Hi Sinem
I agree with everything Swann said, but just wanted to add two comments:
> Yes it is, but this signal is highly correlated to the BOLD signal in
> adjacent cortex, you ll probably loose if you include it as a regressor of
> non interest.
There are two things to consider: one is the degree to which signals
may be correlated in the raw data you've collected. The second is the
amount of additional smoothing you've applied. For example, if you
smooth your data at 8mm, then signal from gray matter voxels is going
to be smoothed into regions where, anatomically, you might label
voxels as "CSF". However, it's unlikely you'll want to disregard this
signal. So if you do create an explicit mask, you may want to take
this into account (by eroding the CSF mask to make it smaller, or
smoothing a gray matter mask to estimate the degree of overlap).
Similarly, if you are actually interested in CSF or scalp signal, you
will probably want to ensure that the part you extract is unlikely to
be contaminated by signal from gray matter (at least, as much as
possible).
> At the 2nd level, you may want instead to build a mean CSF template for your
> group, threshold it at let say 0.8 to build a binary mask using imcalc, and
> use it as a mask on your SPMs.
Same as above: if you are interested in CSF signal, just be aware of
the degree to which gray matter signal may be picked up by any mask
you use. As is always the case when using explicit masks, the cluster
extent calculations SPM uses based on random field theory are
influenced by the smoothness of the data; the whole brain is
relatively smooth/spherical, but when one starts looking at individual
tissue classes (GM only, CSF only, etc.) they get less smooth. I
don't know if anyone has quantified this, but I would expect you might
get different results. (Note this doesn't effect the voxelwise t
statistics, so it depends on what you are looking for.)
Hope this helps!
Best regards,
Jonathan
--
Dr. Jonathan Peelle
Department of Neurology
University of Pennsylvania
3 West Gates
3400 Spruce Street
Philadelphia, PA 19104
USA
http://jonathanpeelle.net/
|