search

bids-standard / bids-bep016

BEP016: diffusion derivatives
Creative Commons Attribution 4.0 International
6 stars 7 forks source link

DiffusionScheme field #86

Closed neurolabusc closed 1 year ago

neurolabusc commented 1 year ago

I would like to advocate that BIDS define the DiffusionScheme field from dcm2niix, with the possible values Bipolar and Monopolar. Philips and GE sequences typically use a monopolar Stejskal-Tanner sequence. Siemens user can often select between monopolar or bipolar twice-refocused spin echo sequences. Monopolar sequences enable a shorter echo time yielding better signal to noise, while bipolar sequences reduce spatial distortion in the raw data (though tools like eddy/TOPUP can attenuate this distortion).

Lestropie commented 1 year ago

I think this is outside the scope of BEP016. What I'm not sure about is whether it will be considered within scope of the main specification, or deferred to BEP033 (docuument; preprint; not sure that it was ever published fully).

Personally I would like to see it there as well as I already use it in my own processing: if data were acquired using monopolar, particularly if using a high b-value, then the interleaved b=0's can have residual eddy current distortions, and therefore shouldn't be used for susceptibility field estimation if the corresponding algorithm assumes only rigid body motion between volumes.

I would suggest posting on the main specification repo; please tag me there also so it comes up in my inbox.