Dear all,
Apologies for the simple question, I've just come across a few approaches on jiscmail / the fslwiki which seem a little conflicting in terms of specific contrasts and details like what gets de-meaned and what doesn't (or are perhaps all the same model but executed differently).
Currently I have a standard TBSS experiment and want to look for negative and positive correlations between FA and a clinical variable. I have manually de-meaned the clinical variable and set up the following matrix / contrast files. I have then run this *without* the -D flag in randomise.
Is this a valid setup to look for positive and negative correlations? If so, I take it contrast #1 would identify areas of negative associations (and vice versa)? If this is not the design I need, I would be interested to know what exactly it does test, as contrast #1 appears to generate some significant clusters...
Thank you all very much,
Iain
Matrix
/NumWaves 2
/NumPoints 38
/PPheights 1.000000e+00 2.000000e+01
/Matrix
1.000000e+00 -2.668421e+00
1.000000e+00 -3.684211e-01
1.000000e+00 -1.668421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 -1.268421e+00
1.000000e+00 -1.684211e-01
1.000000e+00 -1.684211e-01
1.000000e+00 -1.468421e+00
1.000000e+00 -1.568421e+00
1.000000e+00 -1.368421e+00
1.000000e+00 1.315789e-01
1.000000e+00 -4.684211e-01
1.000000e+00 -7.684211e-01
1.000000e+00 -1.568421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 -3.684211e-01
1.000000e+00 -1.668421e+00
1.000000e+00 1.631579e+00
1.000000e+00 1.733158e+01
1.000000e+00 1.631579e+00
1.000000e+00 4.315789e-01
1.000000e+00 -1.668421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 -1.468421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 8.315789e-01
1.000000e+00 -4.684211e-01
1.000000e+00 9.315789e-01
1.000000e+00 -1.268421e+00
1.000000e+00 1.431579e+00
1.000000e+00 -1.268421e+00
1.000000e+00 -2.668421e+00
1.000000e+00 1.131579e+00
1.000000e+00 2.931579e+00
1.000000e+00 1.033158e+01
1.000000e+00 1.631579e+00
Contrast
/ContrastName1 IgA Corr 1
/ContrastName2 IgA Corr 2
/NumWaves 2
/NumContrasts 2
/PPheights 2.000000e+01 2.000000e+01
/RequiredEffect 3.844 3.844
/Matrix
0.000000e+00 -1.000000e+00
0.000000e+00 1.000000e+00
########################################################################
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/
|