Hi Darren, I have similar questions regarding the negative connectivity-behavior correlations, did you get any answers to this?
Thanks!
Thanks!
↧
RE: Negative connectivity-behavior correlations
↧
Error in importing ROI nifti files
Hi all,
I'm wondering if I could get some help or an explanation as to why I keep getting the error below, and how I may be able to fix it:
ERROR DESCRIPTION:
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
CONN v.17.f
SPM12 + CSPM DEM Diffusion FieldMap MEEGtools Volumes aal cat12 conn marsbar render3d suit
Matlab v.2018a
storage: 2619.8Gb available
spm @ C:\matlab\spm12
conn @ C:\matlab\spm12\toolbox\conn
What I have done is create many volume of interest's (VOI's) of relevant brain areas involved in a certain neural pathway, from the T1-MRI nifti file of the average of my subject population (which is in MNI space), using MRIcron. I want to do a functional connectivity analysis of these VOI's using the CONN toolbox, however, when I import these VOI's as nifti files into the 'ROI's' section of Setup, the above error keeps popping up. I'm able to go through to second-level analysis without these VOI's; a problem only comes up when I start adding them into the 'ROI's' section.
I will attach the Anterior Insula VOI as an example below for you guys to see what I am inputting.
I would be extremely grateful to anyone who could help me troubleshoot and/or fix this problem.
Kind regards,
Fernando
I'm wondering if I could get some help or an explanation as to why I keep getting the error below, and how I may be able to fix it:
ERROR DESCRIPTION:
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
ERROR: Subject 1 Session 1 ROI 9 file has not been defined
ERROR: Subject 1 Session 1 ROI 20 file has not been defined
ERROR: Subject 1 Session 1 ROI 29 file has not been defined
CONN v.17.f
SPM12 + CSPM DEM Diffusion FieldMap MEEGtools Volumes aal cat12 conn marsbar render3d suit
Matlab v.2018a
storage: 2619.8Gb available
spm @ C:\matlab\spm12
conn @ C:\matlab\spm12\toolbox\conn
What I have done is create many volume of interest's (VOI's) of relevant brain areas involved in a certain neural pathway, from the T1-MRI nifti file of the average of my subject population (which is in MNI space), using MRIcron. I want to do a functional connectivity analysis of these VOI's using the CONN toolbox, however, when I import these VOI's as nifti files into the 'ROI's' section of Setup, the above error keeps popping up. I'm able to go through to second-level analysis without these VOI's; a problem only comes up when I start adding them into the 'ROI's' section.
I will attach the Anterior Insula VOI as an example below for you guys to see what I am inputting.
I would be extremely grateful to anyone who could help me troubleshoot and/or fix this problem.
Kind regards,
Fernando
↧
↧
error while doing preprocessing
why conn ran into unexpected error everytime i try doing preprocessing.
↧
How to add new subjects to existing project with a batch script (or merge)?
Hi,
I've been trying for 2 weeks to figure out how to add in subjects with out having to use the GUI to input the file names. Once I get the files imported, I would like to use the GUI to run preprocessing steps, as it is already set up in my project. This seems like it should be very simple to do, but I have tried 3 different methods and none have been working (see below). Help with any of these methods would be greatly appreciated! Thanks!
~Em
[b]1. [/b] I was going off a post saying that I could merge the new subjects with the project file using this method: https://www.nitrc.org/forum/message.php?msg_id=8613
So far this has not been able to work for me even after taking out the line in the merge file and opening in the GUI as it still gives me these problems:
Merging projects... please wait
loading info from project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched ROI names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched condition names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched first-level covariate names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched second-level covariate names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
There were problems importing the new subject data. See log or Matlab command window for details
[b]The next two methods are in the same script (see below)[/b]
[b]2. [/b]Using the BATCH.Setup function in the script, it gives me the following errors:
Error when using setup:
Brace indexing is not supported for variables of this type.
Error in conn_batch (line 842)
else CONN_x.Setup.structural{nsub}{nses}=conn_file(temp{min(numel(temp),nses)});
Error in Add_sub_190912 (line 79)
conn_batch(BATCH)
[b]3.[/b] When using the New function (these are commented out in the script), it doesn't error, and it seems to be doing something, but it ultimately doesn't add the subject files and they are just the last subjects files as if I just changed the number of subjects with out changing the files.
[b]THis is the script:[/b]
function [BATCH]=Add_sub_190912(sublist)
%This is a script for adding new subjects to the script
%Existing subs
subs_existing=24
len=length(sublist)
clear BATCH
BATCH.Setup.done=1 %setup is already done
BATCH.Setup.overwrite='No'
%BATCH.Setup.nsubjects=27 %set up total number of subjects (adding 3)
%for each subject
for i=1:len;
sub=sublist(i)
numsub=i+subs_existing
%Add in paths for structural data
T1path=strcat('Path_to_T1',sub,'/T1_data_office')
filelistT1=dir(string(T1path)); %create list of files in directory folder
filelistT1={filelistT1.name} %make an array of the file names that can be searched by regular expressions
filelistT1_str=strjoin(filelistT1) %make into single string
T1file=regexp(filelistT1_str,'\<20\d\d...._......T1MPRAGEsagisos........\.nii','match')
T1fullfile=char(fullfile(T1path,T1file))
[b]3. [/b]%BATCH.New.structurals{numsub}=fullfile(T1path,T1file)
[b]2. [/b]BATCH.Setup.structurals{numsub}=fullfile(T1path,T1file)
%for each run of the fMRI data
for r=1:3;
rr=string(r)
fMRIpath=strcat('Path_to_fMRI',sub,'/fMRI_data_office/task2_run',rr)
filelistfMRI=dir(string(fMRIpath)); %create list of files in directory folder
filelistfMRI={filelistfMRI.name} %make an array of the file names that can be searched by regular expressions
filelistfMRI_str=strjoin(filelistfMRI) %make into single string
fMRIfile=regexp(filelistfMRI_str,'\<20......_......ep2dbold2x2x2task2run.........\.nii','match') %get file name
fMRIfullfile=fullfile(fMRIpath,fMRIfile) %make a printed file path
fMRIfullfile=char(fMRIfullfile) %it wants you to convert it to characters from some reason from string
[b]2. [/b]BATCH.Setup.functionals{numsub}{r}=fMRIfullfile
%BATCH.Setup.voxelresolution=1; %This is the default value, probably don't need
[b]3. [/b]% BATCH.New.functionals{numsub}{r}=fullfile(fMRIpath,fMRIfullfile)
end
end
conn_batch(BATCH)
I've been trying for 2 weeks to figure out how to add in subjects with out having to use the GUI to input the file names. Once I get the files imported, I would like to use the GUI to run preprocessing steps, as it is already set up in my project. This seems like it should be very simple to do, but I have tried 3 different methods and none have been working (see below). Help with any of these methods would be greatly appreciated! Thanks!
~Em
[b]1. [/b] I was going off a post saying that I could merge the new subjects with the project file using this method: https://www.nitrc.org/forum/message.php?msg_id=8613
So far this has not been able to work for me even after taking out the line in the merge file and opening in the GUI as it still gives me these problems:
Merging projects... please wait
loading info from project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched ROI names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched condition names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched first-level covariate names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
Mismatched second-level covariate names with project File_Pathnew_subs/190905_newsubs/conn_addsubs.mat
There were problems importing the new subject data. See log or Matlab command window for details
[b]The next two methods are in the same script (see below)[/b]
[b]2. [/b]Using the BATCH.Setup function in the script, it gives me the following errors:
Error when using setup:
Brace indexing is not supported for variables of this type.
Error in conn_batch (line 842)
else CONN_x.Setup.structural{nsub}{nses}=conn_file(temp{min(numel(temp),nses)});
Error in Add_sub_190912 (line 79)
conn_batch(BATCH)
[b]3.[/b] When using the New function (these are commented out in the script), it doesn't error, and it seems to be doing something, but it ultimately doesn't add the subject files and they are just the last subjects files as if I just changed the number of subjects with out changing the files.
[b]THis is the script:[/b]
function [BATCH]=Add_sub_190912(sublist)
%This is a script for adding new subjects to the script
%Existing subs
subs_existing=24
len=length(sublist)
clear BATCH
BATCH.Setup.done=1 %setup is already done
BATCH.Setup.overwrite='No'
%BATCH.Setup.nsubjects=27 %set up total number of subjects (adding 3)
%for each subject
for i=1:len;
sub=sublist(i)
numsub=i+subs_existing
%Add in paths for structural data
T1path=strcat('Path_to_T1',sub,'/T1_data_office')
filelistT1=dir(string(T1path)); %create list of files in directory folder
filelistT1={filelistT1.name} %make an array of the file names that can be searched by regular expressions
filelistT1_str=strjoin(filelistT1) %make into single string
T1file=regexp(filelistT1_str,'\<20\d\d...._......T1MPRAGEsagisos........\.nii','match')
T1fullfile=char(fullfile(T1path,T1file))
[b]3. [/b]%BATCH.New.structurals{numsub}=fullfile(T1path,T1file)
[b]2. [/b]BATCH.Setup.structurals{numsub}=fullfile(T1path,T1file)
%for each run of the fMRI data
for r=1:3;
rr=string(r)
fMRIpath=strcat('Path_to_fMRI',sub,'/fMRI_data_office/task2_run',rr)
filelistfMRI=dir(string(fMRIpath)); %create list of files in directory folder
filelistfMRI={filelistfMRI.name} %make an array of the file names that can be searched by regular expressions
filelistfMRI_str=strjoin(filelistfMRI) %make into single string
fMRIfile=regexp(filelistfMRI_str,'\<20......_......ep2dbold2x2x2task2run.........\.nii','match') %get file name
fMRIfullfile=fullfile(fMRIpath,fMRIfile) %make a printed file path
fMRIfullfile=char(fMRIfullfile) %it wants you to convert it to characters from some reason from string
[b]2. [/b]BATCH.Setup.functionals{numsub}{r}=fMRIfullfile
%BATCH.Setup.voxelresolution=1; %This is the default value, probably don't need
[b]3. [/b]% BATCH.New.functionals{numsub}{r}=fullfile(fMRIpath,fMRIfullfile)
end
end
conn_batch(BATCH)
↧
RE: How to remove first-order derivatives of realignment parameters in conn_batch_humanconnectomeproject's
[b]Dear Alfonso,[/b]
The problem solved, the "batch.Denoising.confounds" command works perfect!
Thank you!
Best,
Mohammad
The problem solved, the "batch.Denoising.confounds" command works perfect!
Thank you!
Best,
Mohammad
↧
↧
Effect size in seed-to-voxel multiple regression analysis?
Dear CONN experts,
I carried out a multiple regression analysis to define effects of a behavioral covariate on seed-to-voxel functional connectivity of a seed. Between-subjects contrast was:
All Subjects, behavioral covariate, other covariates
[0 1 0]
My question is how to define effect size of this effect, which would be interpretable in terms of "low" or "great effect size". When I open effect sizes (Fisher-transformed z-values) in CONN, it is 0.08. But what does it mean? Is it a correlation coefficient between connectivity values and the covariate? Is this effect large, medium or low?
To clarify, I performed post hoc ROI-to-ROI analysis to derive connectivity values between the predefined seed and result cluster; the average connectivity value was -0.05. When I calculate Pearson r between the connectivity values and the behavioral covariate, the r is 0.37. So what is the correct measure of the effect size in this case?
Would be grateful for any help,
Yana
I carried out a multiple regression analysis to define effects of a behavioral covariate on seed-to-voxel functional connectivity of a seed. Between-subjects contrast was:
All Subjects, behavioral covariate, other covariates
[0 1 0]
My question is how to define effect size of this effect, which would be interpretable in terms of "low" or "great effect size". When I open effect sizes (Fisher-transformed z-values) in CONN, it is 0.08. But what does it mean? Is it a correlation coefficient between connectivity values and the covariate? Is this effect large, medium or low?
To clarify, I performed post hoc ROI-to-ROI analysis to derive connectivity values between the predefined seed and result cluster; the average connectivity value was -0.05. When I calculate Pearson r between the connectivity values and the behavioral covariate, the r is 0.37. So what is the correct measure of the effect size in this case?
Would be grateful for any help,
Yana
↧
ICA analysis across multiple conditions
Hi,
I have a methodology question about how ICA works. In CONN, I understand that ICA networks are determined by concatenating each subject and each condition temporally, and then using Fast ICA to find the independent spatial components based on this dataset.
In my experiment, I have a baseline resting state scan, a post motor performance task resting state scan, and a post motor learning task resting state scan. We expect there to be differences in functional connectivity in the 3rd post learning condition. My question is, how does the change in functional connectivity, which changes how voxels in the third condition are associated, affect the identification of spatial components. If for example, in the first two scans, source A talks to source B, but in scan 3 source A talks to source C. Will ICA be able to identify this shift in network connectivity? Or should I do ICA on each condition separately and compare relevant networks?
Best,
Andria
I have a methodology question about how ICA works. In CONN, I understand that ICA networks are determined by concatenating each subject and each condition temporally, and then using Fast ICA to find the independent spatial components based on this dataset.
In my experiment, I have a baseline resting state scan, a post motor performance task resting state scan, and a post motor learning task resting state scan. We expect there to be differences in functional connectivity in the 3rd post learning condition. My question is, how does the change in functional connectivity, which changes how voxels in the third condition are associated, affect the identification of spatial components. If for example, in the first two scans, source A talks to source B, but in scan 3 source A talks to source C. Will ICA be able to identify this shift in network connectivity? Or should I do ICA on each condition separately and compare relevant networks?
Best,
Andria
↧
Including FieldMap in preprocessing
Dear all,
I'm a naive CONN user, so I'm sorry if my question is very basic.
Since we acquired FieldMap phase&magnitude files, I would like to include them in the first step of my resting-state data preprocessing. How can I do it? Is there a way to select them via GUI? I have FieldMaps for each subject.
Thank you for your help,
Bye! :)
I'm a naive CONN user, so I'm sorry if my question is very basic.
Since we acquired FieldMap phase&magnitude files, I would like to include them in the first step of my resting-state data preprocessing. How can I do it? Is there a way to select them via GUI? I have FieldMaps for each subject.
Thank you for your help,
Bye! :)
↧
Generate resting state images using CONN toolbox
We are searching for steps to generate a resting state graph using CONN toolbox.
Is there any specific documentation that we can refer to ?
Is there any specific documentation that we can refer to ?
↧
↧
Data File Missing -- Can I get its full path?
After some changes to the filesystem, CONN reports a missing file ("CG.nii").
Is there a way I can recover the full path of that missing file, rather than just the filename?
Is there a way I can recover the full path of that missing file, rather than just the filename?
↧
RE: Error in Preprocessing
I continue to get similar error messages during processing. The following is after running Denoising step. I was able to get past the preprocessing error by manually inputting realigned files and rp_Funct.txt covariates and then continuing with preprocessing pipeline. But I'm not sure what these error messages mean and whether they are affecting the later analysis steps.
Error after Denoising:
ERROR DESCRIPTION:
Error using matlab.ui.control.UIControl/set
Invalid or deleted object.
Error in conn_disp (line 157)
set(CONN_h.screen.hlogstr,'value',[]);
Error in conn (line 878)
conn_disp('fprintf','saved %s\n',localfilename);
Error in conn (line 4361)
conn('save',filename);
Error in conn (line 5910)
conn gui_setup_save;
Error in conn_menumanager (line 120)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN18.b
SPM12 + DEM FieldMap MEEGtools
Matlab v.2017b
project: CONN18.a
Error after Denoising:
ERROR DESCRIPTION:
Error using matlab.ui.control.UIControl/set
Invalid or deleted object.
Error in conn_disp (line 157)
set(CONN_h.screen.hlogstr,'value',[]);
Error in conn (line 878)
conn_disp('fprintf','saved %s\n',localfilename);
Error in conn (line 4361)
conn('save',filename);
Error in conn (line 5910)
conn gui_setup_save;
Error in conn_menumanager (line 120)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN18.b
SPM12 + DEM FieldMap MEEGtools
Matlab v.2017b
project: CONN18.a
↧
RE: first level threshold
Hello,
I am also interested in analyzing stats at the first-level. I tried running the code that Alfonso posted in this thread but got the following error:
"Struct contents reference from a non-struct array object.
Error in untitled2 (line 7)
filepath=CONN_x.folders.preprocessing;"
Did anybody have any luck with running this code or find another way to produce multiple-corrected maps at the first-level?
Thanks!
I am also interested in analyzing stats at the first-level. I tried running the code that Alfonso posted in this thread but got the following error:
"Struct contents reference from a non-struct array object.
Error in untitled2 (line 7)
filepath=CONN_x.folders.preprocessing;"
Did anybody have any luck with running this code or find another way to produce multiple-corrected maps at the first-level?
Thanks!
↧
Error message in slice viewer
I have received an error message, whe I choise functional tools: slice viewer (I have matlab 2019b). I get the following error:
Error using strfind
PATTERN must be a string scalar or character vector.
Error in contains (line 36)
tf = ~cellfun(@isempty, strfind(s, pattern));
Error in matlab.ui.internal.prepareFigureFor
Error in rotate3d (line 171)
matlab.ui.internal.prepareFigureFor(f, mfilename('fullpath'));
Error in conn_slice_display (line 451)
set(rotate3d,'ActionPostCallback',{@conn_slice_display_refresh,'position'});
Error in conn (line 1765)
elseif val==3, fh=conn_slice_display(CONN_x.Setup.structural{nsub}{nsesstemp(1)}{1},temp1,[],[nan 0],sprintf('dataset %d',nset));
CONN18.b
SPM12 + DEM FieldMap MEEGtools
Matlab v.2019b
project: CONN18.b
storage: 388.3Gb available
spm @ C:\tools\spm
conn @ C:\tools\conn
Thanks
Error using strfind
PATTERN must be a string scalar or character vector.
Error in contains (line 36)
tf = ~cellfun(@isempty, strfind(s, pattern));
Error in matlab.ui.internal.prepareFigureFor
Error in rotate3d (line 171)
matlab.ui.internal.prepareFigureFor(f, mfilename('fullpath'));
Error in conn_slice_display (line 451)
set(rotate3d,'ActionPostCallback',{@conn_slice_display_refresh,'position'});
Error in conn (line 1765)
elseif val==3, fh=conn_slice_display(CONN_x.Setup.structural{nsub}{nsesstemp(1)}{1},temp1,[],[nan 0],sprintf('dataset %d',nset));
CONN18.b
SPM12 + DEM FieldMap MEEGtools
Matlab v.2019b
project: CONN18.b
storage: 388.3Gb available
spm @ C:\tools\spm
conn @ C:\tools\conn
Thanks
↧
↧
How to compute a correlation between FC within a network and a quantitative data set.
Hello all,
I would like to test the following: association between FC within a resting state network (DMN for example), and ratings of how people feel. I have my RS networks calculated, and am able to see how to export values of clusters, but am unable to figure out how to get a value for the [i]entire [/i]RS network. Is this possible in conn? Any help would be much appreciated,
Best,
Natasha
I would like to test the following: association between FC within a resting state network (DMN for example), and ratings of how people feel. I have my RS networks calculated, and am able to see how to export values of clusters, but am unable to figure out how to get a value for the [i]entire [/i]RS network. Is this possible in conn? Any help would be much appreciated,
Best,
Natasha
↧
Distribution of voxel-to-voxel connectivity values looks very similar pre- and post-denoising
Hi all,
I've run my preprocessing steps and about to move on to denoising, but I've noticed something strange. The distribution of connectivity values in the original looks very similar to the distribution predicted for after denoising. Whereas in the tutorial videos, the distributions are hugely skewed before denoising, the distributions of my original data already all have roughly a mean value of zero [i]before[/i] denoising (see attached plots).
Does this indicated an error in the preprocessing steps somewhere, or a problem with my scans? Or is it nothing to worry about?
Note: I didn't use the default preprocessing pipeline. My analysis is volume-based but calls for subject-space, unsmoothed data, with confounds covaried out. We're preprocessing and denoising in CONN but using FSL and R for the rest. My pipeline was as follows:[list][*]Functional realignment and unwarp (subject motion estimation and correction)[*]Functional slice-timing correction[*]Functional outlier detection (ART-based identification of outlier scans for scrubbing)[*]Functional Direct Coregistration to structural without reslicing (rigid body transformation)[*]Functional Segmentation (Grey/White/CSF segmentation)[*]Structural Segmentation (Grey/White/CSF estimation)[/list]Many thanks in advance!
Best,
Willem
I've run my preprocessing steps and about to move on to denoising, but I've noticed something strange. The distribution of connectivity values in the original looks very similar to the distribution predicted for after denoising. Whereas in the tutorial videos, the distributions are hugely skewed before denoising, the distributions of my original data already all have roughly a mean value of zero [i]before[/i] denoising (see attached plots).
Does this indicated an error in the preprocessing steps somewhere, or a problem with my scans? Or is it nothing to worry about?
Note: I didn't use the default preprocessing pipeline. My analysis is volume-based but calls for subject-space, unsmoothed data, with confounds covaried out. We're preprocessing and denoising in CONN but using FSL and R for the rest. My pipeline was as follows:[list][*]Functional realignment and unwarp (subject motion estimation and correction)[*]Functional slice-timing correction[*]Functional outlier detection (ART-based identification of outlier scans for scrubbing)[*]Functional Direct Coregistration to structural without reslicing (rigid body transformation)[*]Functional Segmentation (Grey/White/CSF segmentation)[*]Structural Segmentation (Grey/White/CSF estimation)[/list]Many thanks in advance!
Best,
Willem
↧
RE: denoising (compcor) is shifting all correlation histogram means to zero
Dear functional conn comunity,
I have been wondering the same thing. Denoising seems to shift the distributions towards having mean 0 (although in most cases there is a longer tail on the positive side). This will induce "spurious" negative correlations as well as somewhat change the output of weighted graph theory measures i imagine.
I am not sure what part of the denosing may produce this (e.g., simple movement within the scanner may produce higher correlations). I might investigate myself if I find time; but in the mean time any comment or discussion is welcome.
Happy connectivity
Peter
I have been wondering the same thing. Denoising seems to shift the distributions towards having mean 0 (although in most cases there is a longer tail on the positive side). This will induce "spurious" negative correlations as well as somewhat change the output of weighted graph theory measures i imagine.
I am not sure what part of the denosing may produce this (e.g., simple movement within the scanner may produce higher correlations). I might investigate myself if I find time; but in the mean time any comment or discussion is welcome.
Happy connectivity
Peter
↧
ICA analyses
Dear Conn users,
I have a few questions of clarification regarding ICA analyses. I would like to specify that I have explored Conn's forum on many occasions, but I have not found a definite answer to my interrogations.
Here's the scenario: I'm looking at resting-state data from two populations (patients and controls), and I am investigating for different functional network connectivity between these two groups.
I ran an ICA network analysis and selected 6 components of interest. But now the challenge of group comparison for these components has appeared.
For what I read on the forum, if I compare groups directly for a component at the second level analysis, I understood that I get the difference between my groups of the connectivity between the Network/Component and every voxel in the brain. Thus, I should only obtain voxels outside of my Component network, right? Which is the same results that I will obtain if I reinject my ICA component in the ROI (as non-thresholded image, by default) a the setup stage. Is that correct? And if yes ... can you help me interpreting this result ?
Thus, if I need to compare the difference between groups inside the component of interest, what analysis should I run?
Also, if I want to directly compare components to other components, I should use ROI-to-ROI analysis. But, is Conn toolbox integrates a temporal lag in its analysis (like described in Arbabshirani, HBP, 2013)? And if not, is there a way of doing it?
Last question: Is it correct to the use non-binary masks created by Conn by using "ICA parcellation ROI file" for the type of analysis we are interested here?
I hope I haven't lost you on the way with all these questions.
I hope that a conn-artist will help me!!!
Thank you in advance,
Johanna
I have a few questions of clarification regarding ICA analyses. I would like to specify that I have explored Conn's forum on many occasions, but I have not found a definite answer to my interrogations.
Here's the scenario: I'm looking at resting-state data from two populations (patients and controls), and I am investigating for different functional network connectivity between these two groups.
I ran an ICA network analysis and selected 6 components of interest. But now the challenge of group comparison for these components has appeared.
For what I read on the forum, if I compare groups directly for a component at the second level analysis, I understood that I get the difference between my groups of the connectivity between the Network/Component and every voxel in the brain. Thus, I should only obtain voxels outside of my Component network, right? Which is the same results that I will obtain if I reinject my ICA component in the ROI (as non-thresholded image, by default) a the setup stage. Is that correct? And if yes ... can you help me interpreting this result ?
Thus, if I need to compare the difference between groups inside the component of interest, what analysis should I run?
Also, if I want to directly compare components to other components, I should use ROI-to-ROI analysis. But, is Conn toolbox integrates a temporal lag in its analysis (like described in Arbabshirani, HBP, 2013)? And if not, is there a way of doing it?
Last question: Is it correct to the use non-binary masks created by Conn by using "ICA parcellation ROI file" for the type of analysis we are interested here?
I hope I haven't lost you on the way with all these questions.
I hope that a conn-artist will help me!!!
Thank you in advance,
Johanna
↧
↧
RE: Pending jobs after SGE processing
[color=#000000]Hi Eugenio,[/color]
[color=#000000]I am curious about this, it seems like perhaps some issue with file permissions? (e.g. if CONN is not able to delete .dmat files generated by the SGE scheduler in your projects' directory). I know you already posted that this was solved, but I was wondering if you can tell me more details about why this was failing (just to see if I can try to avoid other users from running into a similar scenario in the the future)[/color]
[color=#000000]Thanks![/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Eugenio Abela:[/i][quote]Hi all
I've performed the Setup step using a SGE computing environment to separate into several jobs (n=20, total of 180 subjects). This runs smoothly, jobs are (apparently) merged, and I can move on to the Denoising step. However, once I want to start Denoising, Conn tells me there are still pending jobs, and starts merging again. This then takes ages, and Conn then saves, but jobs remain pending. I'm at loss as to why this happens. Has anyone run into similar issues?
Thanks for any pointers!
Eugenio[/quote]
[color=#000000]I am curious about this, it seems like perhaps some issue with file permissions? (e.g. if CONN is not able to delete .dmat files generated by the SGE scheduler in your projects' directory). I know you already posted that this was solved, but I was wondering if you can tell me more details about why this was failing (just to see if I can try to avoid other users from running into a similar scenario in the the future)[/color]
[color=#000000]Thanks![/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Eugenio Abela:[/i][quote]Hi all
I've performed the Setup step using a SGE computing environment to separate into several jobs (n=20, total of 180 subjects). This runs smoothly, jobs are (apparently) merged, and I can move on to the Denoising step. However, once I want to start Denoising, Conn tells me there are still pending jobs, and starts merging again. This then takes ages, and Conn then saves, but jobs remain pending. I'm at loss as to why this happens. Has anyone run into similar issues?
Thanks for any pointers!
Eugenio[/quote]
↧
RE: Aberrant FC distributions following denoising
Hi Alfonso,
The dof after denoising are between about 9 and 13. Would you consider that too low? Notably, it is not the case that the runs with the fewest dof are also the runs with strange histograms. If I do not filter the 24 realignment regressors, then the dof jump up to about 30, and the peaked histograms appear more normal. However, structured noise also reappears in the after-denoising carpet plot. I should also note that the dof are similar in the SPM preprocessing version. Looking forward to your input
Thank you,
Walter
The dof after denoising are between about 9 and 13. Would you consider that too low? Notably, it is not the case that the runs with the fewest dof are also the runs with strange histograms. If I do not filter the 24 realignment regressors, then the dof jump up to about 30, and the peaked histograms appear more normal. However, structured noise also reappears in the after-denoising carpet plot. I should also note that the dof are similar in the SPM preprocessing version. Looking forward to your input
Thank you,
Walter
↧
Task-based gPPI a ROI-to-Voxel questions
Hi all,
I have read all the related posts in gPPI and I have got lots of my question answered.
I am using gPPI in CONN (seed-voxel) to compare between two groups (Young and Old) the association between FC and fractional anisotropy (FA).
My 2nd level covariates might look like:
Young: [1 1 1 1 1 1 0 0 0 0]
Old: [0 0 0 0 0 0 1 1 1 1]
young_FA: [.1 .2 .3 .3 .4 .5 .6 0 0 0 0]
old_FA: [0 0 0 0 .1 .2 .3 .3 .4]
To get and compare between young and old in the association between FC and fractional anisotropy (FA), I have set the contrast [0 0 1 -1] to the between subject contrasts and [1-1] for in between condition contrasts. I have computed the average connectivity values in REX and the result looks like the attached figure. My questions are:
1: Whether my approach is right in answering the association between FA and FC and be able to show the underlying difference b/n young and old
2: from the effect size, the beta values appears very large. Is that acceptable for such comparison?
3: for extracting the stats in the [tool:calculator] I still get T stats for [0 0 1 -1], even-though in the second level analysis it was an ANCOVA model.
Any thoughts are appreciated.
Thank you!
Abiot
I have read all the related posts in gPPI and I have got lots of my question answered.
I am using gPPI in CONN (seed-voxel) to compare between two groups (Young and Old) the association between FC and fractional anisotropy (FA).
My 2nd level covariates might look like:
Young: [1 1 1 1 1 1 0 0 0 0]
Old: [0 0 0 0 0 0 1 1 1 1]
young_FA: [.1 .2 .3 .3 .4 .5 .6 0 0 0 0]
old_FA: [0 0 0 0 .1 .2 .3 .3 .4]
To get and compare between young and old in the association between FC and fractional anisotropy (FA), I have set the contrast [0 0 1 -1] to the between subject contrasts and [1-1] for in between condition contrasts. I have computed the average connectivity values in REX and the result looks like the attached figure. My questions are:
1: Whether my approach is right in answering the association between FA and FC and be able to show the underlying difference b/n young and old
2: from the effect size, the beta values appears very large. Is that acceptable for such comparison?
3: for extracting the stats in the [tool:calculator] I still get T stats for [0 0 1 -1], even-though in the second level analysis it was an ANCOVA model.
Any thoughts are appreciated.
Thank you!
Abiot
↧