Errant streamlines at back of brain

[Elizabeth Castro]

Hi Frank,

I am doing two analyses with my data, one which required GQI reconstruction and the other which requires QSDR reconstruction.

The first part used GQI reconstruction and the tractography looks great. I am using the same NIFTI files for the QSDR reconstruction, but am having some issues with the tractography. Namely, I am getting some vertical blue lines at the back of the brain that I cannot get to go away.

This data has been topup and eddy corrected in FSL and was the same data used for GQI reconstruction. I have tried flipping/swapping my b-table, eroding the mask used for reconstruction, flipping my image, both the Chen and Hou versions of DSI Studio,  changing the tracking parameters, and performing additional eddy/topup correction in DSI Studio before reconstruction, but to no avail.

I have attached an image of my GQI reconstructed tractography and the QSDR reconstructed tractography for a subject. I did check all the R values and visually inspected my QSDR reconstructions. They all have R values above 60 and I ran outlier detection on them so it does not seem to be an alignment issue, more that there is "extra" brain in the front and back of the brain where I think these errant streamlines may be arising from. 

[Frank Yeh]

This is likely due to eddy distortion or motion. FSL eddy won't work perfectly for all data. 
I can help confirm if the SRC (.sz) file can be uploaded to the DSI Studio website.

--
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 visit https://groups.google.com/d/msgid/dsi-studio/ef389e4e-b954-4e94-96f3-87b08bb5f01dn%40googlegroups.com.

[Elizabeth Castro]

Thank you, Frank! I have uploaded data for one of my subjects. I included the eddy corrected data, the src file, and the gqi and qsdr reconstructed files. I am confused why the GQI reconstruction would look fine for this subject, but not the QSDR reconstruction. This is a similar issue across all of my subjects. The GQI reconstructed tractography looks great, but the QSDR is lacking. 

Let me know what you think!

Best,

Elizabeth