Dear Jonas,
Answers inline with questions below:
> On 22 Jul 2021, at 13:49, Jonas Hoffmann <[log in to unmask]> wrote:
>
> Dear FSL experts and developers,
>
> we analyze associations of continuous questionnaire scores and FA, MD, RD using tbss randomise.
> We tested multiple tracts, multiple DTI parameters, and multiple questionnaire dimensions. Our questions are:
>
> 1) We assume that randomise outputs "tcfe_corrp_tstat" are controlled for multiple comparisons inside of a single tract (mask), but not for all tracts (whole brain), is that correct?
The randomise outputs are controlled ( on a per-contrast level ) across the entire mask passed to randomise with the -m option. The default option is a “whole-brain” mask ( mean_FA_skeleton_mask ) however if you have run randomise multiple-times with different masks, then the results will be controlled on a per-mask (tract) basis.
> 2) Is it recommended to additionally control the resulting p-values for multiple comparisons (and with which method)? Our literature review revealed mostly no (reported) correction of results, some using Bonferroni to correct for a number of tracts.
If you genuinely have a separate hypothesis for each tract tested ( assuming multiple-invocations of randomise as in the above answer ) then no further correction is needed, and you can report each result as being corrected at whatever significance level was chosen ( e.g. 0.05 ). If you have a more general hypothesis across the set of tracts, i.e. that some or all of the tracts will fit your hypothesis, then Bonferroni correction would be needed to avoid inflating the false-positive rate.
Hope this helps,
Kind Regards
Matthew
>
> Thank you very much for your help!
>
> Best, Jonas
>
> ########################################################################
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1
>
> This message was issued to members of www.jiscmail.ac.uk/FSL, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=FSL&A=1
This message was issued to members of www.jiscmail.ac.uk/FSL, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
|