t1-linear method does nothing

[V-Krause]

I have successfully converted my ADNI data to BIDS using the adni-to-bids method and would now use the t1 linear method to apply the pet linear method later.

After a few small error fixes, I now have the problem that everything should seem to work correctly, but nothing comes of it.

The pipeline has been running for hours now, but there is neither a result nor any visible progress. I have often tried only one picture, which should take 6 minutes, but even after a few hours there was no result.

In this picture I tried it with 24 Images, but it stuck at this point.

So what can I do?

[omar-rifai]

Hello @V-Krause, can you give me an example image (subject/session ID) you failed to run individually so that I can try running it myself?

Thanks,

[V-Krause]

Of course.

I tried it with:
subject 033_S_1016
session M48

[omar-rifai]

Hello,
I can not seem to reproduce the issue. T1-linear seems to run fine on this subject-session (in about 6 minutes) with Clinica 0.4.1 (latest) Is this the version you are using?

Regards

[V-Krause]

Hello,
yes I use Clinica 0.4.1.

But I have to say that I had some problems getting ANTs to work and before that I always got the error that the n4biascorrection is not recognized. But I got this fixed after a clean reinstallation. However, I have the assumption that it might be because of that. Maybe you can give me a quick guide on what steps to take to get this installed properly? Or should I do a complete reinstallation of all?

Thank you for your fast answers!

[mdiazmel]

Hi,

But I have to say that I had some problems getting ANTs to work and before that I always got the error that the n4biascorrection is not recognized. But I got this fixed after a clean reinstallation. However, I have the assumption that it might be because of that. Maybe you can give me a quick guide on what steps to take to get this installed properly? Or should I do a complete reinstallation of all?

Yes, it seems that there is a problem with your ANTs installation. Try to uninstall it and install it again. We're currently proposing a conda recipe to facilitate ANTs installation in Conda Forge. In the meanwhile, you can try to install it using our Channel:

 conda install -c aramislab ants 

(inside the Conda environment where you installed Clinica).

Best,

[V-Krause]

@mdiazmel
The installation work fine, but now I get this error and don't know what I do wrong, because ANTs is a folder.

[mdiazmel]

@mdiazmel
The installation work fine, but now I get this error and don't know what I do wrong, because ANTs is a folder.

To set the environment variables for ANTs, you need to point it out to the place where Conda install the package, for your case it should be something like this:

ANTSPATH="/home/vinc/minconda3/env/clinicaEnv/bin"
export PATH=${ANTSPATH}:${PATH}

after doing this, you can verify if the binaries of ANTs are installed by running antsRegistration -h, from the command line, for example.

[V-Krause]

Okay, I reinstall my whole VM, because of some issues and now that worked. But now I have the problem with the n4biascorrection again.

In the previous VM I fixed it with an extra path for the n4biascorrection. Is there another solution or is that correct?

[mdiazmel]

Could you show us the content of the one of the nipype logs, please?
(Type in your terminal nypipecli crash /home/vinc/crash-xxxxxxxxxxxxxxxxxxxxxxxxxx.pklz)

[V-Krause]

Here is the first.

[omar-rifai]

I think the issue is the space character you have in your path as indicated by the path N4BiasFieldCorrectionis trying to look in. Try to rename your folder to "ADNI_REST" and it should solve your problem !

[V-Krause]

Okay now it works. Thanks to everyone for the quick and helpful answers!

[omar-rifai]

Great! Gl with your study.