Hi all
I have to model an event related task (TR=3000ms) testing conditions composed by the following: (their duration in ms):
fix point (1000)- cue stimulus (200)- blank(100)- target stimulus (until subject response)-subject response-second target (until subject response)-2nd response
we are interested in the interactions between cue and target, considering the reaction times of the first response as a primary information as we observe response facilitation or slowing according to cue-target congruency.
what happens after the first response is of no interest.
reading previous post over this topic. I saw different possibilities, but i'm not sure which is the most correct.
considering using a three column format.
I assume that cue and target represent one event as their onset is separated by 300ms. correct?
A) model each condition as single event.
onset: cue stimulus onset
duration: from cue to subject response
intensity: 1s
like https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;1c7d1bad.0911
B) model 2 EV for conditions
EV1:
onset: cue stimulus onset
duration: ????????
intensity: 1
EV2:
onset: ???????
duration: ?????
intensity: reaction time
EV1 and EV2 are orthogonalized.
this according to this reply :
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;4528eb2.0910
in this case. which are the two durations and the onset of EV2?
from another post, I read that also events of no interest must be modeled. so should target2 be modeled also ?, as EVtarget2+ EVresponse2 or simply EVtarget2?.
when contrasting different conditions.
in A case I would simply EV1:1....EV4:-1 and the reverse.
in case B: EV1:1, EV2:1,......EV7:-1,EV8:-1 ????
thanks in advance
cheers
Alberto
|