Free-water correction and acquisition with different b-values.
Hello, I need your help and your opinion. I would like to run a free-water correction of DTI by using dipy. Only to be sure: is it possible to use a single-shell data, with a b=1000? Additionally, I have other data with 26 directions (I do not like to use <30 directions, but I have this kind of data). The 'issue' is that the b-values are: *0 50 250 450 650 850 100 300 500 700 900 100 300 500 700 900 150 350 550 750 950 200 400 600 800 1000* Therefore, all b-values are different. I have never worked with this kind of data before. In your opinion, is it possible to have a correct DTI fit and analysis (and also free-water correction) with this data? Thanks, Maurizio Bergamino ------------------------------------------------------------------- Maurizio Bergamino, Ph.D Research Imaging Specialist Keller Center for Imaging Innovation Barrow Neurological Institute / Saint Joseph's Hospital and Medical Center Phoenix, Arizona, USA Email: maurizio.bergamino@gmail.com
Hi Maurizio, Thank you for your email. Let me answer each of your questions inline below: On Tue, Aug 18, 2020 at 1:35 PM Maurizio Bergamino < maurizio.bergamino@gmail.com> wrote:
Hello, I need your help and your opinion. I would like to run a free-water correction of DTI by using dipy. Only to be sure: is it possible to use a single-shell data, with a b=1000?
This is possible, and there is some work in progress to have that integrated to get it integrated here: https://github.com/dipy/dipy/pull/2087. This should be rather close to integration (the author of the PR has already published some of that work here: https://arxiv.org/abs/2007.06359). If you feel comfortable working with a dev branch of the software, you can pull that down and give it a go.
Additionally, I have other data with 26 directions (I do not like to use <30 directions, but I have this kind of data). The 'issue' is that the b-values are: *0 50 250 450 650 850 100 300 500 700 900 100 300 500 700 900 150 350 550 750 950 200 400 600 800 1000*
Therefore, all b-values are different. I have never worked with this kind of data before. In your opinion, is it possible to have a correct DTI fit and analysis (and also free-water correction) with this data?
I don't think that having multiple different b-values is a particular concern for DTI fitting, so long as the values are within a range that makes sense for the model. In your case, I would worry that the measurements with very low b-values (i.e., < 500 s/mm^2) would also reflect other effects (e.g., perfusion) so might leave those out. I also don't think there is any concern trying to fit the free water version -- even the one that is currently already integrated -- it should work so long as you have multiple distinct b-values, which you do. Hope that helps, Ariel
Thanks, Maurizio Bergamino
------------------------------------------------------------------- Maurizio Bergamino, Ph.D Research Imaging Specialist
Keller Center for Imaging Innovation Barrow Neurological Institute / Saint Joseph's Hospital and Medical Center Phoenix, Arizona, USA Email: maurizio.bergamino@gmail.com _______________________________________________ Dipy mailing list -- dipy@python.org To unsubscribe send an email to dipy-leave@python.org https://mail.python.org/mailman3/lists/dipy.python.org/ Member address: arokem@gmail.com
Dear Ariel, thank you for your email. I will try to run the FW for DTI single-shell. For now, I am running the pre-processing of the data, but I am curious to try the software Maurizio Il giorno mer 19 ago 2020 alle ore 08:57 Ariel Rokem <arokem@gmail.com> ha scritto:
Hi Maurizio,
Thank you for your email. Let me answer each of your questions inline below:
On Tue, Aug 18, 2020 at 1:35 PM Maurizio Bergamino < maurizio.bergamino@gmail.com> wrote:
Hello, I need your help and your opinion. I would like to run a free-water correction of DTI by using dipy. Only to be sure: is it possible to use a single-shell data, with a b=1000?
This is possible, and there is some work in progress to have that integrated to get it integrated here: https://github.com/dipy/dipy/pull/2087. This should be rather close to integration (the author of the PR has already published some of that work here: https://arxiv.org/abs/2007.06359). If you feel comfortable working with a dev branch of the software, you can pull that down and give it a go.
Additionally, I have other data with 26 directions (I do not like to use <30 directions, but I have this kind of data). The 'issue' is that the b-values are: *0 50 250 450 650 850 100 300 500 700 900 100 300 500 700 900 150 350 550 750 950 200 400 600 800 1000*
Therefore, all b-values are different. I have never worked with this kind of data before. In your opinion, is it possible to have a correct DTI fit and analysis (and also free-water correction) with this data?
I don't think that having multiple different b-values is a particular concern for DTI fitting, so long as the values are within a range that makes sense for the model. In your case, I would worry that the measurements with very low b-values (i.e., < 500 s/mm^2) would also reflect other effects (e.g., perfusion) so might leave those out. I also don't think there is any concern trying to fit the free water version -- even the one that is currently already integrated -- it should work so long as you have multiple distinct b-values, which you do.
Hope that helps,
Ariel
Thanks, Maurizio Bergamino
------------------------------------------------------------------- Maurizio Bergamino, Ph.D Research Imaging Specialist
Keller Center for Imaging Innovation Barrow Neurological Institute / Saint Joseph's Hospital and Medical Center Phoenix, Arizona, USA Email: maurizio.bergamino@gmail.com _______________________________________________ Dipy mailing list -- dipy@python.org To unsubscribe send an email to dipy-leave@python.org https://mail.python.org/mailman3/lists/dipy.python.org/ Member address: arokem@gmail.com
Hi Maurizio, Just out of curiousity, are the b-values you mentioned from the OASIS dataset? If that is the case, I have tried fitting it with DTI from DIPY and the results looked reasonable to me! Regards, Shreyas ________________________________ From: Ariel Rokem <arokem@gmail.com> Sent: Wednesday, August 19, 2020 11:56 AM To: Maurizio Bergamino Cc: dipy@python.org Subject: [External] [Dipy] Re: Free-water correction and acquisition with different b-values. This message was sent from a non-IU address. Please exercise caution when clicking links or opening attachments from external sources. Hi Maurizio, Thank you for your email. Let me answer each of your questions inline below: On Tue, Aug 18, 2020 at 1:35 PM Maurizio Bergamino <maurizio.bergamino@gmail.com<mailto:maurizio.bergamino@gmail.com>> wrote: Hello, I need your help and your opinion. I would like to run a free-water correction of DTI by using dipy. Only to be sure: is it possible to use a single-shell data, with a b=1000? This is possible, and there is some work in progress to have that integrated to get it integrated here: https://github.com/dipy/dipy/pull/2087. This should be rather close to integration (the author of the PR has already published some of that work here: https://arxiv.org/abs/2007.06359). If you feel comfortable working with a dev branch of the software, you can pull that down and give it a go. Additionally, I have other data with 26 directions (I do not like to use <30 directions, but I have this kind of data). The 'issue' is that the b-values are: 0 50 250 450 650 850 100 300 500 700 900 100 300 500 700 900 150 350 550 750 950 200 400 600 800 1000 Therefore, all b-values are different. I have never worked with this kind of data before. In your opinion, is it possible to have a correct DTI fit and analysis (and also free-water correction) with this data? I don't think that having multiple different b-values is a particular concern for DTI fitting, so long as the values are within a range that makes sense for the model. In your case, I would worry that the measurements with very low b-values (i.e., < 500 s/mm^2) would also reflect other effects (e.g., perfusion) so might leave those out. I also don't think there is any concern trying to fit the free water version -- even the one that is currently already integrated -- it should work so long as you have multiple distinct b-values, which you do. Hope that helps, Ariel Thanks, Maurizio Bergamino ------------------------------------------------------------------- Maurizio Bergamino, Ph.D Research Imaging Specialist Keller Center for Imaging Innovation Barrow Neurological Institute / Saint Joseph's Hospital and Medical Center Phoenix, Arizona, USA Email: maurizio.bergamino@gmail.com<mailto:maurizio.bergamino@gmail.com> _______________________________________________ Dipy mailing list -- dipy@python.org<mailto:dipy@python.org> To unsubscribe send an email to dipy-leave@python.org<mailto:dipy-leave@python.org> https://mail.python.org/mailman3/lists/dipy.python.org/ Member address: arokem@gmail.com<mailto:arokem@gmail.com>
Hi Shreyas, yes. The data are from OASIS, 24, and 26 directions. For now, I have used only the data with 65 directions, because I do not like to use less than 30 directions. But, I have never seen and worked with this kind of acquisition before and I was curious if it was good for fit and free-water. Thanks for the information, Maurizio Il giorno mer 19 ago 2020 alle ore 13:58 Fadnavis, Shreyas Sanjeev < shfadn@iu.edu> ha scritto:
Hi Maurizio,
Just out of curiousity, are the b-values you mentioned from the OASIS dataset?
If that is the case, I have tried fitting it with DTI from DIPY and the results looked reasonable to me!
Regards, Shreyas ------------------------------ *From:* Ariel Rokem <arokem@gmail.com> *Sent:* Wednesday, August 19, 2020 11:56 AM *To:* Maurizio Bergamino *Cc:* dipy@python.org *Subject:* [External] [Dipy] Re: Free-water correction and acquisition with different b-values.
This message was sent from a non-IU address. Please exercise caution when clicking links or opening attachments from external sources.
Hi Maurizio,
Thank you for your email. Let me answer each of your questions inline below:
On Tue, Aug 18, 2020 at 1:35 PM Maurizio Bergamino < maurizio.bergamino@gmail.com> wrote:
Hello, I need your help and your opinion. I would like to run a free-water correction of DTI by using dipy. Only to be sure: is it possible to use a single-shell data, with a b=1000?
This is possible, and there is some work in progress to have that integrated to get it integrated here: https://github.com/dipy/dipy/pull/2087. This should be rather close to integration (the author of the PR has already published some of that work here: https://arxiv.org/abs/2007.06359). If you feel comfortable working with a dev branch of the software, you can pull that down and give it a go.
Additionally, I have other data with 26 directions (I do not like to use <30 directions, but I have this kind of data). The 'issue' is that the b-values are: *0 50 250 450 650 850 100 300 500 700 900 100 300 500 700 900 150 350 550 750 950 200 400 600 800 1000*
Therefore, all b-values are different. I have never worked with this kind of data before. In your opinion, is it possible to have a correct DTI fit and analysis (and also free-water correction) with this data?
I don't think that having multiple different b-values is a particular concern for DTI fitting, so long as the values are within a range that makes sense for the model. In your case, I would worry that the measurements with very low b-values (i.e., < 500 s/mm^2) would also reflect other effects (e.g., perfusion) so might leave those out. I also don't think there is any concern trying to fit the free water version -- even the one that is currently already integrated -- it should work so long as you have multiple distinct b-values, which you do.
Hope that helps,
Ariel
Thanks, Maurizio Bergamino
------------------------------------------------------------------- Maurizio Bergamino, Ph.D Research Imaging Specialist
Keller Center for Imaging Innovation Barrow Neurological Institute / Saint Joseph's Hospital and Medical Center Phoenix, Arizona, USA Email: maurizio.bergamino@gmail.com _______________________________________________ Dipy mailing list -- dipy@python.org To unsubscribe send an email to dipy-leave@python.org https://mail.python.org/mailman3/lists/dipy.python.org/ Member address: arokem@gmail.com
participants (3)
-
Ariel Rokem -
Fadnavis, Shreyas Sanjeev -
Maurizio Bergamino