 |
 |
Hello, I am seeking feedback on how I chose to set up contrasts for a simple analysis. My project is a structural comparison of grey matter between 2 groups (I am not analyzing functional data; just T1-weighted images). I set up a comparison of regional GM volume between groups. My hypothesis is that experimental group (Group #2) grey matter < control (Group #1). I set up a T-test with an additional covariate being the specific grey matter volumes. If I didn't have that additional covariate, I would simply use a contrast of [-1 1]. But now that I have this additional covariate, should my contrast be different? I.e. does the addition of covariates impact my contrast specification? Likewise, in another test, I will have 2 additional covariates: the grey matter volume and a score on a psychometric test. If my hypothesis is that higher scores are correlated with lower GM volume, is it still appropriate to use the [-1 1] contrast? Assistance is very much appreciated. Thank you, Lori Bruce Boston University School of Medicine