Medical Sciences, Tehran, Iran
Sara Yousefi Javan - Computer Engineering, Islamic Azad University of Mashhad, Mashhad, Iran

The purpose of this paper is to utilize convolutional neural networks to detect Alzheimer's disease (AD) in comparison to mild cognitive impairment (MCI) and normal control (NC). It has become increasingly important to diagnose AD in recent years because of the increase in life expectancy around the world. As a result of MCI, the patient's mental abilities are irreversibly impaired, which can lead to Alzheimer's disease and other forms of dementia. In order to stop its progression, and to treat it, this disorder has received special attention from many researchers. Biochemical tests and psychological tests are commonly used to diagnose the disease. In order to diagnose Alzheimer's disease, magnetic resonance imaging (MRI), which studies changes in the structure of the human brain, is one of the proposed approaches. The purpose of this paper is to preprocess brain magnetic resonance images (MRIs) with the use of the SPM toolbox, then segment the brain's gray matter (GM) and feed it into the CNN algorithm. The ADNI dataset is used in this article. With an accuracy of over ۹۹% in this test, we were able to classify the three categories of standard control (NC), Alzheimer's disease (AD), and mild cognitive impairment (MCI).

Alzheimer’s disease, Mild Cognitive Impairment (MCI), brain Magnetic Resonance Imaging (MRI), Normal Control (NC), convolutional neural network (CNN).