Identification of Amnestic Mild Cognitive Impairment Using Multi-Modal Brain Features: A Combined Structural MRI and Diffusion Tensor Imaging Study

Article type: Research Article

Authors: Xie, Yunyan^{a; 1} | Cui, Zaixu^{b; 1} | Zhang, Zhongmin^{c; 1} | Sun, Yu^a | Sheng, Can^a | Li, Kuncheng^d | Gong, Gaolang^{b; *} | Han, Ying^{a; e; *} | Jia, Jianping^{a; e; f; *}

Affiliations: [a] Department of Neurology, Xuan Wu Hospital of the Capital Medical University, Beijing, China | [b] State Key Laboratory of Cognitive Neuroscience and Learning & IDG/Mc Govern Institute for Brain Research, Beijing Normal University, Beijing, China | [c] Department of Neurology, Hongqi Hospital, Mudanjiang Medical College, Mudanjiang, Heilongjiang, China | [d] Department of Radiology, Xuan Wu Hospital of the Capital Medical University, Beijing, China | [e] Center of Alzheimer’s Disease, Beijing Institute for Brain Disorders, China | [f] Beijing Key Laboratory of Geriatric Cognitive Disorders and Neurodegenerative Laboratory of Ministry of Education of the People’s Republic of China, Beijing, China

Correspondence: [*] Correspondence to: Jianping Jia, Department of Neurology, Xuan Wu Hospital of the Capital Medical University,Beijing 100053, China. Tel.: +86 10 83198730; Fax: +86 10 83171070; [email protected]

Correspondence: [*] Correspondence to: Ying Han, Department of Neurology, Xuan Wu Hospital of the Capital MedicalUniversity, Beijing 100053, China. Tel.: 13621011941; [email protected]

Correspondence: [*] Correspondence to: Gaolang Gong, PhD, State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Research, Beijing Normal University, Beijing 100875, China. Tel.: +86 1058804678; Fax: +86 1058806154; [email protected]

Note: [1] These authors contributed equally to this work

Abstract: Identifying amnestic mild cognitive impairment (aMCI) is of great clinical importance because aMCI is a putative prodromal stage of Alzheimer’s disease. The present study aimed to explore the feasibility of accurately identifying aMCI with a magnetic resonance imaging (MRI) biomarker. We integrated measures of both gray matter (GM) abnormalities derived from structural MRI and white matter (WM) alterations acquired from diffusion tensor imaging at the voxel level across the entire brain. In particular, multi-modal brain features, including GM volume, WM fractional anisotropy, and mean diffusivity, were extracted from a relatively large sample of 64 Han Chinese aMCI patients and 64 matched controls. Then, support vector machine classifiers for GM volume, FA, and MD were fused to distinguish the aMCI patients from the controls. The fused classifier was evaluated with the leave-one-out and the 10-fold cross-validations, and the classifier had an accuracy of 83.59% and an area under the curve of 0.862. The most discriminative regions of GM were mainly located in the medial temporal lobe, temporal lobe, precuneus, cingulate gyrus, parietal lobe, and frontal lobe, whereas the most discriminative regions of WM were mainly located in the corpus callosum, cingulum, corona radiata, frontal lobe, and parietal lobe. Our findings suggest that aMCI is characterized by a distributed pattern of GM abnormalities and WM alterations that represent discriminative power and reflect relevant pathological changes in the brain, and these changes further highlight the advantage of multi-modal feature integration for identifying aMCI.

Keywords: Alzheimer’s disease, amnestic mild cognitive impairment, classification, diffusion tensor imaging, structural magnetic resonance imaging, support vector machine

DOI: 10.3233/JAD-150184

Journal: Journal of Alzheimer's Disease, vol. 47, no. 2, pp. 509-522, 2015