Quantcast
Channel: NITRC CONN : functional connectivity toolbox Forum: help
Viewing all 6864 articles
Browse latest View live
↧

MVPA posthoc contrast best practice question

January 31, 2020, 7:52 am
$
0
0
Good afternoon,

I have run an MVPA analysis which includes a measure of IQ as a nuisance variable. I then extracted the significant ROIs and am running posthoc seed-to-voxel analyses. On these posthoc analyses is it best to keep the exact same subject effects contrast, which would include my nuisance variables that I included in the MVPA? Or, since the MVPA was already run with the nuisance variables should I only include the variables of interest and not include any nuisance variables in the seed-to-voxel posthoc?

Thank you for the info!
↧
Search

MVPA Regression?

January 31, 2020, 8:51 am
$
0
0
Is it appropriate to use MVPA in a simple regression (correlation) or multiple regression model?  Any particular issues associated with this sort of model or point references using MVPA in CONN with these sorts of analyses?

Thx,
Jeff
↧
↧

ALFF and Results Explorer: Positive and Negative Contrasts for Two Group Two Condition Analyses, Exact Opposite Results Problem

January 31, 2020, 11:43 am
$
0
0
Hello everyone, 

I am running an amplitude of low frequency fluctuations (ALFF) analysis between two groups (healthy older adults -HO, mild cognitively impaired adults - MCI) under two conditions (dual-task, flexion). I am having difficulty interpreting the results in the results explorer between the two groups and two conditions. It seems like the cluster results of the groups and conditions are the exactly the same but flipped. Is one result more correct? 

Also, what do the positive and negative contrasts mean? Is the negative contrast just the result of the other group being compared. Thank you so much! 

Further explanation of the issue. 
Groups/Subject effects (2): 
Healthy older (HO)
Mild cognitively impaired (MCI) 

Conditions (2):
Dual-task (DT)
Flexion (F)

Between-subject contrast: 
HO>MCI

Between-conditions contrast: 
DT>F

Results: Looking at the effects of the dual-task>flexion between HO>MCI, shows positive activation for ClusterA, but flexion>dual-task, MCI>HO, also shows positive activation for ClusterA.

Is this possible? Is one result more relevant or appropriate to report? Thanks again! 

Best,
Coco Tirambulo 
University of Arizona
↧

Hanging on Step 5/7 Importing ROI data

February 3, 2020, 8:55 am
$
0
0
I've tried running the standard preprocessing pipeline now twice on my data of 32 subjects with resting state scans of up to 30 minutes. I received a message both times that it had successfully completed the preprocessing, but when I click "done" to go to the next step it starts the import process. It seems to have no problem with the initial steps (1: data completeness; 2: segmentation; 3: import conditions/covariates; 4: import functional data), but when it gets to step 5 (importing ROI data), it just seems to hang. Step 4 actually took quite a long time, but it gave me a status bar telling me which subject was being imported, and I could find the file that was being slowly written during the import. With Step 5, I am not getting any sort of update on which subject it is working on, and even after waiting a few hours I cannot find any files that have been created or edited since the step began. 

I'm running Matlab 2019 R2019b with Conn 18b. 

Does the import ROI just take a really really long time and I should just wait? Or might there be some undetected error that is causing it to hang? 

Thank you,
Josh
↧

Batch for preprocessing with a text-only linux environment

February 3, 2020, 9:01 am
$
0
0
Hello experts!

I am trying zu develope a BATCH script which i can use to preprocess (and later also denois) a data set with about 320 subjects at rest (to big for a personnel computer, so i need to use our cluster with text-only linux).

setting up and creating a conn project is working. unfortunately the preprocessing wont work. I wrote the following scrip to test it on my personnel computer. how do i implement the preprocessing step? the following doenst work. pop up windows aren't possible!!


which BATCH.Setup.preprocessing lines do i need to add?


addpath '/home/langhamt/spm12';
addpath '/home/langhamt/conn18b/conn';
addpath '/home/langhamt/PROTECT-AD/P4';

clear BATCH;
BATCH.filename='/home/langhamt/PROTECT-AD/P4/conn_Batch_test.mat';
BATCH.Setup.isnew=1;
BATCH.Setup.nsubjects=2;
BATCH.Setup.RT=2;
BATCH.Setup.acquisitiontype=1;
BATCH.Setup.functionals{1}{1}='/home/langhamt/PROTECT-AD/P4/testdata/patients/sub-1013/Ma2/func/RestingState/4D.nii';
BATCH.Setup.functionals{2}{1}='/home/langhamt/PROTECT-AD/P4/testdata/patients/sub-1017/Ma2/func/RestingState/4D.nii';
BATCH.Setup.structurals{1}='/home/langhamt/PROTECT-AD/P4/testdata/patients/sub-1013/Ma2/anat/anon_t1mprsagp2iso2141013Ma2s003a1001.nii';
BATCH.Setup.structurals{2}='/home/langhamt/PROTECT-AD/P4/testdata/patients/sub-1017/Ma2/anat/anon_t1mprsagp2iso2141017Ma2s003a1001.nii';
BATCH.Setup.analyses=[1,2,3];
BATCH.Setup.voxelmask=1;

BATCH.Setup.preprocessing.steps={'default_mni'};

conn_batch(BATCH);

conn save;




The following error occurs:

Error using fprintf
Invalid file identifier. Use fopen to generate a valid file identifier.

Error in conn_disp (line 162)
for n=1:numel(str), fprintf(fh,'%s\n',str{n}); end

Error in conn_process (line 27)
case 'setup_preprocessing', conn_disp(['CONN: RUNNING SETUP.PREPROCESSING STEP']); conn_setup_preproc(varargin{:});

Error in conn_batch (line 1099)
conn_process('setup_preprocessing',steps,'subjects',SUBJECTS,OPTIONS{:});

Error in BATCH_test_script (line 102)
conn_batch(BATCH);
>>
↧
↧

Mask seed-to-voxels analysis

February 3, 2020, 1:35 pm
$
0
0
Hello,

I noticed that significant clusters that I found with a seed-to-voxels analysis included sometimes a lot of voxels that were 'not labelled' (using the standard Harvard-Oxford atlas). Since I would like to find results only in the grey matter, I changed the analysis mask in the Setup, and put the referenceGM.nii file first but noticed there were still unlabelled voxels in the significant clusters (and noted that the mask included a lot a voxels that in average are considered as part of the WM - if I understood correctly CONN select all of them as long as the probability is >0, right?). Then I tried with the atlas.nii file, but same results, there were voxels outside of the atlas.

Would someone know how to really restrict the 2nd level analysis to voxels identified as belonging to the grey matter - or at least those included in an atlas?

Thanks in advance,
Lena
↧

MANCOVA interaction two drugs vs. two time points with 5 MVPA components

February 4, 2020, 4:21 am
$
0
0
Hi,

I'm trying to emulate the analysis performed in the paper "The association between antidepressant treatment and brain connectivity in two double-blind, placebo-controlled clinical trials: a treatment mechanism study" By Wang et al published in Lancet Psychiatry 2019.
In the methods section they state the following:
"[i]In the MVPA-derived maps, each voxel is populated with 64 factor scores summarising the voxel's connectivity with all other voxels in the brain. Following convention, we retained the first five of these 64 components, to yield MVPA-derived maps with each voxel ascribed five values indexing its whole-brain connectivity (the distribution of variance [80–85%] in whole-brain functional connectivity can be explained by five principal components; appendix).[/i]
[i]We then did a repeated-measure multivariate ANCOVA (MANCOVA) with each participant's five factor scores at each voxel (ie, MVPA-derived maps) as the dependent variables and with treatment and time as the two independent variables. Participants from both studies were included in the MANCOVA. Treatment was coded as either active treatment (duloxetine or desvenlafaxine) or placebo, and time was coded as either study baseline (ie, pretreatment) or endpoint (ie, post-treatment). We included RCT (duloxetine vs desvenlafaxine), age, and sex as covariates. The treatment-by-time interaction was isolated.[/i]"


I also have data including an active drug vs. placebo, with rsfMRI data before and after treatment. I have ran this through the Conn toolbox v18.b and I just want to confirm that I have set the GLM up appropriately to get my treatment*time interaction.
In the between-subjects contrast I have placebo>active drug [1 -1]
In the between-conditions contrast I have baseline<followup [-1 1]
In the between-measures contrast I have selected all 5 MVPA components and then selected the 'any effects (F-test)' [1 0 0 0 0; 0 1 0 0 0; 0 0 1 0 0; 0 0 0 1 0; 0 0 0 0 1]

Is this correct?

Many thanks,
Will
↧

Error at First-Level Analyses

February 4, 2020, 6:54 am
$
0
0
Hi all,

I've started running a new data set and I keep getting the following error message while running first-level analyses:

ERROR DESCRIPTION:

Error using conn_process (line 1947)
Incorrect number of ROIs for subject 2 condition 1. Please re-run Denoising step (all subjects/seeds)
Error in conn_process (line 42)
case 'analyses_gui_seedandroi',disp(['CONN: RUNNING ANALYSIS STEP (ROI-to-ROI or seed-to-voxel analyses)']); conn_process([10,11,15],varargin{:});
Error in conn (line 6535)
else conn_process('analyses_gui_seedandroi',CONN_x.Analysis); ispending=false;
Error in conn_menumanager (line 120)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});

Has anyone run into this before? I've done just about everything, including starting completely from scratch. I know it says there is an issue with the denoising step but I've re-run it several times. Does anyone have a fix for this?
↧
Search

RE: Mask seed-to-voxels analysis

February 4, 2020, 1:05 pm
$
0
0
[color=#000000]Hi Lena,[/color]

[color=#000000]Voxel-based second-level analyses are often restricted to a general "brainmask" search volume, but very rarely to a more specific gray-matter search volume (that is more typical of surface-based analyses instead). There are many reasons for this, mainly to accommodate functional and anatomical variability across subjects, but also to better match parametric Random Field Theory assumptions about the spatial properties of random effects, as well as to minimize differences in sensitivity across different spatial locations in the presence of highly non-homogeneous masking such as gray-matter masks. That said, a more reasonable alternative would be to run your second-level analysis normally but then apply a more restrictive mask (e.g. gray matter mask) to your results directly. One simple way to do this would be to use the 'SPM display' button in CONN, and then select in SPM a masking->from file option.[/color]

Hope this helps
Alfonso

[i]Originally posted by Lena Trebaul:[/i][quote]Hello,

I noticed that significant clusters that I found with a seed-to-voxels analysis included sometimes a lot of voxels that were 'not labelled' (using the standard Harvard-Oxford atlas). Since I would like to find results only in the grey matter, I changed the analysis mask in the Setup, and put the referenceGM.nii file first but noticed there were still unlabelled voxels in the significant clusters (and noted that the mask included a lot a voxels that in average are considered as part of the WM - if I understood correctly CONN select all of them as long as the probability is >0, right?). Then I tried with the atlas.nii file, but same results, there were voxels outside of the atlas.

Would someone know how to really restrict the 2nd level analysis to voxels identified as belonging to the grey matter - or at least those included in an atlas?

Thanks in advance,
Lena[/quote]
↧
↧

RE: Connectivity in Subject Space

February 4, 2020, 3:26 pm
$
0
0
Hi CONN Users,

I have reconstructed the structural image for each subject using Freesurfer. Also, I have an atlas for each subject and want to process connectivity using each subject's atlas. What would be the best way of doing this using CONN?
Thanks,
Tuner
↧

RE: using ART with slice timing corrected files

February 4, 2020, 7:52 pm
$
0
0
Hi Alfonso,

I'm using batch scripting and entered my functional volumes as such: 
batch.Setup.functionals{1}{1}='/pathToData/Raw/fmri_001.nii';

If I want to remove the first 4 volumes in the preprocessing step, can I do something like "batch.Setup.removevolumes = 4"? Do you know what the correct syntax is? 

Thanks,
Melanie
↧

RE: ALFF and Results Explorer: Positive and Negative Contrasts for Two Group Two Condition Analyses, Exact Opposite Results Problem

February 5, 2020, 3:17 pm
$
0
0
[color=#000000]Hi Coco,[/color]

[color=#000000]Right, that is perfectly fine, you are looking at the group-by-task interactions, and the "HO>MCI x DT>F" interaction is exactly the same as the "MCI>HO x F>DT" interaction. Even more, since one typically focuses on two-sided / two-tailed effects, the results would also be the same when using "HO>MCI x F>DT" or "MCI>HO x DT>F". A significant result here simply means that the differences in connectivity between DT and F are not the same when looking only within HO subjects vs. when looking only within MCI subjects. In order to properly interpret these results you should simply display the connectivity values for each of these four cases rather than relying only in the sign of the interaction effect.[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Coco Victoria Tirambulo:[/i][quote]Hello everyone, 

I am running an amplitude of low frequency fluctuations (ALFF) analysis between two groups (healthy older adults -HO, mild cognitively impaired adults - MCI) under two conditions (dual-task, flexion). I am having difficulty interpreting the results in the results explorer between the two groups and two conditions. It seems like the cluster results of the groups and conditions are the exactly the same but flipped. Is one result more correct? 

Also, what do the positive and negative contrasts mean? Is the negative contrast just the result of the other group being compared. Thank you so much! 

Further explanation of the issue. 
Groups/Subject effects (2): 
Healthy older (HO)
Mild cognitively impaired (MCI) 

Conditions (2):
Dual-task (DT)
Flexion (F)

Between-subject contrast: 
HO>MCI

Between-conditions contrast: 
DT>F

Results: Looking at the effects of the dual-task>flexion between HO>MCI, shows positive activation for ClusterA, but flexion>dual-task, MCI>HO, also shows positive activation for ClusterA.

Is this possible? Is one result more relevant or appropriate to report? Thanks again! 

Best,
Coco Tirambulo 
University of Arizona[/quote]
↧

RE: Test if BOLD signal in a specific ROI is correlated with my PET signal of a certain ROI

February 5, 2020, 3:21 pm
$
0
0
[color=#000000]Dear June,[/color]

[color=#000000]Selecting 'AllSubjects' and your 'PETvalues' covariate in the subject-effects list, and entering a [0 1] between-subjects contrast, will evaluate the correlation across subjects between the PET values and functional connectivity[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by June van Aalst:[/i][quote]Dear all,

I would like to investigate if my resting state BOLD signal in a certain ROI is correlated with the PET signal in a certain ROI.

For this analysis, I have resting state fMRI data of 20 subjects.

I can add for each ROI the ROI PET values as a second level covariate, but how do I test if these values are correlated with my resting state signal?

Thank you for your help!
June[/quote]
↧
↧

RE: Input to SVD must not contain NaN or Inf.

February 5, 2020, 3:23 pm
$
0
0
[color=#000000]Hi Matt,[/color]

[color=#000000]That is peculiar. It may be caused by having one of your first-level covariates containing either infinite or missing values (NaN). Which first-level covariates do you have in your conn project (and how were they entered/defined)?[/color]

[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by mgarlinghouse:[/i][quote][b]Hi Group,[/b]

[b]Has anyone encountered this error [see below] on a Mac before?  It occurs during step 5 / 7 when importing ROI Data.  I have traditionally used a windows machine and older versions of Matlab and CONN and have not encountered this error.......?[/b]

[b]Any thoughts on the topic are much appreciated.[/b]

[b]Best,[/b]

[b]Matt[/b]

----------------------------------------------------------------------------------------

Error using svd
Input to SVD must not contain NaN or Inf.

Error in pinv (line 18)
[U,S,V] = svd(A,'econ');
Error in rex>rex_do (line 746)
proj=eye(size(tdata,1))-[cov1,0*cov0]*pinv([cov1,cov0]); % removes covariates keeping scale unchanged
Error in rex (line 180)
[params.ROIdata,params.ROInames,params.ROIinfo.basis,params.ROIinfo.voxels,params.ROIinfo.files,params.ROIinfo.select,params.ROIinfo.trans]=rex_do(params,1);
Error in conn_rex (line 8)
[varargout{1:nargout}]=rex(varargin{:});
Error in conn_process (line 858)
else [data{nroi1},namesroi{nroi},params]=conn_rex(Vsourcethis,Vmask{nroi}{min(nses,nsesstemp)},'summary_measure','eigenvariate','dims',CONN_x.Setup.rois.dimensions{nroi},'conjunction_mask',mask,'level',level,'scaling',scalinglevel,'select_clusters',0,'covariates',entercovariates,'fsanatomical',fsanatomical,'output_type',outputtype,'output_rex',filenamerex,'output_folder',filepath);
Error in conn_process (line 16)
case 'setup', conn_disp(['CONN: RUNNING SETUP STEP']); conn_process([0:4,4.5,5]);
Error in conn (line 5262)
else conn_process(processname); ispending=false;
Error in conn_menumanager (line 120)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
[b]CONN18.b[/b]
[b]SPM12 + DEM FieldMap MEEGtools[/b]
[b]Matlab v.2018b[/b]
[b]project: CONN18.b[/b][/quote]
↧

RE: Error at vdm file generation

February 5, 2020, 3:32 pm
$
0
0
[color=#000000]Hi Ariana,[/color]

[color=#000000]Could you check whether perhaps you have entered a single file (B0_mag1.nii) instead of three files (B0_mag1, B0_mag2, and B0_phase) in the fmap secondary dataset for subject 2?[/color]

[color=#000000]Single fmap files are interpreted as pre-computed fieldmap files, while in your case what you have (if I understand correctly) are two magnitude and one phase-difference files from a double-echo sequence[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Ariana Familiar:[/i][quote]Hello,

I receive this error during VDM file generation for fieldmap distortion correction. I have separate files for the two magnitude and one phase images, and I used "read double-echo timing from BIDS / .json files" and "automatically determine" fieldmap type during setup. Please let me know how I can fix this - any insight would be appreciated!


[u]ERROR DESCRIPTION:[/u]

Error using *
Arguments must be 2-D, or at least one argument must be scalar. Use TIMES (.*) for elementwise multiplication.

Error in FieldMap (line 1659)
IP.vdm.vdm = IP.blipdir*IP.tert*1e-3*IP.fm.fpm;

Error in FieldMap_create (line 159)
[IP.vdm, IP.vdmP]=FieldMap('FM2VDM',IP);

Error in conn_setup_preproc (line 2463)
VDM = FieldMap_create(char(fmap),{filename},pm_def); %[ET1,ET2,0,ERT,-1]

Error in conn (line 1081)
ok=conn_setup_preproc('',varargin{2:end});

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.2019a
project: CONN18.b
storage: 411.9Gb available


[b][u]ADDITIONAL INFO:[/u][/b]

Preparing functional Creation of voxel-displacement map (VDM) for distortion correction
units of FieldMap /Users/ariana/Documents/face_value/data/session2/02_AS2/ref_run1.nii not found, assuming Hz
Creating vdm file for subject 2 session 1...
FieldMap : /Users/ariana/Documents/face_value/data/session2/02_AS2/B0_mag1.nii
ref : /Users/ariana/Documents/face_value/data/session2/02_AS2/ref_run1.nii
options.INPUT_DATA_FORMAT = PM
options.SHORT_ECHO_TIME = 10
options.LONG_ECHO_TIME = 12.46
options.MASKBRAIN = 1
options.UNWRAPPING_METHOD = Mark3D
options.FWHM = 10
options.PAD = 0
options.WS = 1
options.MFLAGS.TEMPLATE = /Users/ariana/Documents/MATLAB/spm12/toolbox/FieldMap/T1.nii
options.MFLAGS.FWHM = 5
options.MFLAGS.NERODE = 2
options.MFLAGS.NDILATE = 4
options.MFLAGS.THRESH = 0.5
options.MFLAGS.REG = 0.02
options.MFLAGS.GRAPHICS = 0
options.EPI_BASED_FIELDMAPS = 0
options.TOTAL_EPI_READOUT_TIME = 57.0004
options.DO_JACOBIAN_MODULATION = 0
options.sessname = session
options.match_vdm = 1
options.pedir = 2
options.defaultsfilename = /Users/ariana/Documents/MATLAB/spm12/toolbox/FieldMap/pm_defaults.m[/quote]
↧
Search

RE: Problem displaying surface-level analyses

February 5, 2020, 3:37 pm
$
0
0
Hi again Alfonso,

I re-conducted the surface-based analyses using the default surface preprocessing pipeline, as well as smoothing with 40 diffusion steps.

There are problems with displays in the QA plots (see attached).

To note, I removed the last preprocessing step since I entered freesurfer brainmask volumes with segmented grey matter, etc.

Thanks for your help,
Amy
↧

RE: Denoised data: order of niftiDATA_Subject*_Condition000.nii file

February 5, 2020, 3:39 pm
$
0
0
[color=#000000]Hi Christi[/color]

[color=#000000]The nifti*Condition000.nii file maintains the order of the original data (runs/sessions) independent of the order of the conditions (so, in your case, Video1, Video4, Video3, etc. assuming that those correspond with session1, session2, session3, etc.)[/color]

[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Christiana Westlin:[/i][quote]Hi Alfonso,

I am wondering if the niftiDATA_Subject*_Condition000.nii output from the denoising step maintains the order of the original timeseries, or if it becomes concatenated into the order conditions were entered?

In other words, if my original timeseries for one subject was ordered as Video 1, Video 4, Video 3, Video 2, will the denoising timeseries maintain this order or will it concatenate the timeseries for all subjects into the order the conditions were entered (Video1, Video2, Video3, Video4).

Thanks!!
Christi[/quote]
↧
↧

ART Toolbox vs aCompCor

February 5, 2020, 9:30 pm
$
0
0
Dear Alfonso,

Sorry not exactly a CONN related question.
I preprocessed my task-related fMRI data (block-design) using SPM8 and used the ART toolbox to determine potential outliers. I then ran GLM and gPPI analyses.
A reviewer is arguing that using this approach for artefact denoising, the global signal is included as nuisance regressor, which has shown to introduce negative correlations. Is this true? 

He suggested me to use the aCompCor approach instead. As far as I understand it, these 2 approaches are very different and do not address the same issues. Could you clarify?

Thanks for your time.

Best regards,

Yann
↧

Cost function masking with FSL MELODIC

February 6, 2020, 6:42 am
$
0
0
Hello, 
 I have a question about the application of cost function masking in a sample of tumor patients. I would like to run MELODIC tool to analyze my resting state fMRI data. We have already created the single-subjects lesion masks. 

Question 1: 
Is there any chance to include the lesion masks (or the structural image masked) in the FSL MELODIC GUI at the registration step? 

Question 2: 
If not, how can we find the single command lines for doing all the steps includind spatial preprocessing and ICA analysis (for example smoothing or slice-timing?). We didn't find them in the FSL help section. 

Many thanks

Best, 
Sara
↧

RE: Mask seed-to-voxels analysis

February 6, 2020, 8:04 am
$
0
0
Thank you Alfonso for your response, indeed it helps!
↧
Viewing all 6864 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>