Connectivity matrix - threshold

[JB]

Dear Frank,

Thanks again for your previous reply.

I have followed the "Connectivity matrix and graph analysis" workflow with QSDR reconstruction and 10000000 seeds.

When I take a look at the final connectivity matrix and change the value to dti_fa I see that the matrix is almost all connected, suggesting that I should probably change the threshold of 0,00100.

However, I'm not sure which value I should put here. Can you let me know which threshold should be included or how it can be calculated?

Thanks in advance!

[Frank Yeh]

0.001 seems okay to me, but different groups may have a different
approach, as long as you can justify it.

> --
> 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/b3871fb1-fe09-42c9-8379-f8c3619ee844n%40googlegroups.com.

[JB]

Thanks Frank for such a prompt reply!

Please see in attachment a snapshot of the dti_fa connectivity matrix, in my experience matrices derived from deterministic tractography should be much less connected, wouldn't you agree? In fact the same matrix with "count" as entries/value is much less connected (second screenshot).

It feels like something is wrong here. What do you think?

[Frank Yeh]

The dti_fa weighted one may give a high value even if there is just
one track connecting.
DSI Studio will sample the FA along the track regardless of how many
tracks are there.
I would increase the threshold to zero some entries.
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/5889f0e1-fadc-4486-9998-db715f4e89b1n%40googlegroups.com.