Dear Joaquim,

I am planning to conduct a meta-analysis mixed with coordinates and statistical maps to explore the neural basis of a cognitive function. I would like to use SDM as my analytic tool, but still have a few questions:

• Should I copy the coordinates txt files and statistical maps to the working directory, convert statistical maps, and start the procedures from preprocessing?
• Some studies use parametric estimation (e.g., a trial-by-trial parameter reflecting working memory capacity), whereas others use categorical analysis (e.g., high > low working memory condition). As most SDM examples are based on studies using categorical analysis, I am not sure if I could also include those using parametric estimation in a single meta-analysis.

Thank you very much and I look forward to hearing from you!

Best wishes,
Florian

Dear Florian,

Thank you for your email. Regarding the questions:
1. Yes.
2. You can mix them. Note however that they use different methodologies and, as such, their inclusion might be a potential source of heterogeneity. That said, nearly all meta-analysis include studies with methodological differences (different MRI devices, different neuroimaging software or processing parameters, different fMRI tasks, etc), as otherwise the number of studies would be too small. You might have a look at I2 of the peaks to have an idea of the heterogeneity.

Hoping that this helps,

Joaquim

Dear Joaquim,

Thank you for your reply! I have another follow-up question regarding the sample size. Some other software, like GingerALE, seems to suggest including at least 15 studies for a meta-analysis. Is it also the case for SDM in general?
As image-based analysis is usually more difficult and the contrast map contains more information than peaks, does it mean that fewer studies could be acceptable for an image-based analysis compared with the coordinate-based approach?

Many thanks!
Florian