Dear Linda,
Are there any error messages in the command window when you do the
estimation? In principle, SPM 'doesn't know' that the images you
supply come from MEG so any problems probably have to do with
something in the data itself. Also for what you are doing there is no
difference with 2nd level fMRI analysis so I don't see the point of a
separate tutorial.
Best,
Vladimir
On Wed, Mar 9, 2011 at 10:59 AM, Linda Burns <[log in to unmask]> wrote:
> dear vladimir
>
> i appreciate your helpful reply..so far i have been using spm5 since i am in the middle of a project..as i said i do not have great experience with spm at least not as a power user.
>
> i had a look at the code for both estimations. i still do not get the following...
> in the fmri estimation i get beta,rms,rpv,conn, ess images and most importantly the spmF images that form the contrast manager...
> using the meeg estimation it gives me only beta images so no contrast manager is formed...
> as far as i understand there the beta images are outputs of the first level analysis together with rms,rpv, ess and conn images that set the contrasts, which later give the spmF images for the contrasts..
> therefore for some reason the contrasts in the meeg option are not created..
> what does this say about my data ?
> on one hand i can only suppose that if there were no significant voxels then the contrast manager would be formed and no voxels would be displayed for each effect...
> on the other since i do not reach this level there is probably a part in the estimation procedure that treats my data differently depending on the option fmri or meeg. ..
> how can i recognize this? generally i would expect the frmi estimation to be more strict than the meeg one...
> do you think that what i am observing is a standard response and it probably says that my data for some reason cannot pass some thresholds during the estimation procedure? it would be helpful if there was a prompt that clearly stated that there no significant voxels and analysis cannot continue rather than just giving beta images...do you think that this is a standard response for meeg estimation or does it have any bug or something?nevertheless is it not very convenient when you do not get a warning or error message.. as far as i can imagine thre is something in the creation of the ess and conn images that is different between the two option but is critical, at least for me...
>
> originally i used the fmri option in order to normalize the sam images i created in the ctf space and then i continued using the same option for the 2 level estimation. i found a tutorial how to use factorial design for fmri but nothing on meeg so if followed the frmi option..it would be helpful however to have a tutorial just for meg users. my design is a 2x2 factorial.
>
> thanks
>
> linda
>
>
>
|