I do not know if this helps, but a third grouping (c) on the basis of another genotype that I just ran on the same participants (minus one) gave estimation errors only for two comparisons: one that had come up in (a) and one that had come up in (b).

Many thanks,

Yannis

From: Paloyelis, Yannis
Sent: 05 June 2009 6:17 PM
To: 'MCLAREN, Donald'
Subject: RE: [SPM] model estimation error: There are no significant voxels.

At this round of analyses the same subjects were grouped into:

(a) Patients and controls, which gave estimation errors in 4/28 comparisons

(b) Genotype A and genotype B, which gave estimation errors in 3 of the above and another 2 comparisons.

In a previous round (before excluding some participants), grouping (a) had given errors in 1/4 comparisons in (a) and a further one.

Everything was done in SPM5.

So, from (a) and (b), some comparisons consistently presented problems, whereas others only in (a) or (b) (despite using overall the same con images).

Thanks again J

Yannis

From: MCLAREN, Donald [mailto:[log in to unmask]]
Sent: 05 June 2009 6:07 PM
To: Paloyelis, Yannis
Subject: Re: [SPM] model estimation error: There are no significant voxels.

At the second level, all subjects go into all comparisons?

I'm wondering if there is something wrong with the con files. Was everything done in SPM5/8?

On Fri, Jun 5, 2009 at 11:56 AM, Paloyelis, Yannis <[log in to unmask]> wrote:

Hi,

I did exactly what you describe (SPM processing steps 1-3, and SPM 2^nd-level step 1).

Many thanks for helping me sort this out!

Yannis

From: MCLAREN, Donald [mailto:[log in to unmask]]
Sent: 05 June 2009 5:40 PM

To: Paloyelis, Yannis
Subject: Re: [SPM] model estimation error: There are no significant voxels.

SPM Processing:

(1) create a design matrix (once per subject)

(2) estimate the solution to the design matrix (once per subject)

(3) compute the contrasts

SPM 2nd-level:

(1) Use the con_files from the above process as inputs

Is the above what you did? OR are there multiple design matrices for each subject where you applied an explicit subject mask?

On Fri, Jun 5, 2009 at 11:29 AM, Paloyelis, Yannis <[log in to unmask]> wrote:

Dear Donald,

Thanks very much for your response. I feel though that I did not express myself clearly in my description of the problem.

At first level, I ran a number of contrasts. For all these contrasts, there is a single mask file (per subject).

At second level, I compare my groups on each of these contrasts. For only some of these comparisons, I get the error I described. When I compare patients and controls, it is e.g. comparisons 1 and 7, when I regroup my participants to genotype A and genotype B, it may be 1 and 7 plus 13.

Given that all these comparisons have the same first level masks, and estimations do work for most, would the examination of the first level masks reveal anything for the comparisons that could not be estimated?

Thanks for your help in understanding this J

Best wishes,

Yannis

From: MCLAREN, Donald [mailto:[log in to unmask]]
Sent: 05 June 2009 4:09 PM
To: Paloyelis, Yannis
Subject: Re: [SPM] model estimation error: There are no significant voxels.

Estimation does not require group difference, this error is telling you that there is no voxel with data from all of the images. Just because the estimation completed at the first level, does not mean that all subjects were estimated at the same voxel. Are you using analyze files? To check for alignment, do the following:

In imcalc, select all your mask.img from all the subjects first-level analyses.

In the equation box, enter: sum(X)>(N-.1)

where N is the number of subjects

In interpolation, select Nearest Neighbor

In Options --> Read Data as Matrix, select Yes.

Now, see if there are any non-zero voxels in the resulting file.

Hope this helps.

On Fri, Jun 5, 2009 at 9:48 AM, Paloyelis, Yannis <[log in to unmask]> wrote:

Dear SPM users,

Sorry for reposting, but I did not get any replies the first time. Perhaps I might be more lucky this time J

Running a 2-sample t-test (comparing Groups A & B) on a number of first level contrasts, I get an error message for only two or three of the contrasts during estimation (no files are generated).

If I run the test using the GUI, then the error is just a warning saying:

Error running job: Error using ==> spm_spm at 861

Please check your data: There are no significant voxels.

In file "/storage/data/paloyelisy/SPM/spm5/spm_spm.m" (v946), function "spm_spm" at line 861.

In file "/storage/data/paloyelisy/SPM/spm5/spm_config_fmri_est.m" (v832), function "run_est" at line 394.

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

Done.

If I run the test using a script, the execution of the script stops and I get a similar message:

??? Error using ==> spm_spm at 861

Please check your data: There are no significant voxels.

Error in ==> spm_config_fmri_est>run_est at 394

SPM = spm_spm(SPM);

Error in ==> spm_jobman>run_struct1 at 1587

feval(prog,val);

Error in ==> spm_jobman>run_struct1 at 1597

run_struct1(c.val{i},gui);

Error in ==> spm_jobman>run_struct1 at 1597

run_struct1(c.val{i},gui);

Error in ==> spm_jobman>run_job at 482

run_struct1(c,gui);

Error in ==> spm_jobman at 89

run_job(varargin{2});

Error in ==> T2ModelSpecificationEstimation at 166

spm_jobman('run',jobs)

Would it be possible for someone to explain to me what this is about? Some recent postings on a similar topics suggested this could be an error of images not being lined-up, but in this case estimation was fine for most of the first-level contrasts, so I am not sure how this could be the problem.

Does that simply mean that there were no significant voxels in either group and SPM does not need to compare the groups, or is it something else??

Many thanks for you help!

Best wishes,

Yannis

--
Best Regards, Donald McLaren
=====================
D.G. McLaren
University of Wisconsin - Madison
Neuroscience Training Program
Office: (608) 265-9672
Lab: (608) 256-1901 ext 12914
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (608) 265-9672 or email.