Dear Daniel, dear SPMers-
thanks for the comment. However, my problem was much simpler: The
transfer of onsets from my editor has failed. Apparently, the vi is
still the safest for UNIX / LINUX, and I would not use kedit or others
blindly.
Ciao-
andreas
Daniel Weissman wrote:
>
> Dear Andreas,
>
> It is possible to use 0.5 scan increments for the conditions. In
> my experience, model estimation sometimes collapses when MATLAB runs out
> of memory. You might try estimating with MaxMem set to 2^25 and see if
> that works (that works well for us and we have about 1GB of RAM). MaxMem
> is a parameter that can be set in the spm_spm.m file (it's near the top
> of the file after the comments). Basically, if MaxMem is set too high,
> then MATLAB runs out of memory.
>
> If you have fMRIT0 set to 1, then the first TR within SPM is
> numbered 0. In this case, if you have an event whose onset is coincident
> with the very first TR of a session, then you would enter a 0 into your
> vector of onsets for that trial type. However, this depends on how you
> did slice timing. If you chose the first slice that you acquired as
> your reference slice, then what I said above is correct (i.e., the onset
> of the first TR is numbered 0). If not, then the situation changes
> depending on what you did.
>
> Hope that helps,
>
> :> Daniel
>
>
> > Dear SPMers,
> > specifying efMRI designs - is it possible to use 0.5 scan increments for
> > the conditions? I have tried so, SPM accepted the trial onsets of half
> > scans without any problem but the result was not ok (it collapsed at
> > half of the session modelling). Do I have to start with 0.5 or 1.5 for
> > the primary vector of onset?
> > TIA-
> > andreas
> >
> >
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|