Specifically, if the present QSIprep pipeline can be used to preprocess DTI acquisitions?

QSIPrep can be run on practically any diffusion dataset, but there are some specific pipelines that take advantage of different sequences (e.g. diffusion kurtosis imaging or DKI) that are yet to be implemented. But you can still run a DKI set and preprocess it like a typical diffusion dataset. Also, DTI is not an acquisition, per se. It is a derivative of DWI. That is, you acquire a dwi dataset, and from that calculate the diffusion tensor image. Since QSIPrep’s endgoal is producing probabalistic tractography, typical DTI metrics like FA and MD aren’t produced by default since not every tracking algorithm needs those. But you can take the preprocessed DWI and fit the tensor using Dipy, FSL, or any other of your favorite DWI code bases. Does that make sense?

Best,

Steven

The only acquisitions that QSIPrep can’t preprocess right now are q-space trajectory imaging / gradient tensor imaging / sequences that samples anything more complex than points in q-space.

If you’ve got Cartesian or random sampling be sure to use `—hmc-model 3dSHORE`