Question about temporal coherence

[magdamd]

I have a question regarding the temporal coherence that characterises ROMEO. Does it make sense to use ROMEO with the temporal coherence ON for a set of images that consists of a FLASH image and a EPI image?

And in that case what are the echo times that should be passed? I would guess the following times

• for the FLASH: the time from the RF pulse until half of the 1st readout
• for the EPI: the time from the RF pulse until half of the whole sequence train (when half the k-space has been acquired)

[korbinian90]

I'm assuming you mean the phasegradientcoherence weight This is a difficult question. In the ideal setting, it would make sense, but I'm no sure for real datasets: If there would be no artefacts in the flash and epi images, then the phase should be phi0 + B0 * TE. The temporal coherence (phase gradient coherence) compares the phase gradients of both time points adjusted by TE. The gradient calculation removes the phi0, so that wouldn't be a problem with EPI and FLASH. I think mainly artefacts and geometric distortions between both acquisitions would be a problem.

Yes, I would pass the effective echo time of both sequences. The phase gradients might be especially sensitive to the higher frequencies. That could mean that the outer k-space regions in EPI are more important here, but hard to say.

I think you would need to test it and see if it improves or worsens the result.

Another point: If you use temporal 4D unwrapping together on these two datasets, there might be problems with differing phi0 (phase offsets). Temporal 4D unwrapping assumes the phase proportional to B0 * TE, which means the phase offsets are removed prior. However, it doesn't immediately fail, if this is not perfectly fulfilled.

I hope this helps!