Dear FSL experts, dear Dr. Winkler.

After running group MELODIC and obtaining group-components we ran a dual regression analysis to obtain single-subject spatial maps. We were interested in running a group comparison on 4 of those spatial maps for 4 contrasts in PALM while correcting for modality and contrasts.

Our generated config file looked like this:

# Configuration file for PALM.
# Version alpha110, running in MATLAB 8.3.0.532 (R2014a).
# 19-Mar-2018 10:48:46

-i dr_stage2_ic0001.nii
-i dr_stage2_ic0006.nii
-i dr_stage2_ic0013.nii
-i dr_stage2_ic0014.nii
-m mask.nii
-d designDR.mat
-t designDR.con
-n 5000
-T
-corrmod
-corrcon
-save1-p
-demean
-o dr_stage3

designDR.mat and designDR.con are attached. We had three groups (n1=23, n2=22, n3=24) and age as a covariate and we were interested in contrasts which among others included a combination of two groups into one.

2 questions:

1. Is our PALM command correct? During the permutation several matlab warnings were displayed talking about a matrix being singular. Do we need to specify some additional options?

2. If we want to check the results of the analysis we would need to look into dr_stage3_tfce_tstat_mcfwep_m*_c*.nii files, right? And these we need to additionally threshold at 0.95?

Thank you very much for your advice!

Kind regards,

Filipa and Paul