Hi Sterling,
> for B>A (-1 1 0) wouldnt you want to use a similar approach and
> use a 0 1 1 as the mask ? this should give voxels where B is active and
> also greater than A. using 1 1 2 as a mask seems more restrictive
> because differences between B and A would only be tested where A+B was
> significant.
I think you're right, I hadn't thought this through properly. I was
assuming we wanted to distinguish (+ve B > +ve A) from (-ve B < -ve
A), but actually, now you mention it, we are almost certainly also
interested in (+ve B >> -ev A) for example, if most first level B
contrasts were strongly positive while most of A's were negative.
If A was very large and negative, then my [1 1 2] A+B contrast
wouldn't light up, even if B was significantly positive, so we would
incorrectly exclude B>A in this case.
So as you say, B>A masked with B>0 seems much better. And of course,
similarly, A>B masked with A>0. And likewise in the case of looking
for "less deactivation" B>A masked with B<0 is also better than B>A
masked with (A+B)<0, as the latter could fail for strongly activated A
even though B was deactivated.
My apologies to those on the list that I've misled about this in
previous replies! And many thanks to Sterling for pointing it out.
Looking on the bright side, I think working things out the hard way
like this probably results in a firmer understanding (for me, and I
hope also for anyone who's been quietly following the thread on the
mailing list!)
Thanks, and best regards,
Ged.
|