Quantcast
Channel: NITRC CONN : functional connectivity toolbox Forum: help
Viewing all 6875 articles
Browse latest View live
↧

RE: longitudinal scans

June 10, 2020, 1:43 pm
$
0
0
[color=#000000]Just remember to create your two conditions, 'baseline' and 'followup', in the [i]Setup.Conditions [/i]tab, assigning them to their corresponding scans/sessions, and then proceed as normal.[/color]
[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]

[i]Originally posted by blj192:[/i][quote]Hello there, I am conducting an experiment with 2 groups (control, test) with 2 collected scans a few months apart (baseline, follow-up). Is there any specific things in setup I need to do in order to properly model longitudinal scans? I was wondering if there is anything different from normal analysis in conn to use. I am essentially interested in the contrast of Test (followup>baseline) > Control (followup>baseline). Thanks for the help! And if I should just proceed as normal that is a fine answer as well.[/quote]
↧
Search

RE: ROI construction and 2nd level analysis

June 10, 2020, 1:46 pm
$
0
0
[color=#000000]Hi Ishi,[/color]

[color=#000000]That error indicates that seed-to-voxel group-level analyses are not ready yet (either you have not yet finished running first-level seed-to-voxel analyses, or perhaps you have only run ROI-to-ROI first-level analyses and you simply need to switch to the ROI-to-ROI results in the second-level results tab)[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Ishi Tandon:[/i][quote]Hello experts,

I performed a VBM analysis of 68 individuals (containing 1 control group and 2 patients group- high and low).
Now since I wanted to look at the functional connectivity only of the clusters obtained in vbm, I am using conn.

I have performed initial preprocessing in Conn. However, when I construct ROIs though MNI co-ordinates in CONN and further run denoising, 1st level and 2nd level, MATLAB window shows the following error as 2nd level completes processing -
(I am not sure if I am going wrong with ROI construction)

Not ready to display second-level Analyses
Condition (rest ) has not been processed yet. Please re-run previous step (First-level analyses)
ERROR DESCRIPTION:
Reference to non-existent field 'Y'.
Error in conn (line 8615)
xyz=conn_convertcoordinates('idx2tal',prod(CONN_h.menus.m_results.Y(1).dim(1:2))*(CONN_h.menus.m_results.y.slice-1)+(1:prod(CONN_h.menus.m_results.Y(1).dim(1:2))),CONN_h.menus.m_results.Y(1).mat,CONN_h.menus.m_results.Y(1).dim);
Error in conn (line 7667)
if any(CONN_x.Analyses(CONN_x.Analysis).type==[2,3]), conn gui_results_s2v;
Error in conn_menumanager (line 120)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN19.c
SPM12 + AAL3 DEM FieldMap MEEGtools TFCE cat12 pTFCE suit
Matlab v.2017b
project: CONN19.c
storage: 1436.1Gb available
spm @ /Users/apple/Documents/spm12
conn @ /Users/apple/Documents/conn
ERROR DESCRIPTION:
Reference to non-existent field 'ncovariates'.
Error in conn (line 9395)
if model==1 || length(idx)~=length(CONN_h.menus.m_results.ncovariates) || any(idx~=CONN_h.menus.m_results.ncovariates),
CONN19.c
SPM12 + AAL3 DEM FieldMap MEEGtools TFCE cat12 pTFCE suit
Matlab v.2017b
project: CONN19.c
storage: 1436.1Gb available
spm @ /Users/apple/Documents/spm12
conn @ /Users/apple/Documents/conn

One more question regarding covariate (2nd level). , how will I construct covariates (2nd) if I have 3 groups?[/quote]
↧
↧

RE: Issues with surface-based processing

June 10, 2020, 1:58 pm
$
0
0
[color=#000000]Hi Kaitlin,[/color]

[color=#000000]Regarding (1), one possibility is that those ROIs (e.g. Frontal Pole) might end up having no valid functional data due to signal dropout and inhomogeneity distortions (e.g. near the sinuses, if those have not been properly corrected for using fieldmap information). Perhaps you could check that displaying your ROIs overlaid on top of your functional data? If this error happens widely (across ROIs that have no good reason to have any signal dropout), then that would more likely indicate some problem with the coregistration of your ROIs and your functional data (and the same visual check as above should help you determine whether this is the case here)[/color]

[color=#000000]Regarding (2), that indicates that you have either insufficient subjects to perform second-level analyses (or, less likely, perhaps you may have included too many factors in your second-level model?)[/color]

[color=#000000]Regarding (3): yes, exactly[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]

[i]Originally posted by Kaitlin Cassady:[/i][quote]Hello,

I am trying surface-based preprocessing/analyses for the first time. I ran into a few issues:

1) When I get to the first-level analysis, some ROIs seem to have no time series data (just a flat line; see screenshot). Why is this?

2) Also, if I ignore that (above) and just perform the first-level analysis, I get a warning message when I try to perform second-level analysis that says "WARNING: possibly incorrect model: insufficient degrees of freedom (suggestion: simplify second-level model)." How should I resolve this?

3) Finally, I assume that, under SETUP -> options, that I should use "Surface: same as template (Freesurfer fsaverage)" as the Analysis space (voxel-level) and the explicit mask called "mask.surface.brainmask.nii" as the Analysis mask (voxel-level). Is this correct?

Any help would be greatly appreciated Thank you!
Kaitlin[/quote]
↧

RE: How can I plot effect sizes for an ROI-to-ROI analysis in the standalone CONN version for Windows (18.b)

June 10, 2020, 2:13 pm
$
0
0
[color=#000000]Dear Tsev,[/color]

[color=#000000]In more recent versions (19b or above) the 'plot effects' option is also available in the ROI-to-ROI result explorer window (and it works just the same as the one in the seed-to-voxel results explorer window)[/color]

[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Tsvetoslav Ivanov:[/i][quote]Dear all,

I am analysing the connectivity between the PCC and the rest 31 ROIs from the 8 networks available in the standalone CONN version for Windows (18.b) using a one-sample test. I want to plot the effect sizes. However, the Results Explorer for the ROI-to-ROI analysis tab does not have the option to do it, unlike the Seed-to-Voxel tab, which does have this option. I tried using Export Mask in the Results Explorer for the ROI-to-ROI analysis tab to export a mask containing the 32 ROIs and then upload it through the REX interface in the Seed-to-Voxel tab by enablig REX GUI and then clicking the ROIs button to choose the mask file. 

There are several problems with this however. First, I can only export a mask from the Results Explorer for the ROI-to-ROI analysis tab with the extensions .mat, .txt, and .csv, while the REX gui only accepts .img, .nii, and .tal file formats. Second, even if I try to export a mask rom the Results Explorer for the ROI-to-ROI analysis tab with any of the available extensions, no file is saved in the directory.
 
Do you know if there is a way for me to plot the effect sizes of the connectivity between the PCC and the rest 31 ROIs from the 8 networks available in the standalone CONN version for Windows (18.b)?

I have also tried clicking on the "Plot effects" button after selecting the PCC as a seed/source and all other 31 ROIs as targets in the ROI-to-ROI analysis tab itself. However, although this seems to be plotting the effect sizes, the x-axis has only one tick mark labelled "all subjects" and there is no way of telling which ROI the individual bars refer to as there is no legend either (I tried to upload a screenshot but after I select my file and click update and then close window, nothing happens). Is there a fix for that? I could not find any options that would allow me to plot label the individual bars.[/quote]
↧

Calculating the adjusted connectivity matrix

June 11, 2020, 2:15 pm
$
0
0
Hi all,

I'm making a ROI-ROI analysis using the contrast [1 0 0 0 0] (SUBJECTS, AGE, MOTION, CARDIO, MEAN_CONNECTIVITY). The database has 535 participants with the age ranging 18-88.

The ROI.mat file shows the statistical results [F, dof, h, p, h] according to the contrast, but the connectivity matrix values "y column" is not adjusted by the covariates. I would to calculate the connectivity matrix in which I can see the adjusted connectivity value for each participant. For doing this, I'm considering the subjects-by-connections matrix (Z) of connectivity values, and a subjects-by-one vector of each covariate (X), and I'm computing:

X = [x1, x2, x3, x4, x5]; % design matrix (each column is a covariate)
B = X\Z; % regressor coefficients
B(1,:) = 0;
Zc = Z - X*B; % Z values adjusted at all covariates

(Script Adapted form the original script: https://www.nitrc.org/forum/message.php?msg_id=20577)

When I do that, I get the adjusted matrix, however:

1) The effect of age disappeared from all the connections. For example, If I plot the connectivity between ROI1 x ROI2 against age, the correlation effect is zero. But if I ask for the effect of age [0 1 0 0 0] on the same connection, the results explorer window of Conn shows a positive effect of age for that ROI-ROI connection. Why is there this difference?

2) The script seems to be working well, but I'm not 100% sure of that. The adjusted matrix shows complex numbers for some values, which make the average connectivity of a given pair of ROI slightly different from the column "h" of the ROI.mat (which stores the average correlation for a given pair of ROI). I used the conn_glm script for checking the stats for the adjusted script. The results is almost the same as the result shown in the ROI.mat (Since the complex numbers are making the average/variance slightly, there is a difference too).

Am I doing the right thing here? What should I for getting the subjects-by-connections matrix (Z) of connectivity values adjusted by the covariates and checking its stats?

Thank you very much for your attention.
↧
↧

Help Clarifying What QA_MaxMotion Values are CONN

June 11, 2020, 2:53 pm
$
0
0
Hello,
I would appreciate getting insight on the QA_MaxMotion co-variate in CONN.
I used a Max Motion threshold value of 3 to exclude high motion participants in my study.
I inferred that this was a 3mm max motion cut off. Was I wrong in this?
A reviewer asked about rotation threshold and I realize am not sure how max motion is evaluating both rotation and translation, nor certain on it's units for the conn project.
Best,
Victoria
↧

RE: Output space: Anatomical scan

June 11, 2020, 5:36 pm
$
0
0
[color=#000000]Hi Panos,[/color]

[color=#000000]I would suggest to modify the "preprocessing pipeline for surface-based analyses (in subject space)" to:[/color]

[color=#000000]1) remove the three steps from "functional resampling..." to "functional smoothing..."[/color]
[color=#000000]and 2) add a "functional smoothing (spatial convolution...)" step and place it third to last[/color]

[color=#000000]The resulting pipeline should look like the attached image, and that would leave your functional data in the same space as your anatomical scans (naturally, these will be different for each subject so that will preclude any group-level voxel-based analyses), skipping those steps in the original pipeline involved in projecting your functional data to FreeSurfer cortical surface.  [/color]

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

I'm trying to analyze my rsfMRIs, and would like to have all outputs in the same space and resolution as my anatomical scans.
Specifically, the steps I'm following are:

1. Load the anatomical scans under Setup: Structural and the rs functional scans under Setup: Functional,
2. Go to Setup.Options and click the "Create confound-corrected time-series" as well as change the Analysis Space to "Volume: Same as structurals,"
3. Click on Preprocessing and select: "preprocessing pipeline for surface-based analyses (in subject space),"
4. Set the "Structurals target resolution" to the voxel resolution of my anatomical scans, and
5. When prompted to enter a number of diffusion steps for smoothing, I select the default value of 40.

I was wondering whether the above process is the best way to preprocess/denoise the rsfMRIs so that all outputs are in structural space and resolution. I'm mainly asking because I wasn't sure whether the surface-based approach was the best in my case since I'm not loading Freesurfer segmentations.

Thank you in advance,
Panos[/quote]
↧

RE: ROI.mat - Difference between "y" average and "h:

June 11, 2020, 5:40 pm
$
0
0
[color=#000000] Hi Tiago,[/color]

[color=#000000]The "results.h" values in those ROI.mat files represent the contrast values (between-subjects contrast linear combination of regression coefficients from GLM analysis of the data) while the "results.y" values represent the original/input data to the same GLM (e.g. Fisher-transformed correlation coefficients for each subject). Only in cases where your second-level analyses implement a one-sample t-test (e.g. selecting only the 'AllSubjects' effect), the results.h values will be equal to the average of the results.y values across subjects.[/color]

Hope this clarifies
Alfonso
[i]Originally posted by Tiago Guardia:[/i][quote]Hi all,

As far as I understand, for the ROI.mat file, the y column has the correlation values between ROIs, and the h column has the average of the y values for the particular pair of ROIs. However if I average the y column for a particular pair of ROIs, I'm finding a different value than shown in the h column.

Does anybody know why is that happening?

Thanks[/quote]
↧
Search

Nuisance regression with bandpass filter problem fMRI

June 12, 2020, 2:28 am
$
0
0
Hi CONN experts,

I normally do a regression in my fMRI data to mitigate the effect of nuisance variables/regressors (e.g. movement, physiological artifacts, white matter, csf, etc.). Sometimes I am also interested to do a band pass filter.
I have read that the correct way to do a "Nuisance regression" with a band pass filter is including the band pass in the regression model. So, all the operations can be done in a single step.

The real problem appears when I try to band pass the data at "0.002 - 0.01"  and "0.012 - 0.028".
It seems that band passing at those frequency ranges use at least more than 90% of the degrees of freedom of the data set. Then, if I have more regressors (movement parameters, physiological recordings, etc.) it is not possible or does not make sense to do the regression because the number of regressors exceeds the number of time points.

How could I deal with this problem if I am interested in keeping bandpass in the regression model?

Best regards,
Karel
↧
↧

Preprocessing multiband resting state scan with fast TR

June 12, 2020, 5:11 am
$
0
0
Hi there,

I am a relatively recent CONN user and neuroimaging enthusiast.

I have been tasked to analyse some resting state data acquired over 10 minutes (500 volume, 1.5 TR, acquired multiband and interleave). I am trying to look for an optimal preprocessing routine in CONN. Is the default MNI_fastacquisition.mat the most suitable in this case? Is there anything that I need to alter from the default setup? I am not sure if I have provided enough information but hopefully this is a start. 

Also, there seems to be a debate over whether or not to include slice timing correction, in spite of Sladsky et al 2011 paper, I have been advised to go without STC. Another colleagues advised to do preprocessing with and without (and look at the results). It seems to me the looking at using the preprocessed data to judged on which approach to take is a bit circular. If this is unavoidable, how do I measure objectively what is "good", and if it is avoidable, how do I go about making the decision instead?

Thanks everyone!

Best,
Steve
↧

Recommended Confounds from fMRIprep for denoising in CONN toolbox (resting-state data)

June 12, 2020, 7:51 am
$
0
0
Hello experts,

I am working with resting-state data from ~200 subjects.

After preprocessing it in fMRIprep, I want to apply smoothing, then denoising in CONN toolbox.

What are the recommended confounds to include from fMRI-prep for the denoising step beyond the 6 motion parameters and their derivates, powers, etc.?

I found several recommendations for adding a_Com_corr_xx – but I am wondering how much agreement there is about including them.
Would a specific cut-off based on the variance explained by each of them be useful when selecting the number of a_Com_corr_xx to include? Or is it more beneficial to go with the full set?

I expect that the number of anatomical and temporal Comp_corr will be different for each subject so keeping a similar portion of explained variance, seem like a solution?

Are there any benefits of adding t_Como_corr_xx parameters as well?

Are there other parameters important to add when dealing with resting-state data?

Thank you in advance!!!
Natasza
↧

RE: Error in seed to voxel analysis: Source not found in global source list.

June 12, 2020, 12:47 pm
$
0
0
I'm having the same problem, did you ever find a solution???
↧

FC for task-fMRI

June 14, 2020, 4:55 am
$
0
0
Dear experts,

I'm trying to use conn for ROI-to-ROI analysis with block-design task. It seems to be two approaches for it. 

Approch1: As described in 'Task-residual Functional Connectivity of Language and Attention Networks' and 'A METHOD FOR USING BLOCKED AND EVENT-RELATED FMRI DATA TO STUDY "RESTING STATE" FUNCTIONAL CONNECTIVITY', by denoising the main effect of task ,set a new condition [0, inf] and set the filter to [0.008,0.09], I can conduct a 'task-residual FC' which emulates resting state fMRI FC.
Approch2: Set up conditions [i]task[/i] and [i]rest, [/i]remove the main effect of task and use a high pass filter[0.008, inf]. Then use contrast '[i]task>rest[/i]' to find the task-related connectivities.

I'm wondering what's the difference between these two approches? Besides, if I choose the later one, will it be necessary to use '[i]task>rest[/i]' in 2nd-level analysis? Or simply select condition [i]task[/i] will be enough?


Thanks!

Best,
Wang
↧
↧

Error with subgroup covariates and "NaN"

June 14, 2020, 10:29 am
$
0
0
Dear Alfonso
Dear Conn-Users

When testing this combination in the seed-to-voxel analysis:
-males (coded with "1", otherwise "0")
-females (coded with "1", otherwise "0")
-clinical_score_males (values for males, otherwise "NaN")
-clinical_score_females (values for males, otherwise "NaN")

using [0 0 1 -1] (for a gender by score interaction), this error below appears when trying to run the 2nd level results. This is a subgroup analysis and groups don't overlap. Is it because of the Conn version not accepting "NaN" (15.h), wrong contrast or still something else? I would appreciate any comment.
Thank you. Lucas


ERROR DESCRIPTION:
Error using conn_process (line 3555)
Null design matrix (all subjects have missing data or all selected covariates have zero or NaN values). Please check your second-level covariates
Error in conn_process (line 45)
case 'results_voxel', [varargout{1:nargout}]=conn_process(16,varargin{:});
Error in conn (line 5558)
conn_process('results_voxel','readsingle','seed-to-voxel');
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.15.h
SPM12 + DEM FieldMap MEEGtools
Matlab v.2016b
↧

plot in CONN

June 14, 2020, 1:28 pm
$
0
0
Dear CONN community:

I did SBC analysis with CONN and got the slice view plot and surface plot. Now I got another analysis result from different software for the same dataset, and I am wondering whether I can still use CONN to make the slice and surface plot so that I can compare these plots easily. Attached is result from different software. Thanks for any help and suggestion.
Forrest
↧
Search

RE: Sliding window analyses - issues at denoising step

June 15, 2020, 1:29 am
$
0
0
[color=#000000]Just bumping this up in case the thread got missed.[/color]
[color=#000000]Thank you[/color]
↧

RE: Moderation Analysis

June 15, 2020, 5:51 am
$
0
0
Hello,

To be more specific about this question following the contrasts I ran below to probe a significant moderation in seed-to-voxel analysis, what does it mean theoretically for example to find significant positive functional connectivity in seed-to-voxel analysis at High levels of BehavioralMeasureA, but no significant functional connectivity at Low levels of BehavioralMeasureA or at High and Low levels of BehavioralMeasureB? I am trying to ensure that I am interpreting seed-to-voxel moderation probes appropriately since the dependent variable is significant functional connectivity in different regions and not necessarily a number value (i.e., making it difficult to plot the moderation probes). Is it appropriate to say that the moderation between BehavioralMeasureA and BehavioralMeasureB is contingent on high levels of BehavioralMeasureA? Is there a way to display this graphically?

Thanks,

Marissa
↧
↧

No Significant Results

June 15, 2020, 11:06 am
$
0
0
My apologies if this questions is basic, I've tried searching through the manual and online tutorials and cannot find an answer.

I am using CONN 19.b. My ROI to ROI analyses when displayed in the results analysis window show NAN. When looking at the results explorer window, it says No significant results. I need the p values to report in my dissertation and I can't seem to figure out how to get the exact values.
↧

Testing an interaction: split or not & issues with "0" coding

June 15, 2020, 1:16 pm
$
0
0
Dear Alfonso
Dear Conn-Users

Previous posts suggest the following 2nd level contrast [0 0 1 -1] for testing the gender by score interaction, using "males" , "females", "score_for_males", "score_for_females", respectively. However, how do you test the same interaction, by simultaneously correcting for covariates of no interest, such as age or sites? Doesn't it confuse the Conn with categorical or continuous variables when you put "0" for these e.g. [0 0 0 0 0 1 -1]? I also assume here the scores are coded with either a value or a "0". Does it additionally require spliting each covariate into "age_males", "age_females", "site_1_male", "site_1_female", "site_2_male", "site_2_female"?

I screened the posts for this specific question, but could find no answer and would appreciate any comment. Thank you.

Greetings,
Lucas
↧

Excluding Subjects in Second-Level Covariates

June 15, 2020, 7:05 pm
$
0
0
Hello, 

I saw that this was a thread a couple of years ago and wanted to see if the protocol has changed. I would like to exclude two participants from a second-level analysis after they had been preprocessed. How exactly should I go about doing so? 

I tried entering 'NaN' in place of a number, but CONN could not recognize the string when I did this.

Thank you.
↧
Viewing all 6875 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>