I have noticed that no one ever talks about the magnitude of the
changes in brain activation, only whether they are statistically
significant or not. I guess this is more of the theoretical question
than an FSL question. But, I would like to know whether a ___% change
in BOLD signal in this particular part of they anterior cingulate has
shown up as substantively meaningful in other contexts, beyond just
claims of statistical significance.
For instance, in political science (my home field) good scholars will
not care nearly as much about whether there is any relationship between
some factor X and public opinion. Rather, we care about how big factor
X is substantively and how causally significant it is.
When looking at response latency data early in my work I discovered
that no one bothered to interpret the coefficients substantively, they
just noted statistical significance. Similarly, how often do people
talk about the substantive measure of the changes in BOLD they are
reporting?
Can one get this data from FSL? How so?
I could see it being useful in giving more dimension to the fMRI data
than the pretty activation pictures we love so much. This might
clarify benchmarks of intensity and facilitate clarification in the
study of different causes of intensity of activation. Further, one
could use this data to estimate they power needed for certain effects.
Power analysis can be really useful in grant work as well as deeper
data probing.
Why isn't more substantive quantification (as opposed to mere
statistical reporting) done with fMRI data?
Darren
|