Hi,
> Anyway, following Matthew's lead, I think the issue of localizing
> information in fMRI data is perhaps more interesting. So I should
> probably answer his question:
Thanks - I was hoping that this question would be a bit more central
to the debate, and it's good to get back to it.
> > > ... the thresholded map does not address my assertion 1, and
> > > therefore doesn't address localization. To check - do you
> > > disagree with this? If so, why?
>
> I guess on the whole I don't completely agree with this. Thresholded
> maps do encourage some really flaky inferences, especially if the
> threshold is chosen mechanically and you're looking at a single study.
> Clearly you don't have direct evidence that two areas differ just
> because one exceeds some threshold on some test and the other doesn't,
> so in the absence of a priori hypotheses, single studies will tend not
> to be much help with localization. But you do still have better
> evidence of a difference for one area than for the other. In my view,
> nobody should expect to learn any especially general truths about
> localization of function from a single fishing expedition, whether or
> not the maps are thresholded.
I think I may have misunderstood, but are you saying something like
"it's true that thresholded maps are worse for localization, but then
again, you've got to look across studies"? If so, then surely we
should use the better map, whether we are looking across studies or
not?
> I'm procrastinating on some painful
> paperwork.
I'm waiting for a lift home. Nothing like a mailing list to while
away the time!
See you,
Matthew
|