A Cortical Based Diagnosis System for MCI Based on sMRI Features Fusion

Abstract

Alzheimer's disease (AD) is one of the most neurodegenerative disorders that target central nervous system with statistical results of more than 5 million sufferers among the Americans. According to the literature, discovering the disease in its early stage is considered as one of the main obstacles that face the scientists. The difficulty of this diagnostic task relied on a number of reasons including the variability of the disease's effect among its patients. This paper aims to study mild cognitive impairment (MCI), the type of impairment that found to increase the factor of achieving to AD. According to this study a cortical regions based computer-aided diagnosis (CAD) system can be presented that in turn serve the early diagnosis of AD. This goal is achieved by visualizing the personalized diagnosis of the MCI in each of the cortical regions separately. For this purpose, the proposed CAD system goes into four main stages: 1-preprocessing and cortex extraction, 2-cortex re-construction and shape-based feature extraction, 3-feature fusing, and 4-local/regional diagnosis followed by global diagnosis step. Evaluating the proposed system shows promising results with a maximum performance of 86.30%, 88.33%, and 84.88% for accuracy, specificity, and sensitivity, respectively.