rs-fMRI Combining Cohorts


Does anyone have advice or know a good paper about considerations for multi-site ra-fMRI, particularly if two datasets have different acquisition parameters? In my case, one of the datasets is relatively old, and has a TR of twice the other dataset. Happy to provide more info as needed.


Hi Steven, i have a paper under review right now looking at REHO results across multiple sites. the result is not pretty (Basically everything seems to depend on FWHM and motion and to a lesser extent cardiac/respiratory patterns. i’ll try to remember to link you to it if it gets accepted,

Thanks for the heads up! It seems like my best bet, if anything, would be to try to correct for those (e.g. acompcor) and set a strict and consistent motion scrubbing protocol.