Hi,
We have a design with 2 ev's coding for 2 different stimulus conditions
and a fixation period that is left unmodeled (see example below).
According to the fsl website it looks like each pe represents each ev vs.
fixation (unmodeled baseline). However, the way each ev is set up (see
below) it seems like the baseline against which each condition is measured
is the rest of the task ( i.e. fixation and the other condition, which are
coded as 0). My understanding is that each pe only represents the
variance that is uniquely associated with its ev (i.e. the variance
associated with the other ev has been removed). I'm left unclear about
what my actual baseline is. For example, say there is artifact in the
signal that occurs only during part of the 1st block of condition 2 and
therefore is not highly correlated with ev2. It seems like, then, in
calculating pe1 the variance associated with that artifact would not be
removed by ev2. So, does this artifact variance go into the calculation
of pe1 as part of the baseline against which condition 1 is being compared
even though it does not occur during condition 1 or during fixation?
ev1 ev2
0 0
0 0
0 0
0 0
1 0
1 0
1 0
1 0
0 1
0 1
0 1
0 1
0 0
0 0
0 0
0 0
1 0
1 0
1 0
1 0
0 1
0 1
0 1
0 1
0 0
0 0
0 0
0 0
Thanks for your time,
Jeff
|