Dear Marvin,
The way randomise deals with cluster-thresholding is a little different to FLAME. Since the stats are non-parametric, no prior t->z transform is known, so the cluster-forming threshold is applied directly to the t-stats. For display purposes a basic t->z transform ( using the permutation data ) is applied to the input t-statistics but this does have some limitations - e.g. the maximum possible z-stat will correspond to a p of 1/nPermutations ( here ~3.54 for 1/5000 ).
Hope this helps,
Kind Regards
Matthew
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
On 5 Jul 2023, at 11:56, Marvin Meiering MSB Berlin <[log in to unmask]> wrote:
Dear FSL experts,
point 3 in my previous request regarding the large areas of zeros is not of interest anymore, I now understand why and how this is done. However, I am still wondering why the z stat maps are thresholded at 2.7 (lower bound) and 3.54 (upper bound) and if this is intended? The tstat images are clearly showing much more variance and higher values in areas activated by the task. Reanalysing the data using FLAME reveals z values of up to 9.8 instead of 3.54 using randomise.
Maybe my problem is a too basic, however, a little help would be much appreciated! I uploaded the gfeat directories here: https://drive.google.com/drive/folders/1Bi0ljMM-43gTadfqwtmxcWab936z6CDN?usp=sharing
Thank you for your time and help!
Best,
Marvin
Am 01.07.23, 00:38 schrieb "FSL - FMRIB's Software Library im Auftrag von Marvin Meiering" <[log in to unmask] <mailto:[log in to unmask]> im Auftrag von [log in to unmask] <mailto:[log in to unmask]>>:
Dear FSL experts,
I am not quite sure if I have an issue regarding the output of randomise. I am conducting a regular activity as well as a PPI analysis using task based fMRI. Three groups are compared using randomise. Conducting a one sample t test to test whether the task works and significantly activates the regions of interest, irrespective of group, reveals the following results: minimum z statistic of 2.770329 and maximum z statistic of 3.540119.For some reason are the connectivity maps showing similar stats: minimum z statistic of 2.726552 and a maximum z statistic of 3.540119. Thresholds of the same analysis with a different seed are: minimum z statistic of 2.706481 and a maximum z statistic of 3.540119.
1. Why is the lower threshold set on approximately 2.7, although being prespecified as 3.1 in the GUI?
2. Why is the upper threshold limited to 3.540119?
3. Furthermore, a zstat NIFTI in the stats directory of the activity analysis is showing large areas of zeros instead of floating numbers. This is not the case in the results of the PPI analyses, which also seems kind of odd to me.
All analyses are set up using the FEAT GUI.
The log is showing the following commands for the thresholding in the activity analysis (one sample t test):
/Users/marvin/fsl/bin/fslmaths stats/zstat1 -mas mask thresh_zstat1
echo 231729 > thresh_zstat1.vol
zstat1: DLH=1 VOLUME=231729 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/pstat1 -mas mask thresh_pstat1
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat1 -t 0.95 --othresh=thresh_zstat1 -o cluster_mask_zstat1 --connectivity=26 --mm --olmax=lmax_zstat1_std.txt --scalarname=Z -p 0.05 -d 1 --volume=231729 -c stats/cope1 --empiricalNull=thresh_pstat1 > cluster_zstat1_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat1 -std
/Users/marvin/fsl/bin/fslstats thresh_zstat1 -l 0.0001 -R 2>/dev/null
2.726552 3.540119
Rendering using zmin=2.726552 zmax=3.540119
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat1 2.726552 3.540119 rendered_thresh_zstat1
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat1 -S 2 750 rendered_thresh_zstat1.png
/bin/cp /Users/marvin/fsl/etc/luts/ramp.gif .ramp.gif
mkdir -p tsplot ; /Users/marvin/fsl/bin/tsplot . -f filtered_func_data -o tsplot
The log is showing the following commands for the thresholding in the connectivity analysis (PPI):
/Users/marvin/fsl/bin/fslmaths stats/zstat1 -mas mask thresh_zstat1
echo 257138 > thresh_zstat1.vol
zstat1: DLH=1 VOLUME=257138 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat2 -mas mask thresh_zstat2
echo 257202 > thresh_zstat2.vol
zstat2: DLH=1 VOLUME=257202 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat3 -mas mask thresh_zstat3
echo 257195 > thresh_zstat3.vol
zstat3: DLH=1 VOLUME=257195 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat4 -mas mask thresh_zstat4
echo 256496 > thresh_zstat4.vol
zstat4: DLH=1 VOLUME=256496 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat5 -mas mask thresh_zstat5
echo 256901 > thresh_zstat5.vol
zstat5: DLH=1 VOLUME=256901 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat6 -mas mask thresh_zstat6
echo 257173 > thresh_zstat6.vol
zstat6: DLH=1 VOLUME=257173 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/zstat7 -mas mask thresh_zstat7
echo 257146 > thresh_zstat7.vol
zstat7: DLH=1 VOLUME=257146 RESELS=1
/Users/marvin/fsl/bin/fslmaths stats/pstat1 -mas mask thresh_pstat1
/Users/marvin/fsl/bin/fslmaths stats/pstat2 -mas mask thresh_pstat2
/Users/marvin/fsl/bin/fslmaths stats/pstat3 -mas mask thresh_pstat3
/Users/marvin/fsl/bin/fslmaths stats/pstat4 -mas mask thresh_pstat4
/Users/marvin/fsl/bin/fslmaths stats/pstat5 -mas mask thresh_pstat5
/Users/marvin/fsl/bin/fslmaths stats/pstat6 -mas mask thresh_pstat6
/Users/marvin/fsl/bin/fslmaths stats/pstat7 -mas mask thresh_pstat7
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat1 -t 0.95 --othresh=thresh_zstat1 -o cluster_mask_zstat1 --connectivity=26 --mm --olmax=lmax_zstat1_std.txt --scalarname=Z -p 0.05 -d 1 --volume=257138 -c stats/cope1 -- empiricalNull=thresh_pstat1 > cluster_zstat1_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat1 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat2 -t 0.95 --othresh=thresh_zstat2 -o cluster_mask_zstat2 --connectivity=26 --mm --olmax=lmax_zstat2_std.txt --scalarname=Z -p 0.05 -d 1 --volume=257202 -c stats/cope2 -- empiricalNull=thresh_pstat2 > cluster_zstat2_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat2 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat3 -t 0.95 --othresh=thresh_zstat3 -o cluster_mask_zstat3 --connectivity=26 --mm --olmax=lmax_zstat3_std.txt --scalarname=Z -p 0.05 -d 1 --volume=257195 -c stats/cope3 --empiricalNull=thresh_pstat3 > cluster_zstat3_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat3 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat4 -t 0.95 --othresh=thresh_zstat4 -o cluster_mask_zstat4 --connectivity=26 --mm --olmax=lmax_zstat4_std.txt --scalarname=Z -p 0.05 -d 1 --volume=256496 -c stats/cope4 --empiricalNull=thresh_pstat4 > cluster_zstat4_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat4 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat5 -t 0.95 --othresh=thresh_zstat5 -o cluster_mask_zstat5 --connectivity=26 --mm --olmax=lmax_zstat5_std.txt --scalarname=Z -p 0.05 -d 1 --volume=256901 -c stats/cope5 --empiricalNull=thresh_pstat5 > cluster_zstat5_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat5 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat6 -t 0.95 --othresh=thresh_zstat6 -o cluster_mask_zstat6 --connectivity=26 --mm --olmax=lmax_zstat6_std.txt --scalarname=Z -p 0.05 -d 1 --volume=257173 -c stats/cope6 --empiricalNull=thresh_pstat6 > cluster_zstat6_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat6 -std
/Users/marvin/fsl/bin/fsl-cluster -i thresh_zstat7 -t 0.95 --othresh=thresh_zstat7 -o cluster_mask_zstat7 --connectivity=26 --mm --olmax=lmax_zstat7_std.txt --scalarname=Z -p 0.05 -d 1 --volume=257146 -c stats/cope7 --empiricalNull=thresh_pstat7 > cluster_zstat7_std.txt
/Users/marvin/fsl/bin/cluster2html . cluster_zstat7 -std
/Users/marvin/fsl/bin/fslstats thresh_zstat1 -l 0.0001 -R 2>/dev/null
0.000000 0.000000
/Users/marvin/fsl/bin/fslstats thresh_zstat2 -l 0.0001 -R 2>/dev/null
2.878166 3.540119
/Users/marvin/fsl/bin/fslstats thresh_zstat3 -l 0.0001 -R 2>/dev/null
2.706481 3.540119
/Users/marvin/fsl/bin/fslstats thresh_zstat4 -l 0.0001 -R 2>/dev/null
0.000000 0.000000
/Users/marvin/fsl/bin/fslstats thresh_zstat5 -l 0.0001 -R 2>/dev/null
0.000000 0.000000
/Users/marvin/fsl/bin/fslstats thresh_zstat6 -l 0.0001 -R 2>/dev/null
0.000000 0.000000
/Users/marvin/fsl/bin/fslstats thresh_zstat7 -l 0.0001 -R 2>/dev/null
0.000000 0.000000
Rendering using zmin=2.706481 zmax=3.540119
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat1 2.706481 3.540119 rendered_thresh_zstat1
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat1 -S 2 750 rendered_thresh_zstat1.png
/bin/cp /Users/marvin/fsl/etc/luts/ramp.gif .ramp.gif
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat2 2.706481 3.540119 rendered_thresh_zstat2
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat2 -S 2 750 rendered_thresh_zstat2.png
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat3 2.706481 3.540119 rendered_thresh_zstat3
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat3 -S 2 750 rendered_thresh_zstat3.png
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat4 2.706481 3.540119 rendered_thresh_zstat4
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat4 -S 2 750 rendered_thresh_zstat4.png
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat5 2.706481 3.540119 rendered_thresh_zstat5
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat5 -S 2 750 rendered_thresh_zstat5.png
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat6 2.706481 3.540119 rendered_thresh_zstat6
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat6 -S 2 750 rendered_thresh_zstat6.png
/Users/marvin/fsl/bin/overlay 1 0 example_func -a thresh_zstat7 2.706481 3.540119 rendered_thresh_zstat7
/Users/marvin/fsl/bin/slicer rendered_thresh_zstat7 -S 2 750 rendered_thresh_zstat7.png
mkdir -p tsplot ; /Users/marvin/fsl/bin/tsplot . -f filtered_func_data -o tsplot
Has anyone observed something similar or could provide an explanation for these results?
Help would be much appreciated, so thank you in advance!
Best,
Marvin
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1 <https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1>
This message was issued to members of www.jiscmail.ac.uk/FSL, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/ <https://www.jiscmail.ac.uk/policyandsecurity/>
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1
This message was issued to members of www.jiscmail.ac.uk/FSL, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1
This message was issued to members of www.jiscmail.ac.uk/FSL, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
