北京生物医学工程

不同大脑分区下注意缺陷多动障碍分类研究

Classification of ADHD children under different kinds of brain partition

作者：叶子骏黄惠芳刘杰

单位：北京交通大学(北京100044)

分类号：R318.04

出版年·卷·期（页码）：2016·35·4（389-395）

摘要：

目的注意缺陷多动障碍(attention deficit hyperactivity disorder, ADHD)是一种常见的影响儿童行为能力的精神疾病，但由于缺乏有力的客观依据判断发病机制，因而临床上诊断与治疗存在一定难度。从脑功能研究ADHD的发病机制是一个热点，特别是从基于感兴趣区域提取特征进行分类研究。为方便找出感兴趣区域，研究者们根据解剖或功能将大脑进行区域划分，常用的脑分区有6种，但还没有研究证明哪一种脑分区最适用于ADHD分类研究。因此本文将对不同大脑分区下注意缺陷多动障碍分类研究，以判断哪种分类效果好。方法本实验采用ADHD-200大赛的数据样本，将数据分为12组，每种大脑分区下包含一组训练集和一组测试集。首先基于静息态fMRI对6种不同大脑分区的训练集进行功能连接计算、特征分类、特征选择，然后利用测试者检验得到6种分类模型的分类性能并做统计分析。结果实验结果显示每种大脑分区下的训练集得到的分类准确率，并比较每种大脑分区下的测试集的分类性能，综合分析总结出AAL大脑分区的分类性能最好，分类准确率达到63.16%。结论在6种大脑分区下，AAL大脑分区是目前最适合用于研究ADHD的大脑分区方法。

Objective Attention deficit hyperactivity disorder (ADHD) is one of the most common diseases in school-age children, yet it is difficult to diagnose and treat because of lacking objective evidence to judge the pathogenesis. To study the pathogenesis of ADHD from the aspect of brain function is a hot spot, especially the research of classification based on the region of interest. In order to find useful region of interest, researchers divide the brain into certain regions based on anatomy or function. There are six kinds of common brain regions, yet no one proves which method is the best for ADHD classification. So this paper studies on ADHD classification under the six brain partitions in order to find the best. Methods The data come from the ADHD-200 competition including train data and test data. Data were divided into twelve groups, each containing a set of training sets and a set of test sets. Firstly, based on the resting state fMRI we make functional connectivity analysis, feature classification and feature selection under six brain partitions by using train data. Then we use the test data to test the classification effect of the classification models and analyze the results. Results The experimental results show that the classification accuracy of the training set is obtained under different brain regions, and then the classification performance of the test sets is compared. In the overall analysis of the classification performance of the six sets data, AAL brain regions has the best performance, and the accuracy is 63.16%. Conclusions In all the six kinds of brain partition, AAL brain partition is the most suitable method for the study of ADHD classification.

参考文献：

［1］Smith SM, Miller KL, Salimi-Khorshidi G, et al. Network modelling methods for FMRI［J］. NeuroImage,2011,54 (2) :875-891.

［2］钱秋瑾, 杨莉, 王玉凤. 儿童注意缺陷多动障碍的研究进展［J］.北京大学学报:医学版, 2007, 39(3):323-328.

Qian Qiujin, Yang Li, Wang Yufeng. Advances on comprehensive research on attention deficit hyperactivity disorder［J］. Journal of Peking University: Health Sciences, 2007, 39(3):323-328.

［3］Cheng Wei, Ji X, Zhang J, et al. Individual classification of ADHD patients by integrating multiscale neuroimaging markers and advanced pattern recognition techniques［J］. Frontiers in Systems Neuroscience, 2012,6:68.

［4］Dey SB, Rao AR, Shah M. Exploiting the brain’s network structure in identifying ADHD subjects［J］. Frontiers in Systems Neuroscience,2012,6:75.

［5］Fair DA, Nigg JT, Iyer S, et al. Distinct neural signatures detected for ADHD subtypes after controlling for micro-movements in resting state functional connectivity MRI data［J］. Frontiers in Systems Neuroscience, 2013,6:80.

［6］Sato JR, Hoexter MQ, Fujita A, et al. Evaluation of pattern recognition and feature extraction methods in ADHD prediction［J］. Frontiers in Systems Neuroscience, 2012,6: 68.

［7］Sidhu GS, Asgarian N, Greiner R, et al. Kernel principal component analysis for dimensionality reduction in fMRI-based diagnosis of ADHD［J］. Frontiers in Systems Neuroscience, 2012,6:74.

［8］Bohland JW, Saperstein S, Pereira F, et al. Network, anatomical, and non-imaging measures for the prediction of ADHD diagnosis in individual subjects［J］. Frontiers in Systems Neuroscience, 2012,6:78.

［9］Colby JB, Rudie JD, Brown JA, et al. Insights into multimodal imaging classification of ADHD［J］. Frontiers in Systems Neuroscience, 2012,6:59.

［10］Zou QH, Zhu CZ, Yang Y, et al. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF［J］. J Neurosci Methods, 2008,172(1):137-141.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录