We have a dataset with 35 subjects and 5 conditions/scan. We found an
activation of interest in a complex contrast at the third (group) level
Because we wanted to understand the characteristics of this activation, We
created a mask using this activation cluster (forming a "functional ROI",
so to speak). We then used Featquery to extract the activity in this ROI
at the individual subject level. We extracted filtered_func_data, all 5
PEs (corresponding to conditions) and our complex contrast of interest; we
asked for transform to percentages.
The Featquery results are confusing in that all the individual PEs (5
conditions) are negative on average, including the Crosshair condition
which is one of the 5. The complex contrast is positive, which is good,
and in line with the underlying PEs.
So we don't undestand how all 5 conditions in our scan can have negative
PEs. If all PEs are negative, what is the baseline used by Featquery in
this ROI? Is it possible that we did something wrong to lead to these
results, or that we we misreading the output?