Hi,
How are you telling the difference between it running
slowly and it "freezing"? It will certainly take longer
to process such data, especially when setting up the
timeseries model at the beginning. The only limit on
the FSL processing is the limitation of the nifti files,
which is a limit of 32676 timepoints. So you are well
under this.
In a situation like this, memory is also extremely
important, and if you are low on memory then the
process may start swapping, which will make it a
lot slower too.
Anyway, one way of seeing how long it is taking
would be to start by analyzing just a portion of
the timepoints (say 500) and see how this gets
on with your system. Then progressively increase
this so that you have an idea of how long it will
take (e.g. try 1000, then 2000, then 4000 points).
I hope this helps.
All the best,
Mark
On 30 Aug 2010, at 22:16, Jianing Shi wrote:
> I have a question about the FEAT functionality in FSL.
>
> We are carrying out a fast acquisition of fMRI using the following
> set of parameters:
> TR duration = 50 ms
> number of TRs = 9800
> Single slice acquisition.
>
> A first-level FEAT analysis froze when I tried to analyze the fMRI
> data.
>
> I am wondering what is the maximum number of time points FSL can
> support in the first-level analysis?
>
> Thanks!
>
|