Dear Leonardo,
I guess that what you mean by “a different TR” is that the spins are pristine when acquiring the first volume? If so I assume that what you see in the first volume is relatively higher intensity in CSF and a little more “anatomical information” in the image compared to later b0 volumes?
Just how different that volume is to subsequent b0 volumes will depend on the TR of your sequence. My guesstimate is that if TR is ~4 sec they are sufficiently similar that you can just use it as is and not worry about it.
If you TR is smaller than that it might start to become a problem and you might have to change your processing pipeline a little. There is nothing inherent in eddy that says that you must start with a b0 volume. The only thing that is special with the first b0 is that it is used internally as the reference space for all b0 volumes. In the same way the first dwi is used as the reference space for all dwi volumes. For the user this will typically be of very little consequence _unless_ one specifies --dont_peas, in which case eddy will assume that those two reference spaces are identical (i.e. that the subject has not moved between the first b0 and the first dwi). Obviously that assumption makes more sense if the first volume is a b0 and the second a dwi, i.e. if there is a very short time between the acquisitions.
However, there is nothing to stop you from reorder the input data to eddy. Let us say for example that your first b0 volume is volume 15 in your 4D data. You can then extract volumes 15 and 16 (assuming volume 16 is a dwi) and stick them first in your data. That means that you now have a data set where the first b0 and the first dwi (original volumes 15 and 16) were acquired very close together in time. You just need to take care that you also change your bvals and bvecs files to reflect that reordering. From the perspective of eddy it doesn’t matter if you move or duplicate volumes 15 and 16. If you duplicate them you need to remove them again before running bedpost/dtifit/something.
Having said that, in the next release eddy will have the option to dynamically track the susceptibility-induced field, and that will work better if that assumption (no movement between first b0 and first dwi) is approximately true (and for various complicated reasons the re-ordering trick will not work for that). So for going forward I recommend everyone to have a few (I usually say 3) b0 volumes at the start of the data. The reason to have more than one is to make it very likely that you have at least one initial b0 volume that is not “broken” by intra-volume movement.
Jesper
> On 27 Aug 2018, at 23:33, Leonardo Tozzi <[log in to unmask]> wrote:
>
> To Whom it may concern,
>
> We are acquiring multiband diffusion data on a GE scanner, which has a set b0s interspersed throughout the acquisition. At first, we thought we had also a b0 volume at the start of the scan, but it turned out that this volume was a b0 calibration scan with different TR compared to the rest of the volumes. Since we did not want to introduce a volume with different properties in the analysis, we cut it and tried to first process our data without it. However, when we tried to do so using the diffusion processing pipeline from the Human Connectome Project, the processing failed, since it assumes the data to start with a b0 volume.
> My questions are: how is eddy in particular affected by not having a b0 at the start of the file? Is it still able to use the b0s interspersed in the acquisition to correct for motion/eddy currents or is the procedure invalid? In the HCP pipelines in particular (I was advised to post here from the HCP mailing list for my diffusion questions), does the motion correction step consider the closest b0 to each volume and if yes, could it be edited to run without a b0 as first scan?
> We can also consider not removing the calibration volume or adding an additional b0 at the start of our scan for all future participants, but we were wondering what the consequences are for the data collected so far.
> Thank you very much for your help,
>
> Yours sincerely,
>
> Leonardo Tozzi
>
> ########################################################################
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|