Hi,

This is something we hope to have ready in the next release, but it is not ready yet.

Hence FDR is the only available option within FSL at the moment.

All the best,

Mark

On 20 Jun 2012, at 00:21, Christine Zakrzewski wrote:

The FIRST page says--In the future we plan to implement cluster-based corrections via randomise. Has this been done yet?
CZ
 

---

Date: Wed, 20 Jun 2012 00:44:02 +0200
From: [log in to unmask]
Subject: Re: [FSL] FIRST, hippocampus, uncorrected results
To: [log in to unmask]

Thank you for your answer but I think that fdr is the only possibility to perform correction for multiple comparison with FIRST (shape analysis of subcortical structures).

Do FSL experts have some advice on the question?

Ali Am@d

Le 19 juin 2012 à 20:43, Christine Zakrzewski <[log in to unmask]> a écrit :

A few options in FSL for multiple-comparisons correction--FDR, TFCE, cluster(in post-stats of feat gui).
You have to perform a multiple-comparisons correction, though.
CZ
 

---

Date: Tue, 19 Jun 2012 13:52:26 +0100
From: [log in to unmask]
Subject: [FSL] FIRST, hippocampus, uncorrected results
To: [log in to unmask]

Dear FSL users and experts,

I compare for a study 2 very highly controlled samples of patients, and I realized shape analysis on hippocampus between my two groups with the new great tool FIRST. 

Our results are very interesting according to our hypothesis.
 Un fortunately, these results don't survive with the FDR correction for multiple comparisons, that is the one available for vertex analysis.

I read on fsl-mailing list, that M. Jenkinson explained that FDR thresholding is "not the most sensitive technique for surface inference".
Thus, I ask myself if correction for multiple comparison is essential for the vertex analysis with FIRST or if it is possible to present uncorrected results for a publication ?

I thank you very much in advance for attention and your answer.

Best regards,

Ali Amad