Progress in
Natural Science

Editorial board
Instructions for Authors
Indexing information
Search by keywords
Current Issue
Advance Search
Search by issue
Times of Access
Times of Download


Visiting Times:
4274479
Visiting Times of Today:
289
Now Online
0

自然科学进展

中国科学基金

Science Foundation in China



Mean mutual information ratio: a new method for ranking components of ICA
Zuo Xinian, Zhu Chaozhe and Zang Yufeng
1. National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100080, China; 2. State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing 100875, China
Abstract:  This paper aims to develop a more robust measure of independency called mean mutual information ratio and propose a new way for ranking components with this measure. In order to evaluate the advantage of mean mutual information ratio, independent component analysis has been performed on task-related and resting-state functional magnetic resonance imaging data. And then, mean mutual information ratio was compared to kurtosis in ranking components. The result demonstrated that mean mutual information ratio gives more accurate and robust measure of independency.
Keywords:  independent component analysis, mutual information, default mode, fMRI
Abstract Hits:   2546
Download:   408
View Full Text  View/Add Comment  Download reader
Address by: Department of Publication/NSFC.
83 Shuangqing Road Haidian District,Beijing CHINA.
Post code: 100085
Tel: 86+10-62327204
Fax: 86+10-62326921