A VNS based framework for early diagnosis of the Alzheimer's disease converted from mild cognitive impairment

Mild cognitive impairment (MCI) is an intermediate stage between age-related cognitive decline. Alzheimer's disease (AD) is a more serious decline in dementia. Early identification of mild cognitive impairment with a high risk of Alzheimer's disease is very important for increasing the success rate of the treatment. In this study, we present a Variable Neighborhood Search (VNS) based framework that uses Magnetic Resonance Imaging (MRI) data to diagnose early conversion from MCI to AD. The proposed framework has been built in three main phases: preparing dataset, feature selection, and classification. After preparing the dataset, a VNS algorithm selects the most predictive MRI features for classification. Then, a Linear Support Vector Machine is utilized to classify the selected features. All data in this study are obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database with 860 subjects, eight different monthly periods, and 286 features in each period. The results obtained from the framework outperform those of previous research in terms of accuracy, sensitivity, and specificity values. The results of this study demonstrate that our framework has a huge potential for early prediction and detection of mild cognitive impairment to Alzheimer's disease conversion.