Hi,

this usually happens when the RAS center is incorrect in your data. It is not really a bug, but we prepared a version that does not use the ras coordinates but the center of mass instead for initial alignment. What operating system are you running?

Cheers Martin

On Nov 20, 2009, at 8:24, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

yup, looks like a bug. Can you arrange to send the data to Martin?

thanks Bruce On Fri, 20 Nov 2009, Hui Zheng wrote:

...

Hi all,

We have a problem in longitudinal processing. The mri_robust_template somehow chopped off a part of the brain (attached image). Any idea how to correct this? With this "partial" template, the longitudinal processing of phase 1 and 2 data were also wrong. This is the second subject I was checking. I hope it is not a common problem. I can provide the subject's data if necessary.

Thanks, Zheng Hui

Hi Zheng,

we will send the right version of mri_robust_template to you.

About the brainmask: The brainmask in the longitudinal processes is taken from the base/template (mapped to each time point), to ensure that all time points use the same brain mask in the longitudinal runs. The base brainmask is the OR of the brainmask of all cross sectional time points. Therefore it might be a little larger than any one of them. We decided to err on that side, rather than cutting away some brain. If the cross sectional brainmasks are OK and the registration to the base is accurate, then the base brainmask will be good and also the brainmask of all longitudinal runs. So, it would be good to check the brainmask of the base/template. For the longitudinal runs that should be sufficient, the brainmasks in the cross sectional dirs do not matter much.

Best, Martin

On Mon, 2009-11-23 at 10:12 +0800, Hui Zheng wrote:

Hi Martin,

we need the version for "Linux centos4 x86_64" and "Intel iMac".

Another issue I noticed is that the brainmask.mgz from subjTPN.long.subjBase doesn't contain the edits I have done for subjTPN and it is larger than the brainmask from subjTPN. This happened to the 4 subjects I have checked. Please advise how to correct for this. Thanks.

Regards, Zheng Hui

On 11/20/09 10:20 PM, "Martin Reuter" mreuter@nmr.mgh.harvard.edu wrote:

...

Hi,

this usually happens when the RAS center is incorrect in your data. It is not really a bug, but we prepared a version that does not use the ras coordinates but the center of mass instead for initial alignment. What operating system are you running?

Cheers Martin

On Nov 20, 2009, at 8:24, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

...

yup, looks like a bug. Can you arrange to send the data to Martin?

thanks Bruce On Fri, 20 Nov 2009, Hui Zheng wrote:

...

Hi all,

We have a problem in longitudinal processing. The mri_robust_template somehow chopped off a part of the brain (attached image). Any idea how to correct this? With this "partial" template, the longitudinal processing of phase 1 and 2 data were also wrong. This is the second subject I was checking. I hope it is not a common problem. I can provide the subject's data if necessary.

Thanks, Zheng Hui