Hi Rysia,
Your commands looks fine to me, and 5000 iterations is good enough to see significant cluster.
You might want to try lowering your threshold to 1.3 (0.05 p value). thanks pratap
Hello,
I have just run a number of permutation and mc-z on data that was previously analysed using qdec. Unfortunately none of my clusters reach significance (in some cases there are no clusters found at all) and I am unsure as to what I am doing wrong. I am running freesurfer v4.1.0 on CentOS. My commands were as follows:
mri_glmfit y y.mgh fsgd ./qdec.fsgd dods --surf fsaverage lh
--fwhm 15 glmdir ./glmdir --C ./contrasts/lh-Diff-patient-CONTROL-Intercept-thickness.mat --sim perm 5000 2 ./glmdir/csd1
followed by:
mri_surfcluster --in ./lh-Diff-patient-CONTROL-Intercept-thickness/sig.mgh csd ./glmdir/csd1-lh-Diff-patient-CONTROL-Intercept-thickness.csd --sum lh-Diff-patient-CONTROL-Intercept-thickness/sig.cluster.sum cwsig lh-Diff-patient-CONTROL-Intercept-thickness/sig.cluster.mgh
I append a log file of mri_glmfit and the output of mri_surfcluster below there are clusters listed but none of them reach significance. It is confusing because I was under the impression that permutation thresholding was less conservative than fdr correction.
I have also run these using sim mc-z 5000 2 - maybe I am not performing enough iterations?
Also: is RFT correction implemented in the latest version of freesurfer?
Any help appreciated,
Rysia
mri_glmfit logfile:
$Id: mri_glmfit.c,v 1.138.2.3 2008/08/27 13:33:05 greve Exp $
cwd /cns_zfs/sphasas
cmdline mri_glmfit --y /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/y.mgh --fsgd /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/qdec.fsgd dods --glmdir /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15 --surf fsaverage rh --label /cns_zfs/sphasas/fsaverage/label/rh.cortex.label --C /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/contrasts/rh-Avg-Intercept-thickness.mat --C /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/contrasts/rh-Diff-OCD-CONTROL-Intercept-thickness.mat
sysname Linux
hostname haemoglobin.cns.image.iop
machine x86_64
user kkvi1114
FixVertexAreaFlag = 1
UseMaskWithSmoothing 1
OneSampleGroupMean 0
y /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/y.mgh
logyflag 0
usedti 0
FSGD /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15/qdec.fsgd
labelmask /cns_zfs/sphasas/fsaverage/label/rh.cortex.label
maskinv 0
glmdir /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15
DoFFx 0
ResidualFWHM 18.998058
SearchSpace 75188.635931
anattype surface
Output of mri_surfcluster:
cwd /cns_zfs/sphasas/qdec/OCD_CONTROL.lh.cortthick.fwhm15
# sysname Linux
# hostname haemoglobin.cns.image.iop
# machine x86_64
# FixVertexAreaFlag = 1
#
# Input ./lh-Diff-OCD-CONTROL-Intercept-thickness/sig.mgh
# Frame Number 0
# srcsubj fsaverage
# hemi lh
# surface white
# SUBJECTS_DIR /software/system/freesurfer/freesurfer-4.1.0/subjects
# Minimum Threshold 2
# Maximum Threshold infinity
# Threshold Sign abs
# AdjustThreshWhenOneTail 1
# Area Threshold 0 mm^2
# CSD thresh 2.000000
# CSD nreps 5000
# CSD simtype perm
# CSD contrast lh-Diff-OCD-CONTROL-Intercept-thickness
# CSD confint 90.000000
# Overall max 2.18803 at vertex 126138
# Overall min -2.94444 at vertex 75849
# NClusters 6
# Total Cortical Surface Area 65416.6 (mm^2)þ
# FixMNI = 1
ClusterNo Max VtxMax Size(mm2) TalX TalY TalZ CWP CWPLow CWPHi NVtxs
1 -2.944 75846 61.98 -17.6 -37.0 -4.2 0.84860 0.84200 0.85500 142
2 -2.672 22094 96.92 -29.6 0.5 -24.2 0.80680 0.79960 0.81400 182
3 2.188 126138 41.24 -22.1 -56.8 40.5 0.87160 0.86560 0.87760 74
4 2.095 81711 45.09 -32.4 -38.2 39.4 0.86860 0.86240 0.87480 105
5 2.061 147756 12.32 -6.2 -61.7 38.4 0.90520 0.89980 0.91040 20
6 2.001 133188 0.60 -32.0 -44.7 59.0 0.91660 0.91160 0.92160 1
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer