sdm-help-list > conjunction analysis setup
Showing 1-12 of 12 posts
Display:
Results per page:
Jan 25, 2018  04:01 PM | Colleen Rollins
conjunction analysis setup
Hello,

I'm wondering whether the "Multimodal" tool in SDM could be used to compare 2 meta-analysis of the same modality (ie. comparing 2 VBM meta-anlayses). For instance, was this the tool used to do the conjunction analysis in "Common and distinct patterns of grey-matter volume alteration in major depression and bipolar disorder: evidence from voxel-based meta-analysis" to compare GM differences between MDD and BP meta-analyses?
If so, I was wondering which file should be loaded as the "p-value image" when you open the "Multimodal" tool. Would this be the *z_p_0.00500_1.000_10_p.nii.gz, the *z_p_0.00500_1.000_10_neg_p.nii.gz, or the *z_p.nii.gz file?

Many thanks,
Colleen
Jan 25, 2018  10:01 PM | Marco Bergera
RE: conjunction analysis setup
I would like to add the question, how one might calculate commonalities of 3 distinct fMRI analyses?
Mar 8, 2018  04:03 PM | Joaquim Radua
RE: conjunction analysis setup
Dear all

the "multimodal" analysis should be understood as an "overlap" analysis, and thus it may be used with images of the same modality.

Indeed, if you uncheck the options "Assume error in p-values" and "Decrease p-values as in a better scenario", it will behave as a simple overlap.

To include 3 meta-analyses, you may first combine two, and then combine the result with the third one. I must say that I would probably recommend, in this case, to make it behave as a simple overlap.

Hope this helps,

Joaquim
Mar 30, 2018  01:03 PM | jing shang - sichuan university
RE: conjunction analysis setup
Dear Joaquim,

If 'multimodal' analysis used with images of the same modality only understand as 'overlap' analysis, how could compare the OCD with other anxiety disorders? Just like the paper published Arch Gen Psychiatry 2010; 67:701-711. Should I use Linear model analyses: comparing two groups, binarized OCD as 1, others as 0, then filter OCD?

Thanks a lot in advances!
Apr 16, 2018  01:04 PM | Ali Amad
RE: conjunction analysis setup
Dear Colleen,

Did you find any information concerning your questions about choosing the "p-value image" ? Should we use the *z_p.nii.gz file or the *z_p_0.00500_1.000_10_p.nii.gz/z_p_0.00500_1.000_10_neg_p.nii.gz files.

I am still working on this question, but if you already found any solution it would be awesome ;-)

Many thanks,

Ali

Originally posted by Colleen Rollins:
Hello,

I'm wondering whether the "Multimodal" tool in SDM could be used to compare 2 meta-analysis of the same modality (ie. comparing 2 VBM meta-anlayses). For instance, was this the tool used to do the conjunction analysis in "Common and distinct patterns of grey-matter volume alteration in major depression and bipolar disorder: evidence from voxel-based meta-analysis" to compare GM differences between MDD and BP meta-analyses?
If so, I was wondering which file should be loaded as the "p-value image" when you open the "Multimodal" tool. Would this be the *z_p_0.00500_1.000_10_p.nii.gz, the *z_p_0.00500_1.000_10_neg_p.nii.gz, or the *z_p.nii.gz file?

Many thanks,
Colleen
Apr 17, 2018  10:04 AM | Joaquim Radua
RE: conjunction analysis setup
Dear Jing Shang

as you say, to compare two disorders you may use a linear model with a binary variable that codes one of the disorders with 1 and the other disorder with 0. The name of the map of interest would be "..._1m0".

Hope this helps,

Joaquim

PS: maybe I'm missing something, but I would not filter.
Apr 17, 2018  11:04 AM | Joaquim Radua
RE: conjunction analysis setup
Dear all,

The file should be preferably the one named, z_p.nii.gz.

Note that as some users found, there was a bug in the multimodal function in operative systems. Please wait until we upload a new, corrected version asap.

With the best wishes,

Joaquim
Apr 11, 2020  12:04 PM | Jiaying Gong
RE: conjunction analysis setup
Dear Joaquim,
  
      I would like to ask if the version 6.21 of sdm-psi can normally do the multimodal function study? 

     So the function of the multimodal is exactly the same as the conjunction analysis in MKDA?
 
     If I want to do two functional studies of the same disease, like VBM and CBF, should I use multimodal or Linear Model?
Apr 11, 2020  02:04 PM | Joaquim Radua
RE: conjunction analysis setup
Dear Jiaying

The version 6.21 has the multimodal function. Please tell us if you encounter any problem with it.
This function shows those voxels that are statistically significant in the two meta-analyses.

The linear model might be useful for other objectives, e.g., for comparing two disorders.

Hope this helps,

Joaquim
Apr 12, 2020  09:04 AM | Jiaying Gong
RE: conjunction analysis setup
Dear Joaquim,
 
Thank you for your prompt reply.

A few days ago, I did a meta-analysis of CBF and VBM for the same disease, and both had results, so I want to do a further conjunction


analysis. But when I did the Multimodal analysis, when I first selected the result file, I didn't find the result file with the suffix *p.nii.gz.

Then I added p manually in front of the original result file suffix *nii.gz. The analysis could go on, but there was no result after running the

Multimodal analysis. As can be seen from the previous CBF and VBM results, there should be some overlap between them. Isn't this

Multimodal analysis doing conjunction analysis? How to explain this result?

Thank you again.
Attachment: Figure .png
Oct 11, 2022  03:10 AM | gaoacu
RE: conjunction analysis setup
dear Joaquim:


        could you explain "comparing two disorders"using linear modal detailed?i read this method in other literatures,but ican't interpret it.


Thank a lot!

gao
Nov 1, 2022  01:11 PM | Lydia Fortea - Instituto de Investigaciones Biomédicas August Pi i Sunyer (IDIBAPS)
RE: conjunction analysis setup
Dea Gong, 

To perform a multimodal analysis you have to select the folder for each computed meta-analysis, and SDM will automatically fill the Model name, be careful what you define in the the output folder, it is there where you will find the results of the multimodal analysis.

Kind regards, 
Lydia