基于有效检测准则的fMRI源信号数目的估计方法

发布时间：2018-06-23 04:00

本文选题：功能磁共振成像 + 独立成分分析　；参考：《安徽大学学报(自然科学版)》2016年01期

【摘要】：独立成分分析方法已经成功地运用于分析功能磁共振成像数据.尽管独立成分分析方法是一个很有前景的数据驱动分析方法,但是在执行独立成分分析之前,需要确定好独立成分的数目.准确地估计出功能磁共振成像数据中独立成分数目对减少过估计或者低估计能起重要作用,目前有许多信息理论准则方法已经广泛运用于独立成分数目估计中,但通常容易出现过估计现象.该文提出一种基于有效检测准则的fMRI源信号数目的估计方法.模拟数据实验结果表明,该方法对不同平滑数据具有很好的鲁棒性.将该方法用于真实的功能磁共振成像数据中,可以减少色噪声下真实数据的过估计,对功能磁共振成像源信号数目的估计表现出了较好的综合性能.
[Abstract]:Independent component analysis (ICA) has been successfully used to analyze functional magnetic resonance imaging (fMRI) data. Although the independent component analysis method is a promising data-driven analysis method, the number of independent components should be determined before the independent component analysis is performed. Accurate estimation of the number of independent components in functional magnetic resonance imaging data can play an important role in reducing overestimation or low estimation. At present, many information theory criteria and methods have been widely used in independent component number estimation. But it is usually easy to estimate. In this paper, a method for estimating the number of fMRI source signals based on effective detection criteria is proposed. The simulation results show that the proposed method is robust to different smooth data. Applying this method to the real fMRI data can reduce the overestimation of the real data under the color noise, and the estimation of the signal number of the fMRI source shows good comprehensive performance.
【作者单位】：上海海事大学信息工程学院;
【基金】：国家自然科学基金资助项目(31170952,31470954) 上海科委基金资助项目(14590501700)
【分类号】：TN911.23;R445.2

【相似文献】