Dear SDM team,

I am beginning an IBMA for which I have all the statistical parametric maps, and I wish to use SDM. SDM seems to be based on calculating Hedge's g and estimating the meta analysis on between-group mean differences. However, the images I need to meta-analyse are t-statistic maps based on a general linear model with a continuous independent variable (a serum analyte ranging from 0 continuously); therefore, there are no "groups" to calculate Hedge's g and compare the mean. I am also planning another large meta-analysis where the independent variable is subject age, and thus I face a similar problem with that project.

Can SDM still be used for this purpose, and if so, how? 

In theory I could use fslmaths to convert the t-statistic maps into effect-size maps that are more suitable to continuous data, such as r or Cohen's f2, and then use these images for the meta-analysis in SDM. But would the SDM algorithm still work in this case, or is it built exclusively for patient vs control study designs?

Thank you for your help.

Best wishes,

Aaron.