Hi taylor !
Thank you for your prompt response. Yes, I am using a gray matter mask. I have upload the GM mask that we usually use in our lab. This mask was obtained from FSL MNI 152 avg and several changes were done on it.
Link to GM mask: [
indianinstituteofscience-my.sharepoint.com]
Can you please let me know if there is any issue with the GM mask attached ?
I also recently started using the GM mask provided in the SSwarper. So, I resampled it using :
3dresample -dxyz 3.0 3.0 3.0 -prefix MNI152_2009_template_SSW_3mm_brain_GM.nii.gz -input MNI152_2009_template_SSW.nii.gz'[4]'
This resampled GM mask shows that a grid of 64x76x64 3.00x3.00x3.00 mm^3 (53104 voxels in mask)
I also used -iter 30000 and ran thrice using the same ACF parameters ( 0.5128 3.8451 11.3839):
3dClustSim \
-mask $temp_path/MNI152_2009_template_SSW_3mm_brain_GM.nii.gz \
-acf 0.5128 3.8451 11.3839 \
-nodec \
-iter 30000 \
-prefix 3dClustSimOutput_censor_MR_ACF.txt
and the pthr=0.001 with alpha=0.05 I get cluster sizes of : 15, 17, 16 . (The full output tables are below, just for reference).
# 3dClustSim -mask /media/cogemolab/home2/templates/MNI152_2009_template_SSW_3mm_brain_GM.nii.gz -acf 0.5128 3.8451 11.3839 -nodec -iter 30000 -prefix 3dClustSimOutput_censor_MR_ACF.txt
# bi-sided thresholding
# Grid: 64x76x64 3.00x3.00x3.00 mm^3 (53104 voxels in mask)
#
# CLUSTER SIZE THRESHOLD(pthr,alpha) in Voxels
# -NN 2 | alpha = Prob(Cluster >= given size)
# pthr | .10000 .05000 .02000 .01000
# ------ | ------ ------ ------ ------
0.050000 221 267 329 375
0.020000 87 104 132 152
0.010000 49 59 75 87
0.005000 30 37 47 54
0.002000 17 21 27 32
0.001000 12 15 19 22
0.000500 9 11 14 16
0.000200 6 7 9 11
0.000100 4 5 7 8
# 3dClustSim -mask /media/cogemolab/home2/templates/MNI152_2009_template_SSW_3mm_brain_GM.nii.gz -acf 0.5128 3.8451 11.3839 -nodec -iter 30000 -prefix 3dClustSimOutput_censor_MR_ACF.txt
# bi-sided thresholding
# Grid: 64x76x64 3.00x3.00x3.00 mm^3 (53104 voxels in mask)
#
# CLUSTER SIZE THRESHOLD(pthr,alpha) in Voxels
# -NN 2 | alpha = Prob(Cluster >= given size)
# pthr | .10000 .05000 .02000 .01000
# ------ | ------ ------ ------ ------
0.050000 238 285 353 408
0.020000 96 115 144 167
0.010000 55 67 83 97
0.005000 34 41 52 60
0.002000 20 24 30 36
0.001000 14 17 21 25
0.000500 10 12 15 18
0.000200 7 8 10 12
0.000100 5 6 8 9
# 3dClustSim -mask /media/cogemolab/home2/templates/MNI152_2009_template_SSW_3mm_brain_GM.nii.gz -acf 0.5128 3.8451 11.3839 -nodec -iter 30000 -prefix 3dClustSimOutput_censor_MR_ACF.txt
# bi-sided thresholding
# Grid: 64x76x64 3.00x3.00x3.00 mm^3 (53104 voxels in mask)
#
# CLUSTER SIZE THRESHOLD(pthr,alpha) in Voxels
# -NN 2 | alpha = Prob(Cluster >= given size)
# pthr | .10000 .05000 .02000 .01000
# ------ | ------ ------ ------ ------
0.050000 228 275 342 395
0.020000 91 110 139 161
0.010000 51 63 79 92
0.005000 32 39 49 57
0.002000 18 23 28 33
0.001000 13 16 20 23
0.000500 9 11 14 17
0.000200 6 8 10 11
0.000100 5 6 7 9
And I also tried running 30000 iteration using the old GM mask (attached file;Grid: 91x109x91 2.00x2.00x2.00 mm^3 (160742 voxels in mask) ), for pthr=0.001 with alpha=0.05 I get cluster sizes of : 49, 46 (The full output tables are below, just for reference)
So, is it an issue with the old GM mask used ?
Also another question, I understand that there will be some stochasticity while calculating the cluster size, but how much variation between runs would be considered a good estimate ?
# 3dClustSim -mask /media/emocoglab/home2/templates/MNI152_T1_2mm_brain_GM_02182017.nii.gz -acf 0.5128 3.8451 11.3839 -nodec -iter 30000 -prefix 3dClustSimOutput_censor_MR_6mm_ACF.txt
# bi-sided thresholding
# Grid: 91x109x91 2.00x2.00x2.00 mm^3 (160742 voxels in mask)
#
# CLUSTER SIZE THRESHOLD(pthr,alpha) in Voxels
# -NN 2 | alpha = Prob(Cluster >= given size)
# pthr | .10000 .05000 .02000 .01000
# ------ | ------ ------ ------ ------
0.050000 748 897 1099 1279
0.020000 297 362 448 521
0.010000 168 206 261 306
0.005000 103 126 161 191
0.002000 58 72 93 111
0.001000 39 49 65 78
0.000500 27 35 45 54
0.000200 17 22 29 35
0.000100 12 16 21 26
# 3dClustSim -mask /media/emocoglab/home2/templates/MNI152_T1_2mm_brain_GM_02182017.nii.gz -acf 0.5128 3.8451 11.3839 -nodec -iter 30000 -prefix 3dClustSimOutput_censor_MR_6mm_ACF_run2.txt
# bi-sided thresholding
# Grid: 91x109x91 2.00x2.00x2.00 mm^3 (160742 voxels in mask)
#
# CLUSTER SIZE THRESHOLD(pthr,alpha) in Voxels
# -NN 2 | alpha = Prob(Cluster >= given size)
# pthr | .10000 .05000 .02000 .01000
# ------ | ------ ------ ------ ------
0.050000 716 859 1072 1231
0.020000 284 345 428 490
0.010000 160 193 247 289
0.005000 97 119 154 180
0.002000 54 67 87 102
0.001000 36 46 59 71
0.000500 25 32 42 50
0.000200 16 20 27 33
0.000100 11 14 19 24