Question about use of confounds in GLM_Flex_Fast4

[Nicolas Lori]

Dear Aaron,

  I am sending this email to ask you about the possibility of using confounds in the the fMRI data-processing in GLM_Flex_Fast4. The reason for the question is that for each of the subjects I would like to include age as a confound (age data is stored in a Excel file), and I could not find any description in the GLM_Flex_Fast websites about how to use confounds.

  The data-processing I am doing is working (without confounds), and it is the one below.

  Can you explain me how to include a confound such as age in the data-processing below?

Thank you for your attention,

Nicolas

_________________________

Scans={

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_007/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_007/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_007/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_009/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_009/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_009/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_023/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_023/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_023/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_024/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_024/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_024/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_025/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_025/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_025/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_031/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_031/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_031/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_032/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_032/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_032/GM1/1st_level_newB/con_0009.nii',

%Controls

%Patients

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_005/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_005/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_005/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_010/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_010/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_010/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_011/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_011/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_011/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_013/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_013/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_013/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_014/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_014/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_014/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_017/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_017/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_017/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_018/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_018/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_018/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_019/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_019/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_019/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_027/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_027/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_027/GM1/1st_level_newB/con_0009.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_030/GM1/1st_level_newB/con_0007.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_030/GM1/1st_level_newB/con_0008.nii',

'/media/neuroimaging/Untitled/GCC_tasks/MRI_GCC_030/GM1/1st_level_newB/con_0009.nii',

};

%7 Controls and 10 Patients

clear dat;

dat.fn=[];

dat.Group=[];

%dat.Run=[];

dat.Condition=[];

dat.SS=[];

sub=1;

cond=1;

for ii=1:numel(Scans)

if(cond>3) cond=1; sub=sub+1; end

if(cond==1) dat.Condition{ii,1} = 'Sit1';end;

if(cond==2) dat.Condition{ii,1} = 'Sit2';end;

if(cond==3) dat.Condition{ii,1} = 'Sit3';end;

dat.fn{ii,1} = Scans{ii};

dat.SS{ii,1} = num2str(sub);

%dat.Run{ii,1} = sub;

cond=cond+1;

end

dat.Group=cell(51,1);

dat.Group(1:21) = {'Control'};

dat.Group(22:51) = {'Patients'};

mkdir '/media/neuroimaging/Untitled/GCC_tasks/GCC_IFIS/GM_Output/MixedDesign_LS_LScomp_cond_v1'

clear I;

I.Scans=dat.fn;

I.Model = 'Group*Condition + random(SS|Condition)';

I.Data = dat;

I.Mask='/media/neuroimaging/Untitled/GCC_tasks/aal.nii';

I.OutputDir = '/media/neuroimaging/Untitled/GCC_tasks/GCC_IFIS/GM_Output/MixedDesign_LS_LScomp_cond_v1';

I.RemoveOutliers = 0;

I.DoOnlyAll = 1;

I.estSmooth = 1;

%I.PostHocs={};

GLM_Flex_Fast4(I);

__________________________

Nicolás Francisco Lori, PhD

University of Minho

ICVS | School of Medicine

Campus Gualtar

4710-057 Braga

Portugal

Professional email: nicol...@med.uminho.pt

[Aaron Schultz]

Any variable can be added to the model to control for a confound.  For instance x ~ y + age will control for age when assessing the relationship between x and y.  Importantly in the context of repeated measures, adding in static covariates has no effect.  If you have 4 conditions per subject, the age is the same for all four observations, and will no effect on the within subject contrasts.  If you have a repeated covariate, things can get a little complicated as you would be running a repeated measures ANCOVA.  If you are interested if the difference between two conditions is related to age, then you need to make a contrast of those conditions and then run a between subjects analysis on the within subject differences.

-Aaron