I agree that there are difficulties in interpreting activation differences
in the absence of behavioral differences. This is particularly
problematic in circumstances where the "control" and "activation" tasks
are not a careful parametric manipulation (i.e., "cognitive subtraction"
issue). I also agree that behavioral measures reflect a combination of
processes that are more distal to the underlying neural bases. Thus,
interpretation of task-related differences in activation in the presence
OR absence of performance difference is not a 1-to-1 mapping in EITHER
case. It simply feels more parsimonious when activation differences are
associated, in some way, with behavioral performance differences.
Having said this.... Though clearly post-hoc, is it possible that the
delay periods used in the study might have invoked STRATEGIC differences
in how the tasks were performed, perhaps giving rise to the observed
activation differences?
wmp
=====================
William M. Perlstein, Ph.D. Phone: (352) 265-0680 ext. 46846
Assistant Professor FAX: (352) 265-0468
Beeper: (352) 334-4433
e-mail: [log in to unmask]
Departments of Clinical & Health Psychology
and Psychiatry
University of Florida
Health Science Center
PO Box 100165
Gainesville, FL 32610-0165
Adjunct Assistant Professor
Department of Psychiatry
Western Psychiatric Institute and Clinic
University of Pittsburgh School of Medicine
Pittsburgh, PA 15213
On Thu, 31 May 2001, Greig de Zubicaray wrote:
> Thanks Joe and Russ for these thoughtful replies,
>
> I posed the question because there is a chronometric tradition within
> experimental/cognitive psychology that might disagree with the relatively
> stronger inference being ascribed to neuroimaging data here (some of my
> cognitivist colleagues have already taken me to task about this), given that
> the presence of a cognitive process of interest hasn't been established in
> the traditional sense.
>
> Regards,
>
> Greig
>
>
>
> on 31/05/01 5:54 PM, Joseph T. Devlin at [log in to unmask] wrote:
>
> > I agree with Russ -- the imaging could well be sensitive to effects not seen
> > in RT data. I think the many studies which have demonstrated "implicit"
> > neural processing are good examples of this. This extra sensitivity comes
> > from the fact that RT (or error data) is a very indirect measure of the
> > underlying neural/cognitive processes because it provides a single score for a
> > combination of processes. rCBF or BOLD are also indirect measures of neural
> > activity but you get many thousands of measures and each is more closely
> > related
> > to regional neural activity than any behavioural measure. Consequently your
> > imaging analysis may be more sensitive to "implicit" processes, which in many
> > cases are interesting/revealing.
> >
> > - Joe
>
|