sorry about that. Please see below:
Basically, I am attempting to account for family structure in my analysis of HCP data and wanted to make sure that I am using PALM properly.
1) To generate the Exchangeability Blocks files, I ran hcp2blocks.m as directed on the instructions page (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/PALM/ExchangeabilityBlocks#EBs_for_data_of_the_Human_Connectome_Project). I didn’t get an error running the script but not sure whether the numbers in row #2 make sense so wanted to make sure that what I have is correct. If the output is incorrect, any pointers as to why this would occur? Here is a snapshot of what was generated (last column of subject IDs isn’t pasted here):
-1 212 -61 100
-1 2012 -99 1000
-1 2013 -3 1000
-1 22 128 10
-1 32 169 10
-1 32 148 10
-1 32 129 10
-1 42 132 10
-1 2002 91 1000
-1 2012 -43 1000
-1 212 -27 10
-1 32 148 10
-1 23 71 10
-1 2002 23 1000
2) I ran PALM using the following (input file, disgn.mat, design.con and mask file as generated with FEAT):
palm -i filtered_func_data.nii -o PALMoutput -d design.mat -t design.con -m mask.nii -n 5000 -C 2.5758 -ise -corrcon -logp -eb EB.csv
Actually, the results were identical regardless of whether I used EB.csv file or not. What was a bit surprising to me is that PALM results were somewhat different from the output I obtained with randomise. I was under the impression that the two programs (randomise and PALM) should produce the same results. Please correct me if I am wrong.
Would appreciate any pointers.
Regards,
Ekaterina
|