Hi,

I'm having trouble with the output of 3dLME (I think)

I've been using it to investigate difference sin resting state connectivity and am running it as follows:

3dLME -prefix restingstate.patAndCtrl.group.and.gender.L_whole_amygdala.3mm.lme.bucket.20151102-1251PST
-jobs 8
-mask mask.grey.patAndCtrl.union.masked+tlrc.HEAD
-model 'Group*Gender+Full'
-ranEff '~1|Subj'
-qVars 'Full'
-SS_type 3
-num_glt 4
-gltLabel 1 'PAT-NCL' -gltCode 1 'Group : 1*PAT -1*NCL'
-gltLabel 2 'M-F' -gltCode 2 'Gender : 1*M -1*F'
-gltLabel 3 'PAT.F-PAT.M' -gltCode 3 'Group : 1*PAT Gender : 1*F -1*M'
-gltLabel 4 'NCL.F-NCL.M' -gltCode 4 'Group : 1*NCL Gender : 1*F -1*M'
-dataTable @dataTable.patAndCtrl.group.and.gender.L_whole_amygdala.3mm.txt

For the sake of brevity some of the lines in the data table are as follows (trailing \ omitted):

age and Full are already mean centered (though Full shows a
between-group difference on a t-test). The numbers at the beginning of
the lines are NOT in the file but are placed there by R's print
command

     Subj Group        age        Full                                                                     InputFile
41  141_A   NCL -0.5210804   2.7623762 141_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
65  165_A   NCL  0.9789196  -4.2376238 165_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
95  326_A   NCL  1.3388097  11.7623762 326_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
106 337_A   NCL -3.0210804  21.7623762 337_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
149 380_A   NCL -2.2408606  21.7623762 380_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
174 403_A   NCL  0.7893592   1.7623762 403_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
13  113_A   PAT -1.2545969  -0.2376238 113_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
20  120_A   PAT -0.1941573 -21.2376238 120_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
82  313_A   PAT  0.6986998  -6.2376238 313_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD
135 366_A   PAT -2.1666848   6.7623762 366_A/rsfc/L_whole_amygdala.3mm/L_whole_amygdala.3mm.z-score.masked+tlrc.HEAD

The issue I'm having is that the results in the Group F stat subbrik
(left of image) are substantially different from those displayed in
PAT-NCL contrast (right of image) as can bee seen in the attached
screenshot. If I'm understanding this correctly the Group main effect
the difference between groups collapsed across genders and with Full
held at its mean value. Then the glt 1 should give me the signed
difference between the PAT and NCL group collapsed across genders with
Full held at its mean value. What I don't understand is why there is
such a large discrepency between the two briks when displayed in afni.
Can any one enlighten me, what am I missing?

I hope this renders properly, the forum is rendering badly for me for some reason today.

Thanks,

Colm



Edited 1 time(s). Last edit at 11/02/2015 07:32PM by Colm Connolly.