hi, i'm Luca benedettini, i'm graduating in biomedical engeneering, and i want to thank you for this guide on DTI analysys. It's helping me a lot with my thesis, your explanation is clear and complete, very very good. THANK YOU
Thank you for the great tutorials! Honestly, finding a good tutorial for lab tools is not easy! always demanding a lot of time to go through written reports and then fail on command executions. But you made the whole a lot easier. Thank you again:)
Andrew, thanks for these videos. If I'm doing a multicenter trial. How do I include the centers as regressors in GLM? I have been told to put the number of centers minus 1 for the columns and input binary data. Would you agree?
Thank you so much for your tutorial. I wonder to know what mask should I use for free water corrected maps. Can I use Mean-FA-skeleton? Or I should create mean-FA-freewater-skeleton?
Hello, I love your little childhood anecdotes! Quick question for you... What if I was only interested in looking at the corpus callosum fibers.. How would you specify this and gather the mean FA difference in different subdivisions? Thank you so much.
Dear Jahn Do you have some courses about the FMRIB Diffusion Toolbox.Such as DTI model fit,BEDPOSTX modelling crossing fibres,PROBTRACKX propagating uncertainty in tractography
Thanks for this great tutorials! I am trying to perform an interaction analysis of a categorical variable (3 steps) and one continuous variable. I am wondering how to set up the Glm in TBSS/FSL. Could you help me with this, please ?
Hello, I believe that you need to create a new regressor which is the product of the categorical variable multiplied by the continuous variable. I'm not 100% sure about this, so I would also run it by the FSL listserv. Best, -Andy
Hi Andrew, whole series were so helpful, yet there's a particular thing I did not understand.. What exactly design matrix consists of? my fsl does not opens glm guı so ıs there an another way of creating design.mat and design.con.. like a script command or so? any advice would be helpful, thanks..
Hey Ozge, the easiest way would be to type Glm_gui, fill in the regressors and contrasts as you normally would, and then save it as a design.mat file. -Andy
wow.. thanks for that, İ've just read it.. Now I have a quite more complicated problem: the thing is I have two different images that belongs two different groups and I want to compare these two, which means ı guess, there should be 2 timepoints in my design.mat file and a matrix of 2x2 dimensions. But when I compile, it gives out the error, "number of errors in your design matrix does not match the number of timepoints in your data" have you ever seen such error? What might be the reason.. Thanks again for your kindness.. Özge
Hello, thank you Andrew for your videos, it helps me a lot. i'm just starting with TBSS and i was wondering if there is a routine to extract the results (significant region of the FA skeleton), to have talairach coordinates etc. Thank you beforehand
Hi Eya, Unfortunately, I don't have anything like that on hand. I would start with featquery, and see if you can script it with your DTI data to extract data from your ROIs. -Andy
HI Andrew thank you for your answer, i'm a Tbss beginner and i thaught that there is a kind of output of Fa value (individual and group values) like the other tools i used to use (trakvis), but i just realized that's Tbss provides olnly visual results (brain maps with Fa skeleton) so if i need these value i have to script it myself lol (how great, i'm not an engineer) anyway thank you again for you help Eya
Hi Jlingo unfortunately TBSS output is limited to visual skeleton map and you have to write a script (python etc ..) to extract Fa value from significant cluster voxel. Good luck for your results
Hi Debasish, I'm not exactly sure in this case; these days, I usually use MRtrix for diffusion analysis, in which case you simply extract the data from the relevant cells of the connectome, and then perform whichever test you want. Best, -Andy
hi, i'm Luca benedettini, i'm graduating in biomedical engeneering, and i want to thank you for this guide on DTI analysys. It's helping me a lot with my thesis, your explanation is clear and complete, very very good. THANK YOU
Great illustration! Please make a video on the process/steps to report the results in numbers/coordinates etc from that final image/output.
Thank you for the great tutorials! Honestly, finding a good tutorial for lab tools is not easy! always demanding a lot of time to go through written reports and then fail on command executions. But you made the whole a lot easier. Thank you again:)
Thanks Andrew, very helpful!
Thanks Andrew. Really helpful!
Andrew, thanks for these videos.
If I'm doing a multicenter trial. How do I include the centers as regressors in GLM? I have been told to put the number of centers minus 1 for the columns and input binary data. Would you agree?
Thank you so much for your tutorial. I wonder to know what mask should I use for free water corrected maps. Can I use Mean-FA-skeleton? Or I should create mean-FA-freewater-skeleton?
Hello,
I love your little childhood anecdotes! Quick question for you... What if I was only interested in looking at the corpus callosum fibers.. How would you specify this and gather the mean FA difference in different subdivisions?
Thank you so much.
Dear Jahn
Do you have some courses about the FMRIB Diffusion Toolbox.Such as DTI model fit,BEDPOSTX modelling crossing fibres,PROBTRACKX propagating uncertainty in tractography
Thanks for this great tutorials! I am trying to perform an interaction analysis of a categorical variable (3 steps) and one continuous variable. I am wondering how to set up the Glm in TBSS/FSL. Could you help me with this, please ?
Hello,
I believe that you need to create a new regressor which is the product of the categorical variable multiplied by the continuous variable. I'm not 100% sure about this, so I would also run it by the FSL listserv.
Best,
-Andy
Hi Andrew, whole series were so helpful, yet there's a particular thing I did not understand.. What exactly design matrix consists of? my fsl does not opens glm guı so ıs there an another way of creating design.mat and design.con.. like a script command or so? any advice would be helpful,
thanks..
Hey Ozge, the easiest way would be to type Glm_gui, fill in the regressors and contrasts as you normally would, and then save it as a design.mat file.
-Andy
wow.. thanks for that, İ've just read it.. Now I have a quite more complicated problem: the thing is I have two different images that belongs two different groups and I want to compare these two, which means ı guess, there should be 2 timepoints in my design.mat file and a matrix of 2x2 dimensions. But when I compile, it gives out the error, "number of errors in your design matrix does not match the number of timepoints in your data" have you ever seen such error? What might be the reason..
Thanks again for your kindness..
Özge
Hello, thank you Andrew for your videos, it helps me a lot.
i'm just starting with TBSS and i was wondering if there is a routine to extract the results (significant region of the FA skeleton), to have talairach coordinates etc.
Thank you beforehand
Hi Eya,
Unfortunately, I don't have anything like that on hand. I would start with featquery, and see if you can script it with your DTI data to extract data from your ROIs.
-Andy
HI Andrew
thank you for your answer, i'm a Tbss beginner and i thaught that there is a kind of output of Fa value (individual and group values) like the other tools i used to use (trakvis), but i just realized that's Tbss provides olnly visual results (brain maps with Fa skeleton) so if i need these value i have to script it myself lol (how great, i'm not an engineer) anyway thank you again for you help
Eya
Hi Eya,
Did you by chance figure out how to extract the FA values using FSL? Is FSL's stats capability really limited to visual comparison?
Thanks!
Hi Jlingo
unfortunately TBSS output is limited to visual skeleton map and you have to write a script (python etc ..) to extract Fa value from significant cluster voxel.
Good luck for your results
what to select for paired two groups?
Hi Debasish, I'm not exactly sure in this case; these days, I usually use MRtrix for diffusion analysis, in which case you simply extract the data from the relevant cells of the connectome, and then perform whichever test you want.
Best,
-Andy
Sir
How did you get that all_FA_skeletonized.nii file? From which step was it originated?
Hi Debasish, I think that may have been in an earlier video, such as Steps 5 and 6: czcams.com/video/bxSqZKWeQQc/video.html
Best,
-Andy