Quantcast
Channel: NITRC CONN : functional connectivity toolbox Forum: help
Viewing all 6872 articles
Browse latest View live
↧

RE: Extracting PPI values

November 25, 2016, 9:55 am
$
0
0
Thanks, Alfonso! I successfully exported the csv with beta values per subject.

I do have a question about the non-parametric analyses. I first ran the default 1000 iterations via the GUI. However, I'd like to see if my significant result is still there after 10,000 iterations. When I go back and select 'non-parametric' it outputs the result from the 1000 iterations. Is there a way to change the number of iterations and rerun the analysis?

Second: I have a significant result (p = 10^-6) of negativity connectivity between my seed and a cluster. When I ran the permutations it was no longer there. However, for a different subject effect but the same contrast I have another significant result (p also  10^-6) of positive connectivity between my seed and a different cluster which  is still significant after permutations. I just thought it odd that one would survive non-parametric and another wouldn't. Does it have to do with the positive vs negative connectivity? Might I have to modify the cluster threshold for the negative connectivity?

Finally: I specified 3 conditions in my general Setup: angry, neutral and control. Based on what I read in the forum, if in PPI I selected angry>neutral, does that mean angry-control > neutral - control?


Best wishes,
Erin
↧
Search

RE: One structural scan for each functional scan

November 26, 2016, 4:21 am
$
0
0
Hello Eric,

Yes you can by using the programming API conn_batch, more specifically you can set CONN_x.Setup.structurals{sid}{sessid} (so you can set one structural per session and subject).

For a concrete example, please see my conn subjects loader script:

https://github.com/lrq3000/neuro_experiments_tools/blob/master/matlab/conn_subjects_loader/conn_subjects_loader.m#L269

You will see you can easily set at the same time and with a similar syntax the functional images and the structural images per session.

Hope this can help you,

Best regards,
Stephen
↧
↧

RE: Importing structural and functional images

November 26, 2016, 4:25 am
$
0
0
[i]Originally posted by Eric Hg:[/i][quote]Dear Alfonso,

Thanks a lot for the advice!

I got the functional files as 46x1 and the structural files as 1x46. I hope that's the correct way to load the files in the CONN toolbox.

Best,

Eric[/quote]
[color=#000000]If you are using 4D functional files, the cell array should be the same size for both structural and functional images (ie, 1 per subject and session). If you are using 3D functional files (so volumes are split over multiple files), you should have nb_volumes x nb_subjects x nb_sessions (or the transpose, not sure the correct orientation but the size should be like this).[/color]
↧

RE: Constructing within subject design

November 26, 2016, 4:30 am
$
0
0
I think you can consider Drug/Placebo as groups (so as second-level covariates) because they are independent. Sessions (Conditions tab) is I think meant to be used when there is a dependency between the two set of images, so this is perfect to model pre/post sessions.

I am not sure but I think you can model pre/post in the conditions tab as sessions, and then drug/placebo in the 2nd level covariates (eg, for Drug covariate you set 1 for all subjects that used the drug, then 0 for the rest ; for the Placebo covariate you do the opposite: 1 for subjects using placebo and 0 for the ones with the drug ; in the end you should have 2 covariates that are non-overlapping vectors of 1s and 0s).
↧

RE: Structural data

November 26, 2016, 8:37 am
$
0
0
Hello Tateya,

Yes you can and I think it would be better indeed to have the real structural image for each subject. I guess the default image is SPM's template, but it would be more statistically correct during the analysis to use the structural images instead of the template.

Note that you can specify one structural image per subject on CONN GUI, but using the conn_batch.m API you can programmatically set one structural image per session (so you can have multiple structural images per subject, one for each session).

Hope this can help you,

Best regards,
Stephen
↧
↧

RE: error using load

November 27, 2016, 11:07 am
$
0
0
[color=#000000]Hi Rola,[/color]

[color=#000000]This is curious, the error message suggests that the ROI_Subject019_Condition002.mat file might be corrupted. This can happen, for example, if you reached a "disk-full" condition in a previous run or if the system was hard-reset in the middle of a write operation. I would suggest trying to delete that file and run the analyses again (using the 'do not overwrite' option in order to avoid having to repeat any already-performed steps)[/color]

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

[color=#000000] 
[/color][i]Originally posted by Rola Farah:[/i][quote]Hi all,
I keep getting the following error while running first level analysis.

ERROR DESCRIPTION:

Error using load
Unable to read MAT-file E:\MATLAB\conn_project01\results\preprocessing\ROI_Subject019_Condition002.mat. File might be corrupt.
Error in conn_process (line 1830)
X1=load(filename);
Error in conn_process (line 41)
case 'analyses_gui_seedandroi',disp(['CONN: RUNNING ANALYSIS STEP (ROI-to-ROI or seed-to-voxel analyses)']); conn_process([10,11,15],CONN_x.Analysis);
Error in conn (line 5087)
else conn_process('analyses_gui_seedandroi',CONN_x.Analysis);
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.17.a

Any idea or help is highly appreciated
Thank you[/quote]
↧

Missing sessions and switching GUI - batch

November 27, 2016, 4:18 pm
$
0
0
Hi everyone,

I really hope that someone knows the following:
<span style="white-space: pre;"> </span>1. In a setup with each subject having two session and some of the subjects are missing the follow-up (session 2) is it still possible to include them in the analysis? And will the missing session be a problem?

<span style="white-space: pre;"> </span>2. Is there an easy way to switch between GUI and batch? If I for example define: number of subjects, sessions, TR and load the structural and functional files can I switch to batch?

<span style="white-space: pre;"> </span>3. When is it best to remove the initial functional scans (like the first 3-4)? Is it before the default preprocessing pipeline?

Best, 

Eric
↧

Order of preprocessing steps

November 27, 2016, 10:49 pm
$
0
0
Hello CONN users,

I have a question about "default preprocessing pipeline" which you can choose as a preprocessing method when you run data preprocessing by CONN. I wonder why "functional Slice-timing correction" is not the very first step of its order. Isn't it supposed to be the first step of data preprocessing? I would be very happy if you could tell me the reason of that. I just would like to know why Slice-timing correction is not in the first step. Thank you.

Best regards,
Risa
↧
Search

RE: Loading conn files in SPM

November 28, 2016, 2:26 am
$
0
0
Dear Alfonso

I have a follow-up question on the question form Laurel. I would like to export beta weights (regressor coefficients) ideally as a .txt file to import them into a SPM GLM. Is there a possibility how to get these beta-weights as .txt files? Or do you have another idea how I could do this? The reason I want to do this is to investigate the effect of the denonising with aCompCorr. Therefore, as an alternative solution, do you know a toolbox which implemented aCompCorr in the same way as CONN, that would allow me to do this? 

Thank you very much in advance, 

Isabel
↧
↧

Effects of task confounds

November 28, 2016, 2:45 am
$
0
0
Dear experts,

I am using previously preprocessed and band-pass filtered data in CONN.  I don't want to do any further denoising except for removing potential confounding effects of tasks. I know that leaving confound list empty, changing band-pass filter to [0 inf] and selecting no detrending will completely skip all denoising. However if I keep the effects of task as the only items in the confounds list and still have no filtering/no detrending will this remove these effects?

Thank you,
Isabella
↧

Effects of task in pre-processed data

November 28, 2016, 2:46 am
$
0
0
Dear experts,

I am using previously preprocessed and band-pass filtered data in CONN.  I don't want to do any further denoising except for removing potential confounding effects of tasks. I know that leaving confound list empty, changing band-pass filter to [0 inf] and selecting no detrending will completely skip all denoising. However if I keep the effects of task as the only items in the confounds list and still have no filtering/no detrending will this remove these effects?

Thank you,
Isabella
↧

RE: Analysis step error, minimal denoising

November 28, 2016, 8:47 am
$
0
0
This problem has been solved. You are correct in your diagnosis, Alfonso; motion parameters were in the first-level sources list causing the error. I removed these from the first-level covariates in my batch script and all works well now.

Thank you!
Scott
↧

RE: Denoising Error, Updating Analysis variables

November 28, 2016, 10:21 am
$
0
0
Hey Alfonso,

Thank you for your reply. Attached is the conn_*.mat file as requested. Please let me know if you may need anything else. I'll work on it on my end as well.


Thanks,
William
↧
↧

Conn17a Networks

November 28, 2016, 10:34 am
$
0
0
Is there a reference for how the included networks in CONN 17a are defined?
↧

Error using spm_slice_vol : File too small

November 28, 2016, 12:30 pm
$
0
0
Hello,

I am getting the following error with Conn 17a during "loading ROIs" (with Atlas.nii and Networks.nii as default ROIs). I tried both SPM8 and SPM12 but got the same error. Any help is much appreciated.

Thanks,
Hadi 

--------------------------------------------------------------------
Error using spm_slice_vol
File too small.

Error in spm_read_vols (line 34)
Y(:,:,p,i) = spm_slice_vol(V(i),spm_matrix([0 0 p]),V(i).dim(1:2),0);
Error in conn_process (line 608)
X0=spm_read_vols(V0);
Error in conn_process (line 15)
case 'setup', disp(['CONN: RUNNING SETUP STEP']); conn_process([0:4,4.5,5]);
Error in conn (line 3195)
else conn_process('setup');
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});

CONN v.17.a
SPM8 + ArtRepair Beamforming DEM FieldMap MEEGtools
Matlab v.2014b
storage: 4813.1Gb available
--------------------------------------------------------------------
↧
Search

Dynamic FC Analysis - Error Msg in Conn17a

November 28, 2016, 12:50 pm
$
0
0
Hello,

Thank you for the great toolbox.

We are currently trying to run a dynamic FC ROI-to-ROI analysis using conn17a. We are simply trying to get a hang of the feature, thus we are doing a test run with data from only two subjects.

The initial steps of the analysis seem to work to completion with our ROIs and subject data: denoising comes to completion, then in 'first level analyses' I click 'done' to run the dynamic FC analysis, and immediately I receive the following error:


"Error using eig
Input to EIG must not contain NaN or Inf.
Error in conn_ica (line 47)
[Q,D]=eig(X*X'/(Ns-1));
Error in conn_invPPI (line 174)
[S,W]=conn_ica(reshape(B1,[Nr*Nr Nk1])',[],'dodisp',DOPLOT,'rndseed',RNDSEED);
Error in conn_process (line 3075)
[H,B,H0,B0]=conn_invPPI(Y,Ncomponents,X,Xfilter,1);
Error in conn_process (line 44)
case 'analyses_gui_dyn',disp(['CONN: RUNNING DYNAMIC CONNECTIVITY STEP']); conn_process(14,varargin{:});
Error in conn (line 5135)
else conn_process('analyses_gui_dyn',CONN_x.dynAnalysis);
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.17.a
SPM8 + Beamforming DEM DiffusionSPM8 FieldMap MEEGtools conn wfupickatlas xjview
Matlab v.2013b
storage: 31.4Gb available"

I see that similar issues have been brought up and addressed for previous versions of the toolbox, but unfortunately I was not able to find a solution that worked for me.

I look forward to your response.

Thank you,

[b]Kevin  G. Solar (BS)[/b]
MSc Candidate
Department of Physiology and Pathophysiology
University of Manitoba
Figley NeuroImaging Lab
Kornelsen Lab
Winnipeg Health Sciences Centre
↧

NaN values in Graph Theory Analysis

November 28, 2016, 3:32 pm
$
0
0
Hello Conn experts, 

I'm running a between-groups analyses on task-based functional connectivity and am interested in a particular network (the AON). In particular, I'd like to see if there are any differences in functional parameters in the AON between groups. I am pretty new to functional connectivity and have several questions!

I have attempted to set up my contrasts following this post (https://www.nitrc.org/forum/message.php?msg_id=18769) -- setting up the between-subjects effects as [1 -1] for the group comparison, selecting my condition, and then hitting "graph theory". 

It seems that the between-source contrast column in the second-level design page doesn't have an effect after I click on graph theory--is this correct?

From the graph theory page, I go to the network nodes, and select my ROIs from my network of interest (8 ROIs). However, when I click on "export data," I get many NaN values for multiple subjects in my .csv file. This does not seem correct. Can someone please explain this to me? (I do not have this problem if I select all of the atlas ROIs as nodes). I have kept the default thresholding values. 

Also, when I select only my 8 ROIs, does this mean that each node will have at most 7 edges (connecting to other nodes I selected), or does this look at how well connected the nodes I selected are, to all other nodes in the brain? 

If I am also interested in how well connected the network is (intra-network) compared to another network, in one group compared to another, what would be the best way to go about this?

Finally, is there a way to export the default ROIs so that I get individual nifti images for each one? (I found them in the conn>rois folder, but it comes with all of them in one file). 

Thanks so much for your help!
↧
↧

2nd level connectivity crash (specific FoV)

November 29, 2016, 6:50 am
$
0
0
Dear CONN community,

I am having a hard time understanding the error I obtain in CONN when arriving at the 2nd level analyses step. 
I've used the toolbox on several datasets so far, and it always worked for whole-brain field of view.

This specific set however was acquired using a non-wholebrain FOV, covering only the basal ganglia and temporal cortex.

Seeds were selected based on group-level results (SPM12) and even running a seed-to-seed analysis (hence excluding result computation for all brain voxels), but it still crashes. I attached a printscreen of the error I'm getting (using the latest version of CONN).

Would anybody have a clue here?

Thanks in advance and best wishes
Leonardo Ceravolo
University of Geneva, Switzerland
↧

Changing segmentation input in the gui

November 29, 2016, 10:09 am
$
0
0
Hello,

During preprocessing, the functional segmentation is failing because it cannot read my meanu_seg_sn.mat. I believe it is because my meanu.nii files have multiple volumes. Is there a way to change my segmentation input to the first functional volume, rather than the mean volume, using the gui?

Thank you,

Andrea
↧

Using ART in addition to Despiking?

December 1, 2016, 8:03 am
$
0
0
Hello CONN users,
I've been trying to figure out if it would be okay to use ART in addition to Despiking. I've seen many posts suggesting that ART is superior to despiking, but I'm not sure why both couldn't be used or why this would be troublesome. If ART identifies outliers and produces a file that can remove these individual timepoints, but despiking is a continuous function, why not do both? Is it because despiking following removal of the outliers cause an over "squashing" of the remaining data (sorry for being so simple in my verbiage)? Could one then use a liberal ART threshold to identify the very extreme values, while still benefiting from the "squashing" of the remaining data (or would this cause too much signal loss)? Or am I totally off base?!
Any guidance is much appreciated!
Christina
↧
Viewing all 6872 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>