Dear spmers,
I have a task with 3 conditions. I have tried two approaches in modelling
this data with different results.
In the first model I model all three conditions, then look at the contrast
1 0 0.
In the second model I just model one condition of interest - leaving me
with the contrast of 1.
I am confused as to why I get better specificity and activation from the
latter model with only one stick function, as I thought including as much
information as possible about the task was the correct method?
Thanks
jason
|