Hi all,
I have been looking through the search board but no one seems to have my same problem.
So I have run my TBSS analysis and now I'm looking to run Cluster.
When I open my tbss_tfce_corrp_tstat1 image I have MANY red and yellow regions, which is great!
cluster --in=tbss_tfce_corrp_tstat1.nii.gz --thresh=0.95
When I run the cluster analysis with the corrp file (tbss_tfce_corrp_tstat1) and t=0.95 threshold I only get 1 massive cluster.
Cluster Index Voxels MAX MAX X (vox) MAX Y (vox) MAX Z (vox) COG X (vox) COG Y (vox) COG Z (vox)
1 63458 0.994 128 119 48 88.6 112 89.2
This 1 large cluster stays relatively constant (without breaking up into smaller clusters) all the way until t=0.99.
Am I right to assume a global reduction in FA?
If this is true, then is my only way of determine where the most significant cluster are is by overlaying my tbss_tfce_corrp_tstat1 with an atlas and determining the areas manually?
With this said, when i do the same command with the uncorrect image (with p=0.95 and p=0.99), I end up with 237 small clusters. This does not make sense to me because the corrp image should be more stringent.
cluster --in=tbss_tfce_p_tstat1.nii.gz --thresh=0.99
Cluster Index Voxels MAX MAX X (vox) MAX Y (vox) MAX Z (vox) COG X (vox) COG Y (vox) COG Z (vox)
237 49951 0.998 130 120 48 87.7 110 88.9
236 568 0.998 142 123 89 137 112 104
235 150 0.998 35 116 101 37.1 114 102
234 127 0.998 133 137 84 134 132 86
233 93 0.994 64 158 87 64.2 152 96.2
232 74 0.998 46 107 115 48.2 106 116
231 74 0.998 141 141 87 135 141 86.7
230 72 0.998 76 127 63 70.2 120 62.6
229 68 0.998 45 99 97 41.7 103 100
When I open up my tbss_tfce_corrp_tstat1 image and put in the above co-ordinates (i.e. 130, 120, 48) I am in random or non-significant areas of the brain.
.. Im not sure if anyone else has had this problem, but any advice is helpful :)
Thanks.
Erica
|