Hello all, 

given my aim to get a general (exclusively coordinate-based) aggregation of studies with certain patient-control samples, regardless of specific tasks used, how would I treat multiple experiments within single studies? Do I pool the coordinates of all experiments into one, as described for other fMRI meta-analysis approaches (Turkeltaub et al., 2012) and if so, should duplicate coordinates (possibly with different t-/z-values) or close coordinates (e.g., < 8 mm (Euclidean distance)) be excluded (keeping only the one with the highest t-/z-value)? 

I hope this question was not asked and answered before. I searched the forum, but maybe I have overlooked something. Thank you for any help!

Sincerely
Georg

Dear Georg, 

When a study has several experiments, you should only include one of them (the most suitable to your criteria). Including several experiments in one study can lead to biased results, and usually, the tasks are different.

Best, 

Lydia