DSI data quality

[Mark Pinsk]

Hi Frank

I collected a sample dataset with the 101 grid scheme (bmax=7K, MB=3, 2.4mm voxels), but having some trouble.  Would you mind taking a look at the images? 

1. I cannot create a brain mask in DSI-Studio, and it will not properly load the mask that is output by QSIprep.

2. Both the original un-processed whole brain tracks and the QSIprep processed tracks look pretty bad.

The QSIPrep preprocessing included bias correction, denoising, SHOREline motion correction, susceptibility correction with opposing B0s, and ACPC rotation.

Any thoughts on the images?  Is this what you typically get?  I've uploaded a zip file for you with both original and preprocessed images plus brain mask.

thanks!
Mark

[Fang-Cheng Yeh]

Sorry about the problem.
Could you upload the SRC file?

> --
> You received this message because you are subscribed to the Google Groups "DSI Studio" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/fb7bce20-d3c7-4a77-b41c-6f322c080b2an%40googlegroups.com.

[Fang-Cheng Yeh]

I am checking the file and will get back to you.
Frank

[Fang-Cheng Yeh]

The b-value is too high, and thus the SNR is not enough.
The recommended value at
https://dsi-studio.labsolver.org/doc/how_to_acquire_dmri.html is
b-max=4000, and for 101 direction, the value can be further reduce to
3000.
Frank

On Wed, May 10, 2023 at 8:15 AM Mark Pinsk <mpi...@gmail.com> wrote:
>

[Mark Pinsk]

I was setting 7K to optimize for differential tractography.  Do you think 3-4K is okay for diff tract ?

[Fang-Cheng Yeh]

Yes.

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/fd179c4e-4052-45da-9016-d2372df3fb61n%40googlegroups.com.

[Mark Pinsk]

<thumbs up>

[Mark Pinsk]

Hi Frank

Would you mind examining my two new pilot scans? I've updoaded the SRC files for the data (no preprocessing).
#1- 101 dir, bmax=3K, 2.4mm, MB=3, TE=120ms, 6 minutes
#2- 258 dir, bmax=4K, 2.4mm, MB=4, TE=132ms, 11 minutes

#1 gives 6 bad slices.
#2 gives 1859 bad slices !

thanks
Mark

[Fang-Cheng Yeh]

I checked the data. The MB=4 with bmax=4k does not have enough SNR,
likely due to no enough gradient performance.
I would suggest using MB=3 and bmax=3000 with 258 directions for your system.
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/b2cbad64-b198-4f06-a01f-ed8a577701f0n%40googlegroups.com.

[Mark Pinsk]

Thanks!  I'll try 258dir with MB=3 versus MB=4 with bmax=3000.