Hi All,

 I was going through the 2_funcpreproc.sh script and I don't seem to find any slice-timing correction. Is it the case that people don't do that? I am aware of the "controversy" about whether it's helpful or not, but I see that most publications doing this sort of analysis do include it as one of the initial steps.

 Do people using this scripts generally not use it?

 Thank you for any input

 cheers

 Martin

Hi Martin,

thanks for not taking the scripts for what they are and taking the time to go through them and seeing what they're actually doing :) 

Indeed, we did not include slice-timing correction in the 2_funcpreproc.sh script. The reason for that is that when gathering the data from the sites we did not have access to all imaging parameters. To avoid a delay in sharing the data and putting out the feasability study (Biswal et al., 2010) we decided to go ahead without implementing the slice-timing correction. And yes, there is also the controversy about whether slice time correction has any influence if you're looking at the slow-frequency fluctuations. In terms of the fcon data it would also mean applyin different slice time corrections to different datasets as not all had the same acquisition order. (Since then we added the slice timing information to the table you can find here http://www.nitrc.org/docman/?group_id=296).

However, yes, most people (including us in our papers) implement slice-timing correction. As mentioned in the wiki you are free to add slice time correction to the script, which should be fairly easy. You should do it as one of the first steps, right after removing the first n timepoints.

Let me know if you have any difficulties adding it to your scripts.
Best,
Maarten

Thank you Maarten!

 For the benefit of "onlookers", here is a small line that you can paste in the script, just under the dropping of the first TRs (CAUTION: this is for ASCENDING SLICES -- see notes):

slicetimer -i ${rest}_dr.nii.gz -o ${rest}_dr -r $TR

in case your TR variable is not set you can also use:
slicetimer -i ${rest}_dr.nii.gz -o ${rest}_dr -r `fslval ${rest}_dr pixdim4`

NOTE:
1. It will overwrite the _dr image, so if for some reason you want to keep that 'original' file, you can set the  output to ${rest}_stc.nii.gz but then remember to change the input to the next 2 steps (i.e., 'de-obliquing' and re-orienting) to ${rest}_stc.nii.gz
2. This command is for ASCENDING slices.
     If you want descending add: --down to the above command
     If you want interleaved add: --odd to the above command (I think..)

Cheers

 Martin

Hi Maarten,

I think your release table is very helpful. But not every site has the slice order information. Is it because they didn't provide it or I can find it elsewhere? In the case that the slice order information is missing, is it reasonable to skip this step? 

Thanks,
Xue

Hi Xue,

In the practice of our lab, we generally do slice timing correction if we have the information.
If the information is too hard to get, then we will skip it. we don't believe it's a necessary step in data preprocessing.

Yang

Originally posted by Xue Yang: