Hello FSL message board,

I have a perplexing conundrum that I would appreciate some feedback on.

I'm using a remote associate task in my scan, and modelling periods of correct/incorrect trials in my Feat design. (So for each condition, there are two potential EVs.)

Let's say a participant gets a condition completely correct, such that there isn't an incorrect EV made in my data scripts. When I run their first level Feat, should I:

A) only include the EV for correct answers (no 2nd EV)?

or

B) include the EV for correct answers, and a "null" EV modelling incorrect (a string of zeros in the text file)?

Currently, I've been following option B figuring I should keep everyone's models the same. But our results are turning out a bit funky, so I wanted to see if I was making a wrong judgement in setting up my first levels.

Thank you,

Kevin Japardi