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
|