Anova statistic according to tumor localisation

Forums Forums BrainSuite Statistics toolbox (BSS) Anova statistic according to tumor localisation

This topic contains 9 replies, has 2 voices, and was last updated by  shjoshi 2 weeks, 2 days ago.

  • Author
    Posts
  • #939

    leonidastra86
    Participant

    I want to make the statistic not according to the side (right/left) but according to the tumor localisation (contra/ipsilateral side).

    Is it possible?

  • #940

    shjoshi
    Moderator

    Hello,

    Can you please clarify? The hemispheric analysis is only available for cortical thickness (left or right hemisphere). I’m guessing if you want to analyze tumors, you want to conduct a volumetric analysis?

    Shantanu

  • #941

    leonidastra86
    Participant

    I want to make a statistical analysis of the changes in cortex of the contralateral side in patients with infratentorial tumors.

  • #942

    shjoshi
    Moderator

    Hmm.. This analysis could be done in two different ways, but both of them are not currently supported.

    1. For surface based analysis (cbm), you can merge left and right hemispheres and do a combined cortical analysis.

    2. For volume based analysis (tbm), you can do the whole brain analysis, and then report the results on the contralateral side.

    3. Alternately, one can do an asymmetry analysis by subtracting the thickness of the left from the right hemisphere within subject and then do a group analysis and mask out the location of the tumor (you will find a huge effect there obviously).

    Options 1 and 3 are not currently supported in bssr. But you can run option 2.
    Is that what you had in mind?

  • #943

    leonidastra86
    Participant

    I will tell you what I did and also a second idea, which I think won’t work.

    What I did is : analysed the contralateral hemisphere for both and combined the results – the right hemisphere of the patients with a tumor on the left side and the left hemisphere of the patients with the tumor on the right side.

    In this case, the results are accurate but not combined. I would rather be able to mirror the patients for example with tumor on the left side so that I would be able to do one analysis for all the patients.

    Concerning the last sentence (just above this one) I had the following ideas:

    To re-name the atlas.pvc-thickness_0-6mm.smooth2.5mm.left.mid.cortex.dfs
    a)left to right (for the patients for example with left side tumors) so that all the tumors will be on the same side and the side analysed will be the one that I need (contralateral)

    In this case, I am afraid that the volumetric results won’t be right, because the structures of the left side will have the values of the right side.

    b)in place of left to write ipsilateral and in place of right to write contralateral (or reverse, according each time to each case).

    In this case, I am afraid that bssr won’t even work.

    What do you think?

    I hope you can understand my thoughts.

    Thank you!

  • #944

    shjoshi
    Moderator

    Hi,

    I would not recommend doing either a) or b) above.

    1. For within patients analysis only, one could analyze Laterality (left vs right) as the main effect. You could include Sex as a covariate.

    2. Again for within patients, you can do separate analyses for all patients with left sided tumors and right sided tumors.

    3. If you want an across subject (controls and patients), then you could include laterality as a covariate.

  • #945

    leonidastra86
    Participant

    Thats a good idea you gave me. Thank you!
    I will use laterality as main effect and the symptom that I want to examine as a covariate!

  • #946

    shjoshi
    Moderator

    Great. Good luck with the analysis.

  • #964

    leonidastra86
    Participant

    I have my results and started to write the paper. I have however two more questions:

    1) The yellow/pink colours show an increase in cortical thickness and the blue colour a decrease, is it right?
    2) Is there a way to press on the coloured region and see which structure it is? Some of them are easy to identify but some of them not.

    I thank you so much for your help until now and I hope that sonn I will attend a Brain Suite course to learn much more about the software.

  • #965

    shjoshi
    Moderator

    >> 1) The yellow/pink colours show an increase in cortical thickness and the blue colour a decrease, >> is it right?

    That’s right.

    >> 2) Is there a way to press on the coloured region and see which structure it is? Some of them are >> easy to identify but some of them not.

    We are currently developing this functionality, but I’m afraid it won’t be available at least until a few months.

    And yes, stay tuned for the information on the next BrainSuite course.

You must be logged in to reply to this topic.