Hello FSL PALM users, I have some questions about the cluster extent output when using PALM with CIFTI data. I am running a 3 group ANCOVA model with fMRI contrast images (COPE) generated using the HCP pipelines. My design matrix contains 136 rows (subjects) x 6 columns (3 covariates— age, sex (1=female, 0=male), handedness (1=right hand, 0=left hand), and 3 categorical groupings— low, medium, and high nicotine dependence). I am requesting surface level cluster extent results based off an initial vertex-wise threshold of p<.05 (C=1.96).
#######################################
Here is my palm script for the R.Surface —
#######################################
Running PALM alpha109 using MATLAB 9.0.0.341360 (R2016a) with the following options:
-i ../../../separateCIFTI/Pre_NicNeut_Separated/R_Nic_-_Neut_N136.func.gii
-d ../design_Pre_NicNeut_HSI_wCovars_noProj_N136.mat
-t ../Pre_NicNeut_HSI_corr_noProj_wCovars.con
-f ../Dep_noProj.fts
-demean
-fonly
-n 5000
-Cstat extent
-C 1.96
-s N136_R.midthickness.32k_fs_LR.surf.gii ./N136_R_area.func.gii
-logp
-saveglm
-saveparametric
-o PreNicNeut_N136_R_5k
#######################################
Here are the first few lines of my matrix—
#######################################
%Age,Sex,Hand,HSI-Low,HSI-Med,HSI-High
/NumWaves 6
/NumPoints 136
/Matrix
28 1 1 1 0 0
45 1 1 0 1 0
25 1 0 1 0 0
26 1 1 1 0 0
33 1 1 0 0 1
#######################################
Here is my con file—
#######################################
%Columns are Age,Sex,Hand,HSI-Low,HSI-Med,HSI-High
/ContrastName1 High>Low
/ContrastName2 Med>Low
/ContrastName3 High>Med
/NumWaves 6
/NumContrasts 3
/Matrix
0 0 0 -1 0 1
0 0 0 -1 1 0
0 0 0 0 -1 1
#######################################
Here is my F-test file—
#######################################
/NumWaves 3
/NumPoints 1
/Matrix
1 1 1
#######################################
Questions--
1.) Is the F-test correctly setup to test the hypothesis that the 3 groups differ in their brain activations when adjusting for age, sex, and handedness?
2.) To identify significant clusters, is it necessary to threshold the resulting “clustere_fstat_fwep” map? As I am requesting -logp, do I threshold at > 1.3 to identify clusters that pass a corrected p<.05? The min. and max. of my map is 0–0.9, so I am unsure if this means nothing passes a corrected -logp > 1.3, or, if the p-values are back to a non-log-p scale and I should instead threshold at <.05.
3.) How can I create separate masks for each cluster identified from the “clustere_fstat_fwep” map? As an example, let’s say I threshold the “clustere_fstat_fwep” map and find two clusters, one in left insula, and another one in right IFG. Is there a recommended tool I can apply to this map to obtain two separate masks? I want to use these masks to perform an ROI extraction and conduct pairwise t-tests for each region to identify which groups are driving the effects.
4.) Is there a recommended tool to obtain the mean COPE values within each separate region for each subject? I have been using wb_command -cfiti-stats with freesurfer ROIs, so I presume I would use this tool with the masks isolated from Q3 above.
Thank you!
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|