Hi,
Sorry, the new flirt command-line options aren't yet on the wiki - that will change very soon.
1) No bias correction is done within the epi_reg script. We advise that this is done beforehand.
2) I've just updated the wiki to explain that an existing segmentation is used if it has the same name as the t1brain image but ends with _wmseg. Currently there is no other way of specifying it, but we could add something more explicit in the future.
3) There would be no problem using a segmentation from any other tool (as long as it is aligned to the T1 image still).
All the best,
Mark
P.S. the extended parameters include the derivative of the motion parameters and the squares of the parameters and derivatives.
On 12 Sep 2012, at 08:05, Ted Satterthwaite <[log in to unmask]> wrote:
> Dear FMRIB,
>
> Thanks so much for the release of FSL5.0. I am in particular excited about the integration of BBR for coregistration, but don't fully understand the usage as of yet.
>
> The BBR page suggests the use of a bias-corrected structural image as well as a segmented structural image. The FLIRT wiki page also suggests this via the epi_reg script. However, on the FLIRT wiki and command line help I do not see options for where to specify these images. Specifically:
>
> 1) Should the wholehead_T1 and brain extracted T1 specified in epi_reg be bias corrected first? Or is this done as part of the script?
>
> 2) Where does one specify the T1 segmented image in the usage of epi_reg?
>
> 3) Can you anticipate any problems in using a non-FAST T1 segmentation for #2?
>
> Thanks for your advice as always.
>
> Best,
> Ted
>
>
> PS. Totally unrelated, but I didn't see a description of how the "extended motion parameters" are derived in the new FEAT v6.0wiki. Could you advise?
>
|