Hej,
I am building a first level model in FEAT, but am struggling to get it correct.
Question 1:
Our simple memory task has 4 conditions, 1) an experimental "encoding" phase, 2) a control encoding phase, 3) an experimental decoding phase (when they answer a quesiton), and 4) a control decoding phase (where they also answer a question). In addition, the "decoding" question phase is present for only half of the encoding phases, i.e. half- of the trials are null-events (so there are only 24 experimental decoding, 24 control decoding, but 48 exp encoding and 48 control encoding).
Am I supposed to model these as 4 EVs? And are the ISI's for the null-event sufficient as "baseline"? The ISI's are variable between 1200 and 3200 ms.
When I create a model with 4 EVs it seems from the efficiency estimation that the two encoding EVs are close to rank deficient. Should I perhaps drop the "control encoding condition" or will that cause a problem, as it is a baseline for the experimental encoding only and not for the 2 decoding phases?
Question 2:
Is it necessary to model the button presses of the subjects separately, or will they be levelled out automatically? It does happen that participants miss a question, so I thought it might be necessary. In that case I will end up with 5 EVs (I could model the motor responses in 1 EV for both control and experimental condition)
I thought an alternative for that might be simply using the response time as a duration for the decoding phase, instead of the entire time of question reading, answering and time to the next ISI. Or would the motor response in such a case still confound the results?
I was also wondering if I use the response time, is it a problem that there are "gaps" to the next ISI, or will they just be included as baseline without a problem?
Thank you so much!
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|