用 voxel level FWE 的时候描述 cluster 大小

各位专家,您好

我在处理一些 Repeat Measure ANOVA 数据的时候,用了SPM

在结果数据的时候,我用的 voxel level FWE ,P<0.01 (FWE校正, 因为数据量大效应也挺明显,我把阈值做得比较低严)

然后,为了避免噪音和小的 cluster干扰,我设定cluster size 30, 然后审稿人误认为我是用了 cluster level 的FWE

我不知道怎样Revise 才好,不知道这样是否可以,多谢大家。

 In the current study, the results of whole brain analysis were corrected for multiple comparisons at p < .01, family wise error (FWE) corrected at the voxel level. In order to rule out unreliable small significant clusters, only the cluster lager than 30 significant voxels (p < .01, FWE corrected at the voxel level) were regarded as reliable.

 

Forums:

你用的voxel level的Bonferroni校正吗?

严老师,不好意思我最近做了个手术,刚刚又登录这个论坛

我是用 SPM 软件默认的 voxel-level 的 FWE校正,不知道怎样描述,审稿人今天问我说FWE的详细情况,不知怎样答复为好,多谢指教

Since massive multiple comparisons were corrected with some familywise error rate, this strengthens our confidence in the whole brain connectivity results and allows us to interpret the cluster as precise anatomic findings. The authors will need to state which familywise error rate correction method was uses.  (FWE is a category of methods, not a method itself).

如果是voxel-level,可能是Bonferroni校正。

如果是cluster-level,可能是GRF校正。

你最好在SPM maillist上问一下。

https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=SPM&X=E35F2D01A10705C919&Y=18940258370%40163.com

SPM  Mai list 的答复还是认为 是随机场理论,不知道怎样解释给审稿人为好,多谢!

Dear Bing,

 
The SPM FWE correction is based on random field theory. Below are a couple of links to helpful websites, and you can find lots more by searching for 'random field theory' and 'spm'.
 
Hope this helps!
 
Jim
 
 
 

l

 

This is GRF correction.

Read from here: Chen, X., Lu, B., Yan, C.G., 2018. Reproducibility of R-fMRI metrics on the impact of different strategies for multiple comparison correction and sample sizes. Hum Brain Mapp 39, 300-318.