 |
 |
---------- Forwarded message ---------- From: JISCMAIL LISTSERV Server (16.0) <[log in to unmask]> Date: 10 January 2012 01:06 Subject: Rejected posting to [log in to unmask] To: [log in to unmask] Your message cannot be distributed to the FSL list because it exceeds the maximum message size of 150k. The size of your message was 306,866 bytes. This limit has been set by the list owner and does not necessarily apply to the other lists hosted at JISCMAIL.AC.UK. If you have any questions, please contact the list owner at [log in to unmask] ---------- Forwarded message ---------- From: Colin Reveley <[log in to unmask]> To: "FSL - FMRIB's Software Library" <[log in to unmask]> Cc: Date: Tue, 10 Jan 2012 01:01:19 +0000 Subject: Re: [FSL] reduce volume size/mem use for probtrackX Saad - thanks. Sure, it is unusual data I guess. it is an ex-vivo macaque in a brukker 7T preclinical scanner. there is no transformation from seed to diffusion space. There is an magnetization transfer ratio structural image that is in perfect register with diffusion space, but the FA or mean_F* samples work fine for defining seeds. It's all one space. the voxels are 0.25mm isotropic. the volumes are 256x296x256. Which is all just great, but it is a lot of data. Any advice you can give in general for FSL 4.1.9 to leverage the spatial resolution of this data would be appreciated. The Bval of 4000 and high SNR permitted excellent bedpostx results for 2 fibres. what is a good integration step? (I'm using 0.125) and so on. and in addition, any way to cut memory and computation time. With several target masks and a small (in terms of space, not voxel numbers) seed mask it can take a day, and actually just fails with "Indexing error" or similar. Unfortunately I don't have the whole error because I was using the "screen" program and didn't save the error stream to file.