Dear experts,

I have some doubts about the quandary of covarying in resting-state functional MRI studies.
One class of covariates that have been used in GM and WM analyses are demographic and anatomical variables such as age, sex, and intracranial volume (ICV). However, previous studies suggest that there are other "nuisance variables" less common used (education, IQ, Socioeconomical status, cognitive status), and the inclusion of these variables in a regression model can alter the pattern of findings in such a way that changes that relation of interest due to the removal of meaningful variance from the other predictor(s).
Then, to investigate brain functional connectivity (roi-to-roi) between two-groups with differences in education years, the use of this covariate would be appropiate? Is there any relationship with the average sample size in these studies, and how does it related to the number of covariates used?

Thank you very much in advance,

Best regards,