Dear Anna,
After reviewing your design, it seems fine to me. Please see below for some additional comments.
All the best,
Bryan
> On 17 Mar 2020, at 00:21, Anna Manelis <[log in to unmask]> wrote:
>
> Dear SWE exerts,
>
> We would like to use the SWE tool to conduct a 3 (groups) by 3 (task conditions - within subjects) analysis with several covariates of no interest (age, IQ, gender, scanner - all demeaned) across around 300 subjects. All subjects have all conditions.
>
> We would really appreciate your feedback about the design described below.
>
> The design.sub file has the following structure:
> 1
> 2
> 3
> ...
> 1
> 2
> 3
> ...
> 1
> 2
> 3
> ...
>
> The cope.nii.gz files for each contrast of interest were merged using the same principle (all subjects condition 1, all subjects condition 2, all subjects condition 3).
>
> The design.mat file has the following structure:
> Gr1_Cond1 Gr2_Cond1 Gr3_Cond1 Gr1_Cond2 Gr2_Cond2 Gr3_Cond2 Gr1_Cond3 Gr2_Cond3 Gr3_Cond3 Age Sex IQ Scanner
>
> 1 0 0 0 0 0 0 0 0 4.7 0.34 3.4 0.7
> 1 0 0 0 0 0 0 0 0 4.2 -0.66 12.4 0.7
> ...
> 0 1 0 0 0 0 0 0 0 1.9 0.34 -20.6 -0.3
> 0 1 0 0 0 0 0 0 0 -2.1 -0.66 7.4 0.7
> ...
> 0 0 1 0 0 0 0 0 0 1.4 -0.66 3.4 -0.3
> 0 0 1 0 0 0 0 0 0 2.8 -0.66 7.4 0.7
> .........................................
> 0 0 0 1 0 0 0 0 0 4.7 0.34 3.4 0.7
> 0 0 0 1 0 0 0 0 0 4.2 -0.66 12.4 0.7
> ...
> 0 0 0 0 1 0 0 0 0 1.9 0.34 -20.6 -0.3
> 0 0 0 0 1 0 0 0 0 -2.1 -0.66 7.4 0.7
> ...
> 0 0 0 0 0 1 0 0 0 1.4 -0.66 3.4 -0.3
> 0 0 0 0 0 1 0 0 0 2.8 -0.66 7.4 0.7
> ........................................
> 0 0 0 0 0 0 1 0 0 4.7 0.34 3.4 0.7
> 0 0 0 0 0 0 1 0 0 4.2 -0.66 12.4 0.7
> ...
> 0 0 0 0 0 0 0 1 0 1.9 0.34 -20.6 -0.3
> 0 0 0 0 0 0 0 1 0 -2.1 -0.66 7.4 0.7
> ...
> 0 0 0 0 0 0 0 0 1 1.4 -0.66 3.4 -0.3
> 0 0 0 0 0 0 0 0 1 2.8 -0.66 7.4 0.7
> ...
>
>
> The contrasts and F-tests are as below. The F tests are computed for the main effect of Group, the main effect of Condition and the Group x Condition interaction. The real design also has four "0" after 'Gr3_Cond3' for covariates of no interest. I skip them here to save space. The last three values in each row are for the F-tests. They are presented here together with the contrasts to better illustrate the design, but for the real analysis we saved them in the design.fts
>
> Gr1_Cond1 Gr2_Cond1 Gr3_Cond1 Gr1_Cond2 Gr2_Cond2 Gr3_Cond2 Gr1_Cond3 Gr2_Cond3 Gr3_Cond3 F1_Group F2_Condition F3_GroupxCondition
>
> Gr3_vs_Gr1 -1 0 1 -1 0 1 -1 0 1 1 0 0
> Gr3_vs_Gr2 0 -1 1 0 -1 1 0 -1 1 1 0 0
> Cond1_vs_Cond3 1 1 1 0 0 0 -1 -1 -1 0 1 0
> Cond2_vs_Cond3 0 0 0 1 1 1 -1 -1 -1 0 1 0
> Gr1_Cond1-Gr1_Cond3_vs_Gr3_Cond3-Gr3_Cond1 1 0 -1 0 0 0 -1 0 1 0 0 1
> Gr2_Cond1-Gr2_Cond3_vs_Gr3_Cond3-Gr3_Cond1 0 1 -1 0 0 0 0 -1 1 0 0 1
> Gr1_Cond2-Gr1_Cond3_vs_Gr3_Cond3-Gr3_Cond2 0 0 0 1 0 -1 -1 0 1 0 0 1
> Gr2_Cond2-Gr2_Cond3_vs_Gr3_Cond3-Gr3_Cond2 0 0 0 0 1 -1 0 -1 1 0 0 1
>
>
> The swe command:
> swe -i ${input} -o ${out} -d design.mat -t contrast.con -s subj.sub -f ftest.fts -m ${mask} -x -T --glm_output -R -E -D --corrp --wb -n 999
I do not think this is an issue, but both options -x and --corrp are doing the same thing. Thus using only one would be sufficient.
>
> The *_corrp_* output images report the corrected p-values=1-*_corrp*
Yes, these images contain the FWER-corrected p-values saved as 1- p-values. For your information, if you had used the option --logp, the p-value images would have been saved as -log_10(p) images instead of 1-p images and would have used the labels _lcorrp_ instead of _corrp_.
>
> Any feedback will be great appreciated.
>
> Thank you very much,
> Anna.
>
> ########################################################################
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|