Hi,

you can use either a typical regression analysis (with a dummy IV), structural equation modeling or ANOVA.

One thought: In case neural activation has a causal influence on depression, controlling for depression will induce endogenous selection bias / collider bias.

Controlling by the way means "including it as a predictor" or stratifying it via group comparisons.

Good luck,

--Holger

Hi, If possible describe types (qualitative or quantitative) of the Stimuli, neural activation and familiarity variables, then I can tell my opinion more accurate.

SEM seems to be the right method, but as Seyed Mahdi Amir Jahanshahi said if you clarify your question, you get a better answer.