Different correlational tractography results on two different systems

[Anna Kannonier]

Dear Frank,

I’m encountering a reproducibility problem when running correlational tractography on two different systems, both using the CPU version (July 7, 2025): 

- Laptop: Windows, CPU

- Desktop: macOS, CPU 

I get two different results, even though I'm using the exact same datasets and statistical settings:

- Different tracts are detected

- Different groups show significant effects

I'd be very grateful for any help on how to get consistent results!

Best wishes,

Anna

[Frank Yeh]

This is a rare condition unless ROI is used, which makes the analysis
more sensitive to seed sequence (determined by the random sequence).
I would increase the permutation count until the results converge to
become consistent.

Also, you may check the HTML report to confirm if the method section
is identical.

If still doesn't work, please let me know.

Best regards,
Frank

> --
> 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/f9da25bd-d5d4-4fc0-a901-5a246c5539e5n%40googlegroups.com.

[Anna Kannonier]

Thank you for your help! I increased the permutation count from 4000 to 10 000 and now I'm getting consistent results (for whole-brain analysis).

My sample is relatively small (n = 29, groups 13 vs. 16) and quite heterogeneous. Do you think this could explain why the results were more variable at lower permutation counts?

[Frank Yeh]

Likely there is an outlier.
I would perform 'quality control' on the SRC files to see if there is
any dataset that has low NDC or low contrast.

Best,
Frank

> To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/617fab03-b025-44be-b0dd-43ec0af73dbcn%40googlegroups.com.