Dear TI,
A few remarks:
* You are probably using an earlier version of SPM that only provides
cluster level FWE-corrected p-values for t-tests (and not F-tests).
* For cluster level inference using RFT to be appropriate, the
cluster-forming threshold has to be high, e.g. p<0.001 uncorrected, see:
http://www.scholarpedia.org/article/Statistical_parametric_mapping_%28SPM%29
* For VBM analyses, the assumption of stationary smoothness is unlikely
to be appropriate. This is important for cluster level inference, see:
http://www.sciencedirect.com/science/article/pii/S1053811908000438#sec5
Best regards,
Guillaume.
On 22/12/16 03:04, TI wrote:
> Hi, expert
>
> I want to perform one-way ANOVA with VBM to find differences among three groups and
> carry out cluster-wise FWE correction sertting p at 0.05.
> However I cannot find FWE c data in the result table which is always shown under the table in performing two-sample t-test.
>
> How should I set cluster numbers when performing one-way ANOVA with cluster-wise FWE correction?
> Any help is greatly appreciated.
> Best
> TI
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|