Ziad,

Is 3dUniformize better performed on the average anatomical than on the
original volumes? As in, could one just do 3dUniformize first, followed by
3dvolreg and then 3dmerge? Is there an advantage to intensity normalizing
the mean anatomical? (Does the better SNR result in better intensity
normalization?)

-shruti


-----Original Message-----
From: AFNI Message Board [mailto:AFNI-L@LIST.NIH.GOV]On Behalf Of
phorum@AFNI.NIMH.NIH.GOV
Sent: Thursday, September 18, 2003 12:08 PM
To: AFNI-L@LIST.NIH.GOV
Subject: Re: preparing a surface volume


This message was sent from: AFNI Message Board.
[afni.nimh.nih.gov]
----------------------------------------------------------------

Hi Mark,

It looks like you have all the steps you need but I suggest a reshuffling
of the steps.

1- 3dVolreg to align all volumes
2- 3dmerge to average them
From this point on you just use the averaged scan
3- 3dUniformize to remove instensity non-uniformity
4- 3dIntracranial to remove skull
5- 3dResample to write out data set in LPI and 1x1x1mm3
(see NOTE under section 'Creating the spec files for SureFit:' on SUMA's
documentation page: [afni.nimh.nih.gov])
6- 3dAFNItoMINC and load it to SureFit

This sequence will reduce the number of operations needed and preserves the
data in its original format for volume registration and averaging.

cheers,
-z

----------------------------------------------------------------
Sent using Phorum software version 3.4.1 [phorum.org]

------------------------------------------
Posted to Phorum via PhorumMail