Hi all.
I'm writing in hopes of raising
a discussion on the phenomenon of negative t values with conjunction
analysis.
So my first question is....if my
conjunction analysis (I'm using the Global null conjunction hypothesis)
returns activation patterns with negative t values, does that mean the
activation patterns are actually deactivation? And maybe more importantly are
negative t values even statistically relevant (i.e. can I report them as
actual results?). I find that the probability of getting negative t values
increases with more subjects I include in the conjunction analysis (i.e. all
positive t values with a conjunction of 10 subjects, but a majority of negative
t values with a conjunction of 20 subjects). I also get negative t values when
I'm looking a deactivation patterns (i.e. areas less active during the stimulus
over baseline). So I guess what I'm really asking is how do I interpret the
negative t value results.
My second question revolves
around how to display results with negative t values. I'm able overlay
the blobs of negative t values on the three slices using the overlay slices
button on the Results GUI, but rendering the results on the 3D brain only
displays positive t values and same with using the display_slices command. If I
use the save button to save the results as an .img file, SPM does not
save any of the negative t values....I think SPM somehow writes any t
value of zero or less as an NaN. And so I've been advised that in order for
SPM to save negative t values, it would involve re-writing SPM code which is
definately beyond my scope of expertise.
Anyways, I'm wondering if anyone
else has had this sort of problem with displaying results with negative t
values, and maybe how to overcome it...I'm hoping that I won't be limited to
using the 3 slice overlay button, since every three slices elicits a different t
value colour range.
Thanks in advance for any
discussion.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Nicholas Masse, M.Sc.
Research
Technician
Neurovascular Research Lab
School of Kinesiology
University
of Western Ontario
Rm. 3110 Thames Hall
519.661.2111 x88526
[log in to unmask]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~