Abnormal Changes of Multidimensional Surface Features Using Multivariate Pattern Classification in Amnestic Mild Cognitive Impairment Patients

• DOI: 10.1523/jneurosci.4356-13.2014
• Publication Date: 2014-08-06T16:28:56Z
• Authors: S. Li; X. Yuan; F. Pu; D. Li; Y. Fan; L. Wu; W. Chao; N. Chen; Y. He; Y. Han
• Keywords: Human brain; Univariate

Previous studies have suggested that amnestic mild cognitive impairment (aMCI) is associated with changes in cortical morphological features, such as thickness, sulcal depth, surface area, gray matter volume, metric distortion, and mean curvature. These features been proven to specific neuropathological genetic underpinnings. However, most primarily focused on mass-univariate methods, were generally explored isolation. Here, we used a multivariate method characterize the complex subtle structural changing pattern of anatomy 24 aMCI human participants 26 normal controls. Six extracted for each participant, spatial patterns brain abnormities identified by high classification weights using support vector machine method. The accuracy discriminating two groups was 76% left hemisphere 80% right when all six used. Regions showing subtle, spatially complex, predominately located medial temporal lobe supramarginal inferior parietal lobes. In addition, also found had different contributions even same region. Our results indicated neuroanatomical discriminated individuals from controls truly multidimensional effects features. Furthermore, regions our could potentially be useful clinical diagnosis.