Dear Christian,
Thank you so much for your reply!
Regarding my question 2.b: this is surprising, as I used the raw T1 image.
This file also went through the cross-sectional pipeline successfully.
Do you have any additional thought on that problem?
Thanks again for your time.
Yann
> On 9 Nov 2017, at 3:31 am, Christian Gaser <[log in to unmask]> wrote:
>
> Dear Yann,
>
> On 8 Nov 2017, at 9:44, Yann Quidé wrote:
>
>> Dear Christian and other CAT12 users,
>>
>> I am using CAT12 v1204, SPM v6909, Matlab 8.1 (R2013A) on Mac OSX 10.9.5
>>
>> 1) Cross sectional preprocessing
>>
>> On few subjects, I get this kind of warning:
>> "Internal WMH correction for spatial normalization (WMHCstr=0.50):
>> WARNING: Uncorrected WM lesions (5.16% of the WM)! “
>>
>> The preprocessing continues normally, and the average quality assessment indicates, for example: 86.49% (B).
>>
>> Is there anything to “fix”? Do I need to include the WMH values to any subsequent analysis? To what extend these WMH are considered fine? I attached a copy of the log for one of my subject (file: catlog_1)
>
> This warning can be ignored. CAT12 contains a very basic detection of white matter hyperintensities (WMH), that is intended to internally fill these regions to obtain a spatial registration that is less affected by these WMHs. Of course there exist more appropriate approaches based on the additional use of FLAIR or T2 images, but at least the method tries to roughly detect these WMHs for a less biased spatial registration.
>
>>
>> 2) Longitudinal preprocessing (2 sessions)
>>
>> 2.a) Related to question 1:
>> Surprisingly, while they did not show up after the cross-sectional preprocessing stream, some WMH warnings appeared for some subjects (different from those from the cross-sectional stream!) during the longitudinal preprocessing.
>> Is there anything I can do to potentially solve that?
> No.
>
>>
>> 2.b) Unrelated to question 1:
>> For 1 subject during the longitudinal stream, the preprocessing is working well for the average and Session 1 scan, but crashes for the Session 2 scan. These scans were normally processed during the cross-sectional pipeline.
>
> It looks like you are using skull-tripped images which is never a good idea and sometimes causes these problems. In this case I would strongly recommend to use the original images with skull.
>
> Best,
>
> Christian
>
>> Do you have any suggestion on this error? I attached a copy of the log (file: catlog_2).
>>
>> Thanks for your time.
>>
>> Cheers,
>>
>> Yann
|