Quantcast
Channel: NITRC CONN : functional connectivity toolbox Forum: help
Viewing all 6869 articles
Browse latest View live
↧

RE: display error

June 2, 2016, 10:49 am
$
0
0
Hi thank you very much for this explanations.
 
My data are from monkeys, so I have already made the preprocessing on monkey template.
The coregistration/normalization step is ok, maybe I have to make few adjustments to use the conn toolbox for monkeys ?
 
 
Thank you, best regards,
 
Maeva
 
↧
Search

RE: Reference Slice?

June 2, 2016, 12:16 pm
↧
↧

White Matter Functional Activation

June 2, 2016, 2:18 pm
$
0
0
Hi Alfonso,

I have been using CONN to explore the connectivity differences between athletes and sedentary controls. I tried using a voxel-to-seed approach, with the medial Prefrontal node of the DMN as a seed and found a statistically significant region of difference in what appears to be the genu of the corpus callosum, where athletes had greater correlation with the seed. I made sure to check all the preprocessing outputs to ensure there wasn't any tissue misclassification or registration errors. I also overlayed my seed region onto the GM segments to make sure they were in the correct position and within the gray matter. I also did not change any of the denoising options. At this point, I don't think it is an artifact/error.

A relatively recent review by Gawryluk et al(2014) (Does functional MRI detect activation in white matter? A review of emerging evidence, issues, and future directions) mentioned a few task dependent fmRI studies where the activation was localized to the different sections of the corpus callosum. There has been some evidence of white matter BOLD activation in regions that were structurally connected to cortical areas with significant differences in task based fMRI, but not information is available for resting state fMRI and white matter activation.

I am curious to hear you thoughts on the validity of white matter fMRI activation activation for task based in general and resting state fMRI in particular.

Best,
Prad
↧

RE: strange scatter plot - rank transforming ?

June 2, 2016, 11:46 pm
$
0
0
It definitely helps!

Thanks a lot!

A.
↧

RE: Clarifying contrast interpretation

June 3, 2016, 11:09 am
$
0
0
Thank you very much for the explanation, Alfonso!

I have one more (related) question. If I ran a between-subjects (patients, controls) and between-sources contrast (MPFC, RLP) and again found a significant cluster in occipital cortex, the interpretation is that the difference in MPFC-Occ cortex and RLP-Occ cortex connectivity is different in patients and controls? In this analysis, what values would be represented in the 'import values' function? I am wondering how to further explore these contrast results (e.g., see if the difference is driven by positive or negative correlations in which group).

Thank you again!
Tessa
↧
↧

Problems with preview analysis results on GUI

June 4, 2016, 12:38 pm
$
0
0
Hi all,

I've run my conn batch script up to the 1st level analysis. When I look at the results, for both the denosing and 1st-Level analyses steps the brain template and the processing results in the preview are off (not aligned onto each other, please see file attached). The labeling of the brain regions seems to follow the processing results. When I run ROI-to-ROI analyses, the problem seems to be solved and the preview at the 2nd level shows me the ROIs where they are supposed to be in the brain template.

I'm using Conn 16.a, but this has occured also with previous versions (e.g., 15). Moreover, some months ago I ran similar analyses with another data set and had no such problem. I was wondering what this may depend on.

Any help would be appreciated.
Thank you!

Best,
Gabriele
↧

RE: Setup

June 5, 2016, 9:11 am
$
0
0
Alfonso,

My apologies for not responding back sooner. I was side tracked on some other projects. With the summer here I am trying to get back to this.  I made the suggestion that you made but it didn't remove the warning. The warnings were  created during the Step 5/7 Importing ROI data. I am creating a new project from scratch. I am using data that has been preprocessed through the CONN toolbox as individual projects. I wanted to merge the projects but they were not processed identically due to human error.   

Can you explain to me what the error messages mean?

When I posted the original question I was having problems moving to the First Level Analysis. I was told that Conditions weren't loading. I was trying to track down the problem and started with the Warning Messages. I am not longer encountering problems with the first level analysis so I will just move one. I guess the take away messsage is if you wait long enough problems will resolve themselves. 

Thanks for your help,

Bob
↧

RE: Batch prompt sliceorder

June 6, 2016, 12:31 am
$
0
0
[color=#000000]Hi Alfonso,[/color]

[color=#000000]I figured that it would be something minor. It works without any problems now, so thanks a lot for your quick and very helpful response![/color]

[color=#000000]Clemens[/color]

[i]Originally posted by Alfonso Nieto-Castanon:[/i][quote][color=#000000]Hi Clemens,[/color]

[color=#000000]If you simply change that line to [/color]

[color=#000000]   batch.Setup.preprocessing.sliceorder='interleaved (Siemens)'[/color]

[color=#000000](ie. remove the curly brackets) I believe that should work without a problem.[/color]

Hope this helps
Alfonso

[i]Originally posted by Clemens Mielacher:[/i][quote]Hello everybody!

I'm using a conn batch for my resting-state analyses which is working great except for one small annoyance. Everytime the script tries to perform slice time correction a prompt pops up asking me to specify the slice order. I use the default MNI preprocessing pipeline and the "batch.Setup.preprocessing.sliceorder={'interleaved (Siemens)'}" line but the prompt keeps appearing. I read that this might be an issue of an older conn version, however, I had this problem with both v15h as well as with v16a.

Any help would be very much appreciated.

Clemens[/quote][/quote]
↧
Search

Adding covariates using does not work

June 6, 2016, 4:38 am
$
0
0
I am using to most recent version of Conn, and I am using batch scripts for both preprocessing and setup. I have trouble adding any first or second level covariates to my design: the scripts work fine and produce no errors, but when I open the project using the Conn gui, none of the covariates appear in the design and I have to add them manually using the gui. 

I have preprocessed the data using the default_mni-pipeline.

My first-level covariates are the ART-outliers and the rp_*.txt-file produced by SPM, and I add the covariates as follows:

%% 
Setup.covariates.names={'motion', 'outliers'};
Setup.covariates.files{1}{s}{2}=sprintf('%s/datafiles/nback_run1/rp_EPI_64_p2.txt', main_folder);
Setup.covariates.files{1}{s}{2}=sprintf('%s/datafiles/nback_run1/rp_EPI_64_p2.txt', main_folder);
Setup.covariates.files{2}{s}{1}=sprintf('%s/datafiles//nback_run1_4dnii/art_regression_outliers_wauEPI_64_p2.mat', main_folder,);
Setup.covariates.files{2}{s}{2}=sprintf('%s/datafiles//nback_run2_4dnii/art_regression_outliers_wauEPI_64_p2.mat', main_folder);
%% where 's'=subject number

Is the problem that I am using the SPM.mat file using BATCH.Setup.spmfiles to determine the onsets of my conditions? Someone wrote on the forum that: "Note that if BATCH.Setup.spmfiles has been specified this field will default to a single covariate representing the estimated subject motion for each subject/session." So if I use the SPM-mat-file, I have no other choice but to add the ART-outliers manually as covariates after running the setup?

My second level-covariates represent between-subjects variables such as age. They also do not appear in the gui once I load the project. I add them as follows:
%%
BATCH.Setup.subjects.effects_names={'Gender','Age'};
nsub=length(subjects);
BATCH.Setup.subjects.effects{1}=gender;
BATCH.Setup.subjects.effects{2}=age;
%% where gender and age are n*1 arrays of doubles, such as 

[21
23
19
19
20
22
23
31
32
20]

Thank you in advance for your help!!
↧
↧

Merge function creates different results

June 7, 2016, 11:23 am
$
0
0
Hi Alfonso,

I merged 9 subjects with 3 new ones (both have been preprocessed and analysed using the same script) to create
a new merged project and compared that to running all 12 subject from the ground up with the same script.

I compared the 2nd level ROI-ROI results as well as graph theory results between those two theoretically identical 
data sets and found small differences. Some connections between ROI-to-ROIs were not significant anymore in the merge 
and slight deviations in graph theory measurements like degree and global efficiency could be observed.

Is there any reason merged projects show different results than just combining them manually and running them from the start
as one project together (all the way through the pipeline)?

Thank you for your help,
Philip Brandner

PS: Running CONN v15h and SPM12 with a LFO (resting state) dataset.
↧

RE: Preprocessing failure, problem writing header

June 7, 2016, 1:06 pm
$
0
0
Update: For anyone who experiences this error, unbeknownst to me the computer I was using runs a backup program (Symantec DLO Desktop Agent) that automatically backs up files after any changes, and this conflicted with the SPM preprocessing functions. Once I changed to manual backups the analysis ran smoothly. Definitely a bit of a rookie mistake but I really wanted to make sure no one else found themselves banging their heads against the wall with this error..

All the best,
Jeremy
↧

Variable number of volumes between subjects

June 7, 2016, 1:38 pm
$
0
0
Hi there,

I have a dataset where different subjects have a different number of T2 volumes. This is an intra-subject study, so I compare two conditions where both conditions have the same number of volumes. But the number of volumes is different between subjects.

I could reduce all subjects to the same number of volumes to the minimum count of the "smallest" subject, but the minimum count is quite low (like 150 volumes), spanning only 5 minutes or so, whereas some of my subjects have 1000 volumes.

I would like to keep a maximum of volumes so that I don't miss some effects because of a too short timespan. So I processed all the subjects in CONN while keeping a variable amount of volumes per subject (from 150 to 450, because I arbitrarily set the maximum time gap to x3). This all worked and CONN shows results correctly (Seed-To-Voxel and ROI-to-ROI, I didn't try the other types of analyses).

I wonder if this analysis is technically correct, how CONN does work with a variable number T2 volumes per subject, and how can I interpret it?
I found a paper that is somewhat related that studied the effect of brain scan length on the reliability of resting state results, but not on variable volumes per subject: "The effect of scan length on the reliability of resting-state fMRI connectivity estimates", Rasmus M. Birn et al, 2013, NeuroImage.
↧

RE: Seeing an error while importing SPM.mat

June 8, 2016, 6:27 am
↧
↧

conn gui error in 2nd level results display

June 8, 2016, 7:11 am
$
0
0
Dear Alfonso,

I encountered an error by accident. I rerun a finished voxel-to-voxel analysis by optimizing some parameters but accidentally moved the /results/secondlevel/VoxeltoVoxel/ folder to another place. Then I got:

1. the warning window-'Data not prepared for analyses: Condition (rest ) have not been processed yet...';

2. error message:
'ERROR DESCRIPTION:
Attempted to access CONN_h.menus.m_results.shownconditions(0); index must be a positive integer or logical.
Error in conn (line 5288)
nconditions=CONN_h.menus.m_results.shownconditions(nconditions);
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});

CONN v.16.a
SPM12b + DEM FieldMap MEEGtools suit vbm8
Matlab v.2013a
storage: 150.5Gb available'

3. in the conn gui second level results, only the 'rest' condition of the 'Conditions' column was gone and all contents in the 'Subject effects' and Voxel-to-Voxel Measures are listed correctly.

This persists after moving back the misplaced folders and even I rerun the first level analysis and overwrite all subjects' data.

My guess is that there might exist a bug in the conn gui menu during showing second level results? Any fix? Or if it is not a bug, any idea why and how to avoid it? Thank you in advance!

Best regards,

Yong
↧

conn sec level interaction among Posi and Neg

June 8, 2016, 7:33 am
$
0
0
Dear all,


I am wondering how the Conn toolbox computes the second level analysis (ROI-to-ROI) when the correlation of a pair of ROIs have distinct signal for each group, e.g.: CTR = 0.2 and PAT = -0.3. 
Obviously this would be presented as a significant difference among groups, but how the toolbox will interpretes this relation on the results explorer? As CTR > PAT, PAT > CTR?

I also would like to congratulate the excellent and intuitive toolbox.
Thank you very much!
↧
Search

ERROR first level analyses

June 8, 2016, 10:52 am
$
0
0
Hello all,
I am trying to run a seed-to-voxel first level analysis but I am getting the follow error message:
"Condition (Rest) have not been processed yet. Please re-run previous step( First-level analysis)". 
The condition Rest was added in my setup so I don't know why I am getting this message. The error details are:
"ERROR DESCRIPTION:

Attempted to access CONN_h.menus.m_results.shownconditions(1); index out of bounds because numel(CONN_h.menus.m_results.shownconditions)=0.
Error in ==> conn at 4931
nconditions=CONN_h.menus.m_results.shownconditions(nconditions);
CONN v.15.h
SPM12 + DEM FieldMap MEEGtools conn
Matlab v.2011a
storage: 1688.4Gb available"
I also got another one when trying to re-run the analysis:

ERROR DESCRIPTION:

Reference to non-existent field 'Y'.
Error in ==> conn at 4885
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
Error in ==> conn at 4438
if ~conn_projectmanager('ispending'), conn gui_results; end
Error in ==> conn_menumanager at 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 conn
Matlab v.2011a
storage: 1687.3Gb available
I will be glad if someone could help me with this.
Thank you in advance, Kátia
↧

RE: Denoised Data Output Option Problem

June 9, 2016, 11:35 am
↧
↧

RE: Design matrix illustration

June 9, 2016, 12:34 pm
$
0
0
[color=#000000]Hi Athena,[/color]

[color=#000000]Thanks for reporting this, I believe this is a bug in how the between-sources contrast is being generated, could you please try the attached patch and let me know if that looks ok now? (this patch is for release 16a, simply copy the attached file to the conn distribution folder overwriting the file with the same name there)[/color]

[color=#000000]Hope this helps[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Athena Demertzi:[/i][quote]Hi again,

could you please help me better understand this issue?

Why do I get the same connectivity values in a 2x2 within-subject design when I model eye(2) for between-subjects contrast and eye(2) for between-conditions contrast? I mean, after estimating the F contrasts, I call REX in order to extract mean ROI values (my own predefined ROI). When I plot the contrast estimates an the 90%CI, I get exactly the same values for the pre/post condition in group 1 and the pre/post condition in Group 2. 

What does this mean? I would expect different values across groups and conditions. Can it be related to the fact that I use the main effect of two seeds? Indeed, when I test connectivity with one seed I get different values across conditions and groups.

Thank you in advance for your help, 

Athena[/quote]
↧

RE: CONN doesn't work, displays following error

June 9, 2016, 12:41 pm
$
0
0
[color=#000000]Hi Vinay,[/color]

[color=#000000]This looks perhaps similar to the issue described in this thread (https://www.nitrc.org/forum/message.php?msg_id=17345 )? Basically if you are running CONN on a Windows or Linux machine, but the CONN installation has been copied from an original Mac installation (instead of from a clean install on the new machines), you may end up with a number of Mac-specific files with names starting with ._ which create all sorts of issues (and if so, simply deleting these ._* files should fix this problem). Let me know if that seems to fit your case.[/color]

[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by vinay gupta:[/i][quote]hey, when I wanna anlyze my data via conn; it continoulsy displays following error message :

Could anyone help me with that?

ERROR DESCRIPTION:

Error using svd
Input to SVD must not contain NaN or Inf.
Error in encode_qform0 (line 25)
[U,S,V] = svd(R);
Error in mayo2nifti1 (line 63)
hdr = encode_qform0(mat,hdr);
Error in nifti (line 30)
vol.hdr = mayo2nifti1(vol.hdr);
Error in spm_vol_nifti (line 19)
N = nifti(P);
Error in spm_vol_spm_vol_hdr (line 128)
V = spm_vol_nifti(p,n);
Error in spm_vol (line 61)
v = spm_vol_hdr(deblank(P(i,:)));
Error in conn_getinfo (line 18)
V=spm_vol(filename);
Error in conn (line 606)
[V,str,icon,filename]=conn_getinfo(filename);
Error in conn (line 2812)
conn importrois;
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.16.a
SPM12 + DEM FieldMap MEEGtools Masking conn xjview
Matlab v.2015a[/quote]
↧

Dynamic FC factor loadings

June 9, 2016, 1:39 pm
$
0
0
Hello Alfonso,

We are in the process of preparing results from a dynamic FC analysis and would like to make sure we're correctly interpreting the between-group differences in dynamic factor loadings for FC between two ROIs. From reading the forum and CONN documentation, we think these values refer to a PPI interaction term, but also recently read a paper that seems to use CONN's dynamic FC pipeline and described the values as standard deviation in beta values across sliding windows. Can you please clarify what the dynamic factor loading values represent? Additionally, can you provide some context for what a negative dynamic factor loading value would indicate?

Thank you very much!
Amanda
↧
Viewing all 6869 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>