Hi,

Since I'm very new to FSL and PALM I would appreciate if anyone can answer my questions regarding PALM command and model set-up. 

Thanks!

Shie Rinat

---

From: [log in to unmask]
Sent: February-19-19 1:13 PM
To: [log in to unmask]
Subject: PALM question

Dear FSL experts,

I'm new to FSL and even more so to PALM, and I will appreciate a conformation that the contrasts and PALM command I'm using are accurate.

I'm analyzing data from a resting state study, and I wish to compare ICA networks between three groups, while controlling for age and sex (the main effect of age and sex is not of interest). I plan to use PALM for the group comparison after dual regression, with the tfce and fdr options (as a relatively lenient method to correct for multiple comparisons). 

PALM command:

palm –i dr_stage2_ic0001.nii.gz –i dr_stage2_ic0002.nii.gz –i dr_stage2_ic0005.nii.gz  –i dr_stage2_ic0007.nii.gz –i dr_stage2_ic0008.nii.gz –i dr_stage2_ic0010.nii.gz –i dr_stage2_ic0011.nii.gz –i dr_stage2_ic0014.nii.gz –i dr_stage2_ic0018.nii.gz -d model.mat -t model.con -n 5000 -T -corrmod -corrcon -fdr -save1-p

This is design matrix and contrasts (I have ~20 participants in each group, I used only 9 here for the example):

EV1- Group 1 (1 1 1 0 0 0 0 0 0)

EV2- Group 2 (0 0 0 1 1 1 0 0 0)

EV3- Group 3 (0 0 0 0 0 0 1 1 1)

EV4- Age (demeaned)

EV5- Sex

C1-  1  0 -1 0 0 

C2- -1  0  1 0 0 

C3-  1 -1  0 0 0

C4- -1  1  0 0 0

C5-  0 -1  1 0 0

C6-  0  1 -1 0 0    

(I'm not using an f-test since I read in previous correspondence in this forum that it will be easier to interpret the results if using only t-tests. Is that correct?)

I was also hoping you could explain why is it  recommended to use the logp (and not 1-p for example)?

Thanks!

Shie Rinat

University of British Columbia 

---

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1