一种脑磁定位中背景噪声的去除方法

A method of reducing background noise in the localization of neuromagnetic sources

脑磁源的定位问题是脑磁图(MEG)研究的一个基本问题。在脑磁源的定位研究中,噪声对定位的准确性有很大的影响。自发的脑神经活动所产生磁场与感兴趣的诱发磁场不相关,可将其称为背景噪声。传统的消噪方法无法有效去除这种背景噪声。利用不含信号信息的纯噪协方差矩阵结合MUSIC算法,可以有效地抑制背景噪声而快速获得较好的定位结果。

Locating the sources of brain magnetic fields is a basic problem of magnetoencephalography(MEG). In the study of localization of the neuromagnetic dipole-sources- the noise has a great influence to the accuracy. The field of spontaneous brain activity is not related to the field of evoked brain activity. Therefore- such spontaneous brain activity can be referred to as background noise. The conventional MUSIC algorithm can not effectively reduce the background noise. The method- conbining pure noise covariance matrix with MUSIC algorithm- can restrain the background noise and get a good result very quickly.