Dear Romain,
it is not an artefact “coming” from eddy, but it is one that eddy fails to correct for. It is in the pipeline to augment eddy to fix.
My interpretation of the artefact is that there are eddy currents with long time constants such that the EC from one diffusion encoding extends into one or more subsequent shots. After a couple/few shots with one diffusion encoding the EC reaches a steady state and the EC will be stable. That means that at the transition between one volume (diffusion encoding) and the next there will be a couple of shots during which the old EC decay and the new EC build up to a steady state.
I suspect slice-to-vol could do a little to help, but it wouldn’t solve it. But, as I said, it is in the pipeline.
Jesper
> On 22 Mar 2018, at 08:41, valabregue <[log in to unmask]> wrote:
>
> Dear all,
>
> We notice on some diffusion acquisition protocol with multiband (on our prisma) a slice specific distortion than seems to be due to transient effect of eddy current that will be more intense for the first slices of the next acquired volume.
>
> Cf :
>
> https://github.com/CMRR-C2P/MB/issues/199#issuecomment-374605694
>
> The interpretation of the artifact as coming from eddy as been mention earlier here
>
> https://github.com/CMRR-C2P/MB/issues/68
>
>
> Do you thing those artifacts may be corrected with the slice 2 volume correction in eddy ?.
> My guess is that the answer is no because in case of slice to volume correction you model rigid deformation. (and with eddy current we have local non linear deformation)
>
> If no, is there a chance that those artifacts may be taken into account in a future version of eddy ?
>
> Many thanks
>
> Romain
>
>
> PS the artifact is visible on the sagital view and looks like if it is within volume motion. I wonder how widespread it is.
|