we also have some tools for this, available at
http://spm-toolbox.sourceforge.net. Our
tools are optimized for the extraction of selectively averaged event-related
timecourses, though they will also work for blocked designs. Because we use the
data
from Y.mad to create the timecourse, our tools have both of the disadvantages
that
matthew mentioned (need to set UFp=1, and susceptibility to rounding errors).
cheers
russ
Matthew Brett wrote:
> We are interested in extracting the time-course of activity in a group of
> > voxels around a local maxima. Ideally we wish to extract the time course
> > from all voxels active in a particular cluster. We have set the fMRI_UFp
> > variable to 1 in order to have information about all the voxels and are
> > using the spm_regions.m script version 2.8 (Karl Friston 00/11/22) for
> > extracting the average timecourse in a volume of Interest(VOI) covering
> > the entire cluster.
>
> > Is there any other procedure available for extracting average(along with
> > standard error) time course of a group of all active clusters surviving a
> > particular threshold.
>
> Hi,
>
> We have written a toolbox that does this - MarsBar - http://
> www.mrc-cbu.cam.ac.uk/Imaging/marsbar.html. It is still in alpha, but is
> fairly stable at the moment. One advantage of the toolbox is that you do not
> need to set the UFp to 1 to save all the data, and you don't suffer from the
> small rounding errors incurred by the compression that SPM uses, because it
> gets the data from the raw images. You have a lot of data there, so I can't
> guarantee it will be any quicker than SPM... Your technique would be to make
> an ROI from the SPM analysis (ROI->Build->Activation cluster), and then
> ROI->Extract data. Don't forget to download the bugfix routines in the
> MarsBar directory...
>
> Best,
>
> Matthew
|