||Classification of Alzheimer Disease by Combining Hippocampus Volume and Cortical Thickness from sMRI, Plus APoE ε4 and CSF Biomarkers
||(Uttam Khatri) ; (Goo-Rak Kwon)
|| sMRI; Alzheimer’s disease; SVM-RFE; SVM; NB; KNN; Mild cognitive impairment; CSF; Genetics; Hippocampus
||sMRI measurement is important for characterizing the pathology of Alzheimer’s disease (AD), mild cognitive impairment (MCI), and healthy control (HC). To date, several imaging and non-imaging bio-markers for AD and MCI have been identified. Cortical thickness, hippocampal atrophy, apolipoprotein E gene ε4 (APoE ε4), and cerebrospinal fluid (CSF) biomarkers are believed to be the major indicators for AD and MCI. In this paper, these features have been utilized successfully to identify AD patients from controls. These biomarkers have mostly been used separately, so far. The full possibilities of combining sMRI, cortical thickness, hippocampal volume, APoE ε4, and CSF biomarkers for AD diagnosis might thus yet lead to optimal analysis. Therefore, we combined hippocampal volume, cortical thickness, APoE ε4, and CSF markers to enhance diagnostic classification of AD. For 53 clinically diagnosed AD patients, 103 patients with mild cognitive impairment, and 61 cognitively healthy controls, we obtained cortical thickness, hippocampal volume, APoE ε4, and CSF biomarkers. These four measures were first applied separately and were then combined to predict AD in support vector machine?recursive feature elimination (SVM-RFE) to select the optimal features. They were fed into different classifiers (naive Bayes, k-nearest neighbors, and SVM), and experimental results show that the combination of the different biomarkers performs well, as compared to using separate features individually.