Dear Nader,
I don't think that's the case.  The mapping of a particular activation onto a
particular cognitive function is based upon your theory of the manipulation that
resulted in the activation.  That is, when you design an imaging paradigm you do so
with the goal of isolating the neural processes that are associated with a
particular cognitive process, which is based upon your theory of what processes are
engaged by what tasks.  Although it is always nice to have convergent behavioral
data, it's not necessary.

Take the following example.  I have a group of young adults perform a task that they
can perform relatively quickly without a great deal of effort, say simple mental
arithmetic.  now I take a set of elderly people with mild cognitive impairment, and
pay them a large amount of money such that I temporarily motivate them to perform
the task at the same speed as the young adult.  I should be able to set this up such
that there are no differences in response time or accuracy between the groups.
however, when I look at activation there will clearly be differences, and these
differences can probably be interpreted in a relatively straightforward manner.

More generally, it's clearly the case that fMRI data can be more reliable than RT
data.  Back when I ran RT experiments, I remember being happy if 2/3 of the subjects
would show a particular effect, because RT's are so noisy, at least for the kinds of
tasks that I was working with.  However, it is not uncommon at all to find regions
in an imaging study where 100% of subjects show a particular effect.  One certainly
wouldn't want to peg the interpretability of the more reliable measure (imaging) on
that of the less reliable measure (RT).

Cheers,
Russ

Nader Amir wrote:

> Dear Russ:
>
> I am sure I am being naive, but is it not the case that this obviates the whole
> idea of functional, in fMRI.  In the absence of any reliable behavioral
> indicators, as well as a control condition, how can we be sure that any
> activation
> is the results of that particular function.  I see many papers in the field that
> do this
> so I would really appreciate it if someone explained the logic of looking at
> activation
> across conditions when there are no behavioral differences across conditions.  It
> seems
> to me that in the absence of some reliable external behavioral indicator we have
> circular
> argument for the role of that particular function in activation of a region.
>
> thank you in advance,
>
> Nader
>
> Russ Poldrack wrote:
>
> > I would think that it is quite reasonable to analyze the imaging data across
> > conditions that don't show any differences in RT or accuracy, particularly if
> > there are hypotheses about differential activation across the conditions in
> > question.  For one, it could be the case that there is some cognitive process
> > that varies across the conditions but the behavioral measure is not sensitive
> > enough to find this difference whereas the imaging measure might be more
> > sensitive.
> >
> > cheers,
> > russ
> >
> > Greig de Zubicaray wrote:
> >
> > > Dear all,
> > >
> > > I have a thorny question motivated by a reviewer's comments on one of our
> > > fMRI papers.
> > >
> > > Briefly, we used a memory task with variable delay periods. However, we did
> > > not find significant differences in terms of either accuracy or RT across
> > > the delays. Therefore, we did not look for an effect of delay in the fMRI
> > > data. The reviewer is requesting that we revise the paper to include this
> > > analysis.
> > >
> > > Does anyone know of any references addressing whether it's "good form" to
> > > perform analyses of neuroimaging data in the absence of a significant
> > > behavioural effect? Any views on this issue?
> > >
> > > Any help would be appreciated.
> > >
> > > Regards,
> > >
> > > Greig
> > >
> > > --
> > > Dr Greig de Zubicaray
> > > Centre for Magnetic Resonance
> > > The University of Queensland
> > > Brisbane, QLD 4072
> > > AUSTRALIA
> > >
> > > Tel: +61 (0) 7 3365 4250 [direct]
> > >      +61 (0) 7 3365 4100 [CMR]
> > > Fax: +61 (0) 7 3365 3833
> >
> > --
> > Russell A. Poldrack, Ph. D.
> > MGH-NMR Center
> > Building 149, 13th St.
> > Charlestown, MA 02129
> >
> > Phone: 617-726-4060
> > FAX:  617-726-7422
> > Email: [log in to unmask]
> > Web Page: http://www.nmr.mgh.harvard.edu/~poldrack

--
Russell A. Poldrack, Ph. D.
MGH-NMR Center
Building 149, 13th St.
Charlestown, MA 02129

Phone: 617-726-4060
FAX:  617-726-7422
Email: [log in to unmask]
Web Page: http://www.nmr.mgh.harvard.edu/~poldrack