Hi everyone,
I am working on an emotional go/no-go scanning paradigm. We are interested in brain activity during correct responses to go and no-go stimuli as well as brain activity during omission and commission errors. To accomplish this goal I made vectors for correct go, correct no-go, omission and commission errors for each emotion we studied (happy, angry, fear, and a black box control).
However, this setup has some issues with subjects who scored perfectly or got a score of 0 on one of the tasks would be missing a vector (since they either had no accurate responses for a given emotion and go/no-go state or would have either no omission or no commission errors for a given emotion). To circumvent this problem, I tried to make 'dummy vectors' with timepoints which would then be censored that could 'replace' the missing vectors.
When I tried to run my data using this method we got errors while doing the t-test since there were too many censored-out vectors. The output that resulted didn't make sense (ie there was no brain activation in controls even though I would expect to see some activation in controls during this paradigm). Please let me know if any of you have thoughts regarding how to proceed.