QUESTION: I am interested in using the MyRegression function to examine whether heterogeneous effects (QH) can be predicted by differences across studies. After running the function I see htm files, but there are no MyRegression.nii files that I thought would describe where there were associations. Am I missing files in my output? Are the results in the sdm_main.nii file?

RESPONSE: You can use the "Threshold" button to statistically threshold the regression images and display the results. Additionally, the sdm_main text file refers to the image numbers in the 4D sdm_main.nii file in case you were interested in the unthresholded regression images.

Hello! I have a related question. I'd like to see the unthresholded map for MyMean. Is it the sdm_main map that I should be looking for? I tried but the result from the thresholded map contains area not in the unthresholded map. I wonder whether if I made a mistake.
Thanks

Dear _,

as  you point out, you should look at the sdm_main file (which is a 4D image!). However, I'd like to note that a new version of SDM will be available very soon and in this version the unthresholded maps will be straightforwardly available.

With the best wishes,

Joaquim

Hi Dr. Radua, 
Sorry about the multiple questions but I am still trying to learn how this program works. 
I have a few images as input to the meta-analysis.  

First question, do input images all need to be in the same dimensions? Several images that I had did not map directly on the standard MNI 2mm space ( in fact I am still trying to figure out how to visualize these images on a template).  The program could still run but I am not sure if it was correct. 

Second, the coordinate of one image decreased from right to left on the x plane. Should I convert it by fslswap BEFORE using the conversion button on SDM? (The rest all increased from right to left) 

Third, does SDM 'convert images' create a new image that we are supposed to check for consistency? It sounds from the instruction that 'convert images' creates new images that are compatible with SDM, but I did not see one.

Fourth, as mentioned in my previous post, my thresholded map included area that was not shown in the unthresholded map. What might be the cause of this? Also, shouldn't each voxel has a value in the unthresholded map if I used images as inputs? 

Thank you very much. 
*I attached one of the raw images I have here. 
Peggy

Dear Peggy,

the input images may have different dimensions (as far as they cover all brain) but must be in a stereotactic space (e.g. MNI). If any image is not in a stereotactic space, you should transform the image to MNI (or alternatively Talairach). Please note that this doesn't mean simply swapping dimensions but conducting a proper normalization.

SDM 'convert images' only modifies the header of the images (e.g. to set that the images are t-images with 30 df in MNI space) and additionally may conduct simple conversions from p- or z- values; these modifications are necessary because software packages may not set the header with these parameters. However, as noted before, the image should be in a stereotactic space, and the voxels be p, t or z-values. You can open the converted images with a view such as MRIcron or FSLview, but you should also check that the maxima of the SPM pre-processing report approximately coincides with those reported in the paper.

All areas of the thresholded map should appear in the unthresholded map, in which all brain voxels should have not-null values. Have you checked that the viewing parameters (color range or brightness / contrast) is the same in the thresholded and in the unthresholded maps?

Best wishes,