Deep neural networks for the early diagnosis of dementia and Alzheimer's disease from MRI images

Early diagnosis methods of Alzheimer's disease seem to be necessary due to the high costs of care and treatment, the indeterminacy of existing treatment methods, and the worrying future of the patient. This study was conducted in order to diagnose Alzheimer's disease from MRI images using artificial intelligence. In this study, a computer system for early diagnosis of Alzheimer's disease with the help of machine learning algorithms is presented in the framework of the computer-aided diagnosis process. Conditional random field and Inception deep neural network have been adapted to diagnose this disease on brain MRI images. Since the hippocampus tissue is one of the first tissues to be affected by Alzheimer's disease; Therefore, for the early diagnosis of this disease, first, the hippocampus was determined between other brain tissues and then, according to the extent of this tissue being affected, the disease was diagnosed. Conditional random field was able to extract hippocampus pieces with different shapes from all three brain sections with great accuracy. These pieces are the basis for feature extraction by the deep network. This method was tested on ADNI standard data and its efficiency was shown. The Inception network used is a network pre-trained on the very large ImageNet dataset. One of the important steps is to transfer knowledge to the problem at hand. To facilitate this, data augmentation designed according to the shape and structure of the hippocampus was used. The method implemented in this study achieved 98.51% accuracy in the case of Alzheimer's two-class versus healthy control and 93.41% for the two-class case of mild cognitive impairment versus healthy control, which is an increase of 2.56% and 8.41% respectively. It is compared to competing methods introduced in other articles. The results of this study showed that the use of artificial intelligence according to MRI images is highly accurate in diagnosing Alzheimer's disease.