> 1) I want to use accuracy as a covariate in my MVM/LME. Is LME a better analysis for that or can I still use MVM?
It depends on two factors:
A) What are the explanatory variables (factors) involved in your analysis other than the covariate±?
B) Does the covariate vary within each subject?
> 2) is the format for the covariate file the same as for 3dttest and 3dMEMA?
No. You need to provide the covariate values as a column in the 3dMVM/3dLME script in the table underneath option -dataTable.
Gang