This paper was published in 2013: https://www.sciencedirect.com/science/article/pii/S1053811913007064?via%3Dihub#s0040

I was just wondering if the current version of FreeSurfer can still introduce bias in cortical volume measures if ICV is used as an covariate or this has been fixed in the latest FS version. 

Specific to this issue, if nothing has been updated in FS (since 2013), would you still recommend to ALWAYS use ICV as a covariate while comparing CV?

Thanks,

Sahil