Worsley K J, Cao J, Paus T, Petrides M, Evans A C
Department of Mathematics and Statistics, McGill University, Montreal, Québec, Canada.
Hum Brain Mapp. 1998;6(5-6):364-7. doi: 10.1002/(SICI)1097-0193(1998)6:5/6<364::AID-HBM6>3.0.CO;2-T.
Functional connectivity between two voxels or regions of voxels can be measured by the correlation between voxel measurements from either PET CBF or BOLD fMRI images in 3D. We propose to look at the entire 6D matrix of correlations between all voxels and search for 6D local maxima. The main result is a new theoretical formula based on random field theory for the p-value of these local maxima, which distinguishes true correlations from background noise. This can be applied to crosscorrelations between two different sets of images--such as activations under two different tasks, as well as autocorrelations within the same set of images.
两个体素或体素区域之间的功能连接性可以通过三维PET脑血流量（CBF）或BOLD功能磁共振成像（fMRI）图像中体素测量值之间的相关性来测量。我们建议查看所有体素之间相关性的完整6D矩阵，并寻找6D局部最大值。主要结果是基于随机场理论得出的关于这些局部最大值p值的新理论公式，该公式可区分真实相关性与背景噪声。这可应用于两组不同图像之间的互相关性——例如两种不同任务下的激活情况，以及同一组图像内的自相关性。
