JiscMail Logo
Email discussion lists for the UK Education and Research communities

Help for SPM Archives


SPM Archives

SPM Archives


SPM@JISCMAIL.AC.UK


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

SPM Home

SPM Home

SPM  September 2015

SPM September 2015

Options

Subscribe or Unsubscribe

Subscribe or Unsubscribe

Log In

Log In

Get Password

Get Password

Subject:

Re: Bayesian model estimation for multiple sessions

From:

"Penny, William" <[log in to unmask]>

Reply-To:

Penny, William

Date:

Mon, 28 Sep 2015 15:44:02 +0000

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (76 lines)

Dear Alex,

Yes, you are correct - the current code cannot handle multiple sessions.

Solution 1
------------

You can do Bayesian inference at the second level without doing it at the first level. So, this is probably the most computationally efficient way to go here. Just take the con images from a classical (maximum likelihood) analysis (this will accommodate multiple sessions) and enter them e.g. into a two-sample t-test/other design at the second. Estimate the second level model using classical, and then Bayesian. If you then enter what would be an F-contrast, SPM will make a map of evidence against the null. Or you can make posterior probability maps testing for effects bigger than a specified size (what would otherwise be t-tests).

Solution 2
------------

There may be a benefit of doing a Bayesian analysis at both levels, but the improvement could be marginal. And it may not warrant the very large increase in computer time if you are doing a whole brain analysis. If you're doing analysis on selected brain regions (e.g. using explicit masking) then the overhead will not be so bad. You will have to do it separately for each session and then use imcalc to average over sessions. You can then put these images into a second level analysis. I notice you are saying 'Yes' to model evidence maps. This is quite computationally demanding and unless you are doing model comparisons I would select 'no' - this will save a lot of time.

Best wishes,

Will.







________________________________________
From: SPM (Statistical Parametric Mapping) <[log in to unmask]> on behalf of Alexander Taylor <[log in to unmask]>
Sent: 23 September 2015 10:19
To: [log in to unmask]
Subject: [SPM] Bayesian model estimation for multiple sessions

Hello,

I am running an experiment with 3 runs/sessions for a task based fMRI. I would like to run bayesian model estimation at the 1st level instead of classical modelling. And then subsequently second level bayesian as well.

I am running the analysis from script and have set up the script with 3 different EPIs with appropriate timing .mat files. I have the following selection for my bayesian part of the script:

matlabbatch{2}.spm.stats.fmri_est.spmmat = {[IMAGE_DIR, 'SPM.mat']};
matlabbatch{2}.spm.stats.fmri_est.write_residuals = 0;
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.space.volume.block_type = 'Slices';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.signal = 'UGL';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.ARP = 3;
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.noise.UGL = 1;
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.LogEv = 'Yes';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.anova.first = 'Yes';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.anova.second = 'No';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(1).name = 'Condition1';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(1).convec = [0.5 0 0 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0 0 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0 0 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ];

matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(2).name =  'Condition2';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(2).convec = [1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ];

matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(3).name =  'Condition3';
matlabbatch{2}.spm.stats.fmri_est.method.Bayesian.gcon(3).convec = [0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ];

etc......

After running the script I get the following error:
----------
Warning: spm_vb_ppm_anova only works for single session data.
Failed  'Model estimation'
Improper index matrix reference.
In file "/Applications/spm12b_5918/spm_vb_models.m" (v1143), function "spm_vb_models" at line 23.
In file "/Applications/spm12b_5918/spm_vb_ppm_anova.m" (v4489), function "spm_vb_ppm_anova" at line 37.
In file "/Applications/spm12b_5918/config/spm_run_fmri_est.m" (v5809), function "spm_run_fmri_est" at line 289.

The following modules did not run:
Failed: Model estimation
----------

If I understand this correctly I cannot run this model set up using Bayesian at the 1st level, due to there being more than one session in the design matrix.

My question therefore, is there a way to run 1st level bayesian model estimation with more than one session/run simultaneously? Or is there an alternative way to achieve this, possibly by averaging the bayesian estimation 1st level analysis from 3 different sessions/runs?

Yours Gratefully
A. Taylor

Top of Message | Previous Page | Permalink

JiscMail Tools


RSS Feeds and Sharing


Advanced Options


Archives

September 2019
August 2019
July 2019
June 2019
May 2019
April 2019
March 2019
February 2019
January 2019
December 2018
November 2018
October 2018
September 2018
August 2018
July 2018
June 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
2006
2005
2004
2003
2002
2001
2000
1999
1998


JiscMail is a Jisc service.

View our service policies at https://www.jiscmail.ac.uk/policyandsecurity/ and Jisc's privacy policy at https://www.jisc.ac.uk/website/privacy-notice

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager