Correcting for Signal Drift and Gibb's Ringing Effect

[Ehsan Misaghi]

Hello everyone,

I'm trying to correct a DTI dataset for signal drift and Gibb's ringing effect. I have a few questions about the pipeline:

1. I'm doing the corrections in the following order: a) Correct for Signal Drift b) Correct for Gibb's ringing effect c) Estimate the DTI .mat file d) Motion and EPI distortion correction. Is this order the correct order?

2) My dataset includes 10 b0 images at the beginning followed by 30 b1000 images. I know for signal drift correction it is better to have b0 acquisitions in the middle of the DWI acquisition, but we haven't acquired the data like that. So, I was wondering whether I can still correct my dataset for signal drift with this acquisition method.

3) For Gibb's ringing effect correction, I keep the default values for lambda, # of iterations and step size (100, 100 and 0.01, respectively). Do I have to somehow calculate these parameters and use specific values for my data?

Thanks,

Ehsan

[Sjoerd Vos]

Hi Ehsan,

1) I'd say the order you're doing the processing steps in is perfect! Doing signal drift correction first won't affect Gibbs correction.

2) I'd really check the output figure from the drift correction step. It's not perfect to estimate drift from the first few volumes to then extrapolate this to the rest, and could actually be a bit risky. Since you have ten b0 ims to estimate on and then extrapolate to the remaining 30 I'd expect it's safe enough, but please do check there's nothing fishy going on.

3) don't have too much experience with the Gibbs ringing correction so I hope someone else can help you out there.

HTH,
Sjoerd