An additional point to consider is: what would be the result of
scaling by the maximum value output by waver?
If the stimulus presentation is the same for every run, and the
same for every subject, then max. waver value will be constant.
But consider a random stimulus experiment. In such a case, the
maximum waver output will be dependent upon the timing of the
stimuli. If in one random case, a few similar stimuli are more
or less consecutive, then that could result in a significantly
higher maximum value. Even take a max of 1.5 compared to a max
of 1.2. That's a 25% difference in magnitude, and is possible
under such conditions.
Now consider comparing results from one run to another, and then
from one subject to another. If subject A's response to stimulus
X is 25% higher than that from subject B, what does it mean? It
could simply be the result of some difference in scaling.
Basically, if the scaling facter were to depend on the stimulus
timing (and therefore vary per run and/or subject), it seems like
it could be very difficult to compare resulting beta coefficients.
- rick