基于网络拓扑特性的MCI分类  

Topological Property of Connectivity Network for MCI Classification

作　　者：接标[1] 张道强[1] 

机构地区：[1]南京航空航天大学计算机科学与技术学院,南京210016

出　　处：《数据采集与处理》2013年第5期602-607,共6页Journal of Data Acquisition and Processing

基　　金：江苏省杰出青年基金(SBK201310351)资助项目;高等教育博士学科点专项科研基金(20123218110009)资助项目;南京航空航天大学基本科研业务费专项资金(NE2013105)资助项目;安徽省高校省级自然科学研究(KJ2013Z095)资助项目

摘　　要：阿尔茨海默氏病(Alzheimer′s disease,AD)和轻度认知障碍(Mild cognitive impairment,MCI)是经常发生在老年人中的脑疾病,其主要表现为认知和智力的障碍。机器学习和模式识别方法已经被应用到对AD和MCI的诊断和分类中。最近,研究人员提出利用大脑连接网络实现对疾病的诊断和分类。大部分的研究主要集中在网络中提取一些局部的特性(如聚类系数),并利用机器学习的方法(如支持向量机)来实现对疾病的分类。然而,存在的研究表明AD以及MCI是和一个大规模的脑连接网络相关,而不仅是大脑的若干区域。因此,本文提出一种新的基于网络整体拓扑结构信息的分类方法,并将其用于对MCI疾病的分类。实验结果表明,本文的方法能够对分类结果有重要的改进。Alzheimer＇s disease （AD） as well as its prodromal stage,i.e.,mild cognitive impairment （MCI）,are one of the most prevalent dementia in older adults,which are characterized by cognitive and intellectual deficits.Machine learning and pattern recognition techniques have been applied to the study of AD/MCI.Recently,connectivity-network based classification methods are applied to the study of AD and MCI,i.e.,to identify the individuals with AD/MCI from the healthy controls （HC）.However,most existing methods focus on using some local properties of a connectivity network （e.g.,local clustering）,although other network properties,such as whole topological property,can potentially be used.Moreover,existing studies indicate that AD/MCI is associated with a large-scale highly-connected functional network,rather than a single isolated region.So,by employing kernel based approach,this paper proposes a novel connectivity network based method to extract and quantify whole topological information of connectivity network for improving the classification performance.The proposed method is evaluated on 12 MCI patients and 25 healthy controls,and promising experimental results are obtained.

关 键 词：图核 核方法 连接网络 轻度认知障碍 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]