In the qdec analysis it is hard because we don't really support qdec anymore. You can run the command-line stream, and when you run mri_glmfit-sim to do the clustering, set --cwp 1 to keep all the clusters. At that point you can use mri_segstats with --i sig.mgh and --seg ocn.mgh (where ocn is short for a longer name with ocn in it; see mri_glmfit-sim --help for more details). BTW, it is unusual to compute the mean p-value over a region ( or -log10(p) which is what sig is)

On 10/22/19 6:19 AM, Lucia Batzu wrote:

External Email - Use Caution

Dear Freesurfer Developers,

I'm running a Qdec analysis in which I'm looking for correlations between cortical thickness and a continuous CSF measurement. I'm interested in a specific area in which a cluster of correlation pops up, but doesn't survive cluster-wise correction for multiple comparisons. The details of the cluster are the following:

Contrast: 'rh-Avg-thickness-CSF_protein-Cor', 15fwhm, DOF: 61 ClusterNo: 1 Max: 3.1156 VtxMax: 122763 Size(mm2): 1247.98 TalX: 19.9 TalY: -46.6 TalZ: -9.0 NVtxs: 2022 Annotation: lingual

I would like to know if there is a way to extract the mean p-value [or, better, -log10(p)] of the uncorrected cluster.

Thank you very much in advance,

Best, Lucia

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer