In SDM, I need to combine different t-maps from the same group using "combine images" function if these maps are dependent with each other.

Now, I am wondering how to deal with inter-dependence in coordinate results? For example, in one study, they investigate two contrasts but just report peak coordinate and its statistics in the published paper. How to combine these two interdependent results via "combine images" function?

Dear Fen Liu, 

If you ara handling interdependent results, you should set a higher correlation coefficient when using the "combine images" function, which is 0.3 by default (you shold set > 0.5).

Kind regards, 
Lydia

Dear Lydia,

Thanks for your response. However, I just have peak coordinates rather than statistical images right now. In my case, how to use "combine images" function?

Feng

Originally posted by Lydia Fortea:

Dear FEng, 

You could pre-process the coordenates separated as normal. Then combine the obtained _upper and _lower images for each study. Don't forget to change the sdm_table with the name of the new combined image.

Kind regards, 
Lydia

Thanks Lydia,

But how to combine the obtained upper and lower images for each study? Or I need to put all these two files in the sdm_table?

Best,
Feng


Originally posted by Lydia Fortea:

Dear Feng

From the sdm application, you click on the combine image button. First you select the number of images you want to combine (lower and upper separately) and the assumed correlation (you can leave the deafult value or you can set a higher one). Then, you select the maps you want to combinen (all the path will be directly set fron the folder button), and after you will choose the name of the new image. Note that the combine image with the var file will be saved in the folder of the meta-analysis, so you will have to move it into the pp folder for further analysis. You will also have to change the name of the studies in the sdm_table (e.g. you combine analysis_A1.upper with analyisis_A2.upper = analysis_A.upper, you will have to remove analysis_A1 and analysis_A2 from the sdm table (once you conducted the preprocessing) and set analysis_A.

Kind regards, 
Lydia

Hi Lydia, 

I am also first time using SDM to do a meta analysis. 

I already tried the method you stated above, combined the upper and lower maps separately. I also put them into a new pp folder and want to continue with mean analysis. 

However, I have questions with which parameter files to use? Should I also paste the sdm_nd.txt, sdmpsi_params.xml, sdm_mask.nii.gz and etc files generated from the preprocessing into the new folder? Or is there another way to create those files? 

Now I pasted them to the new folder, and ran the analysis. The first 2 steps can run, but it also creates nothing in the  mi folder in mean analysis, so it reports an error of not be able to find any mi file. 

Or should I do another preprocess for the upper and lower results that I get from the last step? 

I am really confused here, and really hope to get some help. 

Thank you in advance! Hope to hear from you soon! 

Best regards, 
Claire

Dear Claire, 

I'm sorry to not have think about it before. In theory, you should change the sdm_maps file by hand, replacing the lines corresponding to the 2 interdependent studies to the combine upper and lower images. However, taking into account all the different steps and manual changes, maybe it's better to put the coordinates from both analysis in one .txt file or ideally only select one of them. 

Kind regards, 
Lydia