Messy tract bundles when dealing with large tract numbers

[Georgia Bright]

Hi Frank,

I am experiencing some issues with the 'Recognize and cluster' option with differential tractography results with a large number of tracts. I have found that I need to apply many manual fixes (e.g., removing tracts from the corpus callosum that are close association fibres, or tracts recognised as projection fibre tracts but outside the template space for that given tract).

I have found that the corpus callosum and corticostriatal tracts have the most issues.

I would like to reduce the amount of manual fixes needed, but am unsure how to proceed.

Is there a way for me to do the opposite of the 'Delete Repeated Tracks' function for the built-in atlas?

Or should I use the 'Autotrack...' function to increase the number of tracks in the built-in atlas?

Or is there a better way for me to improve the automatic tract recognition function?

[Frank Yeh]

dT only shows fragments of trajectories—the parts that change. In
contrast, recognize and cluster assume full-length tracts. Because of
this mismatch, things tend to break down, especially for long or
complex systems like the corpus callosum and corticostriatal pathways.
So what you’re seeing is expected.

I would suggest increasing the differences threshold to keep the most
significant results.

[Georgia Bright]

Thanks so much!

I am conducting exploratory data analysis on post-neurorehabilitatory changes.

Should I apply the increased threshold to the rest of the cohort, or just for that one participant's results?

Thanks again.

[Frank Yeh]

Sorry for the late reply.
This would risk cherry-picking the results.
I would report results for different thresholds all together instead
of using just one threshold.

> --
> 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/0ccc0013-12c4-4e9b-9028-9389e168c1cen%40googlegroups.com.

[Georgia Bright]

Hi Frank,

No worries at all. Thank you for your continued assistance and advice.

Sounds great, thank you so much.