Correct method for extracting FA and MD values of specific tracts.

167 views
Skip to first unread message

Gauthami Nair

unread,
May 14, 2020, 11:07:45 PM5/14/20
to DSI Studio


Hello sir,
I am very new to DTI and using DSI studiop and I had a few confusions with regard to the correct steps to be used for extracting data. I am currently working on a project with MS group of patients and in my design I needed to statistically calculate the difference in FA, AD,MD and RD values among patients and controls.
Up till now I have used connectometry analysis method.
- I created the fib file database for each of the Index - FA, RD, MD and AD
- Then I proceeded with group connectometry analysis and identified the tracts having significant differences among the two groups.

But I am still not sure on how exactly I need to extract the ROI data. Do I open the initially made connectometry database file and draw the tracts I have identified, extract the ROIs of those tracts and then compare the control and patient group?
Or do I have to extract the ROI's after the group connectometry analysis  has been done. I do hope that I have been able to make sense with what I said. Thankyou for your time.

Fang-Cheng Yeh

unread,
May 14, 2020, 11:23:39 PM5/14/20
to dsi-s...@googlegroups.com
Hi Gauthami,

    You may report the tract that have differences among groups as final findings. You don't need to extract value from ROI.

    If you would like to extract values from these tracks, the documentation is available at  http://dsi-studio.labsolver.org/Manual/diffusion-mri-connectometry#TOC-Scatter-plots 

    Hope this helps.

Best,
Frank 

    

Gauthami Nair

unread,
May 14, 2020, 11:45:47 PM5/14/20
to dsi-s...@googlegroups.com
Dear Sir,

I actually did follow the steps you have mentioned above but I do not know what I am doing wrong. 
When I picked out the specific track ROI's and compared them between the two groups, I actually did not find any difference in the values. I wasn't sure if this was because the data had already been processed in the Group analysis.
When I did a trial method by extracting the ROI from the database file I was finding a significant difference between the two groups. This is what led to my confusion.

Sincerely,
Gauthami

--
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/CAG_QrtL-xRmqZ_-fLYHFXmbLLpoqXnprUdvxdJhRGcWe0c344A%40mail.gmail.com.

Fang-Cheng Yeh

unread,
May 15, 2020, 12:37:57 AM5/15/20
to dsi-s...@googlegroups.com
There could be several affecting factors causing this result.
Could you post the report generated from connectometry analysis?

Gauthami Nair

unread,
May 15, 2020, 12:44:29 AM5/15/20
to dsi-s...@googlegroups.com
sure sir.
I have reports for all 4 of the index's below.

AD between control and Patient.html
FA between control and Patient.html
MD between control and Patient.html
RD between control and Patient.html

Fang-Cheng Yeh

unread,
May 15, 2020, 12:54:05 AM5/15/20
to dsi-s...@googlegroups.com
When I picked out the specific track ROI's and compared them between the two groups, I actually did not find any difference in the values. 

What are the steps you used to picked out the ROI and compared them? 
 
When I did a trial method by extracting the ROI from the database file I was finding a significant difference between the two groups. This is what led to my confusion.


What is the "trial method" you are referring to?
Best,
Frank
 
Sincerely,
Gauthami

On Fri, 15 May 2020 at 08:53, Fang-Cheng Yeh <fran...@gmail.com> wrote:
Hi Gauthami,

    You may report the tract that have differences among groups as final findings. You don't need to extract value from ROI.

    If you would like to extract values from these tracks, the documentation is available at  http://dsi-studio.labsolver.org/Manual/diffusion-mri-connectometry#TOC-Scatter-plots 

    Hope this helps.

Best,
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/CAG_QrtL-xRmqZ_-fLYHFXmbLLpoqXnprUdvxdJhRGcWe0c344A%40mail.gmail.com.

--
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.

Gauthami Nair

unread,
May 15, 2020, 1:44:13 AM5/15/20
to dsi-s...@googlegroups.com
Initially, I did the steps that were mentioned in the manual to compare between the groups and I did not find the significant differences as stated by previous papers in the same field of study.

 The trial method I did was -
- I noted the specific ROI tracts given in the final report from the group analysis
- I loaded the connectometry database file for FA in the Fibre tracking option
- Next, I assigned specific regions from the atlas
- and then after that, I did fiber tracking and took the statistics for the same.

Doing this I was actually able to get a difference in value between the control and patient population, but I wasn't sure if using this methodology was actually correct.

Fang-Cheng Yeh

unread,
May 16, 2020, 11:44:25 AM5/16/20
to dsi-s...@googlegroups.com
Your trial method is also correct. It is an atlas-based fiber tracking to get tracks statistics.

About the no difference in connectometry ROI analysis, I can test it on my side, if you are allowed to upload the connectometry db and the demographics file.

I will see if I can reproduce the problem and fix it.

Best,
Frank


Gauthami Nair

unread,
May 16, 2020, 12:45:31 PM5/16/20
to dsi-s...@googlegroups.com

Oh that’s wonderful to hear that I wasn’t doing the wrong thing.

But I’m still confused about when I do use these results for publication, will it be methodologically accurate creating a comparison on FA and MD values between the two groups using the trial method I mentioned and also putting the report I got from the connectometry analysis together?

Of course there will be no issues in uploading the data but it’ll take me atleast till Monday morning because everything is kept in the lab in the hospital that I’m working in. I’m currently in India.

It could also be that I may be doing an incorrect step after I get the group analysis report. I look forward to seeing the results you get for the same.

Thank you so much for taking so much time and helping me out. I am much obliged.

Sincerely,
Gauthami

Message has been deleted

Fang-Cheng Yeh

unread,
May 18, 2020, 7:59:20 AM5/18/20
to dsi-s...@googlegroups.com
Thanks. I will check and let you know.
Frank

On Mon, May 18, 2020 at 1:57 AM Gauthami Nair <gautham...@gmail.com> wrote:
sir,

I have attached the database for FA and the demographics here for doubt clarifications. This database is sent with special permission from my PI. 
I kindly request that this information remain confidential. 

Thank you so much for your help.

--
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.

Fang-Cheng Yeh

unread,
May 20, 2020, 11:21:32 AM5/20/20
to dsi-s...@googlegroups.com
I ran the analysis, and the extracted values from the connectometry tracks show obvious groups difference with large effect size:
image.png
The same result can be generated following the instruction in http://dsi-studio.labsolver.org/Manual/diffusion-mri-connectometry#TOC-Scatter-plots
Also, as a side note, you may consider adding age and sex as covariates in the connectometry analysis to consider their effect on FA.
The current analysis you have did not have them.

Best regards,
Frank

Gauthami Nair

unread,
May 29, 2020, 12:43:34 AM5/29/20
to dsi-s...@googlegroups.com
Thank you very much for the help sir.
I shall be adding these new covariates and redoing the analysis.
I also got the expected results.
Thank you so much for your time.

Sincerely,
Gauthami

Reply all
Reply to author
Forward
0 new messages