joe sala wrote:
> I guess then there isn't a way to do this more directly with
> 3dDeconvolve... It would be a very helpful option as datasets
> are getting larger and larger, and studies are now not only
> multi-run but are begining to span multiple sessions... could I
> add a feature like this to a very early Christmas Wish list?

Bob/Doug,

Thanks for the script -- it's a nice interim solution to the problem (I'd already penned a script for the problem you hit with sessions that are huge but not this stupidly huge -- making several part-brain masks, running 3dDeconvolve on each mask and reassembling).

That said, I'd like to add this option to my Christmas Wish list as well. Our data are starting to get huge -- stupidly huge -- each subject making 2G of functional data. Those pesky physicists keep making the scanner able to collect more data, feeding our addiction. 3Ddeconvolve is getting unhappy as we can no longer assume we'll a) have enough RAM to hold the whole dataset, and b) have enough hard drive space to hold a bunch of intermediate versions of the dataset. The abilitiy to have 3dDeconvolve read multiple input files would surely help with "b", and the ability to have it make multiple passes through the dataset, working on only a chunk of voxels at a time would surely help "a". Anyway -- just wanted to second the wish-list.

Craig