Group connectometry : extract mean dti_fa value of significant results

Jean Kim

unread,

Mar 31, 2021, 3:16:48 AM3/31/21

to DSI Studio

Dear Dr. Yeh,

Hello, I performed group connectometry (thank you for the great tool btw), and it shows both positive correlational pathways.

To extract the mean FA value of each subject to analyze correlation with behavioral measure, I followed <http://dsi-studio.labsolver.org/Manual/diffusion-mri-connectometry#TOC-Scatter-plots> this description saying load connectometry database and significant tracts, then extract tract statistics to see mean FA value of each subject and more.

However, it shows something unacceptable (?) or hard interpretation values of FA, which I attached to this conversation.

As you can see, positively correlated tracts' mean FA shows very low in the majority. Well, fornix generally shows a lower FA value than other regions, but still I think it is too low.

Is this an error? Or should I recalculate in different ways?

Thank you in advance.

Sincerely,

Jean

report.csv

Frank Yeh

unread,

Mar 31, 2021, 8:34:02 AM3/31/21

to dsi-s...@googlegroups.com

Is you connectometry database created from "qa" or "fa"?

It is likely that it is from "qa" which has values depending on the b-table.

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 on the web visit https://groups.google.com/d/msgid/dsi-studio/eddfa64e-4049-4700-9726-16db3ce41cc4n%40googlegroups.com.

Jean Kim

unread,

Apr 1, 2021, 2:30:57 AM4/1/21

to DSI Studio

If I get your questions correctly, I believe I chose dti_fa to create a connectometry database.

On Step C2: Create a Connectometry Database, I loaded all of qsdr fib files, then I selected dti_fa in Index of Interest and created the database.

I attached the output filename screenshot and what I chose on Step C2.

Thanks a lot for your quick response.

Sincerely,

Jean

connectometry_db.PNG

Step_C2.PNG

Frank Yeh

unread,

Apr 1, 2021, 8:53:39 AM4/1/21

to dsi-s...@googlegroups.com

I see. You may check the original FIB file and see if the FA values are correct.

FA can be lower if there is a partial volume of free water or either acquisitions problems.

Frank

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/1b32c0dc-001f-4ea6-aab8-20c4c19f3f3fn%40googlegroups.com.

Jean Kim

unread,

Apr 2, 2021, 1:54:16 AM4/2/21

to DSI Studio

Dear Dr.Yeh,

Thank you again for your advice.

I loaded the original .fib.gz file used for reconstructing qsdr files to create connectometry database in Step T3:FIber tracking.

Then, I opened the significantly positive tracts, and ran the statistics.

It shows FA values of 0.35, which is definitely different from a previous value I got, 0.07. ( In the previously attached 'report.csv', subject is 'MR_001_0088975_DTI mean_dti_fa').

Should I extract FA values from the original fib.gz files to analyze correlation between behavioral measures? Would it be acceptible?

Sincerely,

Jean

Frank Yeh

unread,

Apr 2, 2021, 9:17:45 AM4/2/21

to dsi-s...@googlegroups.com

Yes, it should also work.

Frank

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/601a9d03-7902-4299-bc21-1d72ce90c203n%40googlegroups.com.

Jean Kim

unread,

Apr 21, 2021, 7:30:56 AM4/21/21

to DSI Studio

Dear Dr.Yeh,

I have tried to export the significant FA values of the connectometry results (pos_corr.tt.gz) from the original FIB file instead of qsdr.

I realized that "pos_corr.tt.gz is registered to hcp1065, but my original FIB file is not aligned to hcp 1065.

So it is quite hard to get the correct FA values.

Is there any suggestion to extract FA values on significant tracts of the connectometry results (pos_corr.tt.gz)?

Sincerely,

YaeJi Kim

Frank Yeh

unread,

Apr 21, 2021, 8:57:53 AM4/21/21

to dsi-s...@googlegroups.com

Hi YaeJi,

You may try this to see if it works for you:
http://dsi-studio.labsolver.org/Manual/diffusion-mri-connectometry#TOC-Scatter-plots

Best,
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/f959427a-232e-4229-a05d-24ffcf035d7en%40googlegroups.com.

Jean Kim

unread,

Apr 22, 2021, 3:28:58 AM4/22/21

to DSI Studio

Dear Dr. Yeh,

I believe that is exactly what I have done before using the original fib file. (as you can find in the earlier conversation in this thread).

There I got an issue having extremely low FA values (I was using dti_fa when reconstructing connectometry database), which is quite unacceptable to interpret ( i.e. FA value of 0.05 on Fornix_L).

You suggest me to load pos_corr.tt.gz file on the original fib file, but as I questioned, the alignment is far apart.

I attached the screenshot of the registration issue when fiber tracking.

'wholebrain.PNG' shows the whole brain fiber tracking of the original fib file, and 'pos_corr.PNG' is the significant results from the connectometry analysis.

- I marked where 'pos_corr.tt.gz' lies on the whole brain image as red.

Sincerely,

Jean

pos_corr.PNG

wholebrain.PNG

Frank Yeh

unread,

Apr 22, 2021, 9:38:18 AM4/22/21

to dsi-s...@googlegroups.com

The cord tracks are in the template space, not subjects native space.

You will need to follow the documentation I sent you to get the right fa values.

Best regards

Frank

To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/eadf44b8-10e6-4d53-9f72-fd4de37f138cn%40googlegroups.com.

Reply all

Reply to author

Forward