Matthew, Daniel, and all:
>I know this is going to seem
very tedious to everyone, but the
> >thresholded map does not address my assertion 1, and therefore
doesn't
> >address localization. I
>
After reading everyone's comments and giving this further consideration,
I've come to a dissenting(?) conclusion about localization. While
neuroimaging does allow reasonable localization inferences to be made, it
really does not address functional localization at all (at least not in a
strict sense).
In my collegiate days, I was taught that "sufficient and
necessary" was the gold standard to demonstrate
functionality. For example, to demonstrate a role of islets of
Langerhans in the pancreas in regulating glucose levels in the blood, it
was necessary to show both that their secretion of insulin was
sufficient to regulate glucose blood levels, and
necessary. Neuroimaging does not meet either of these
criteria; activation maps (whether thresholded or not) only
indicate that the activity of an area is better related to one type of
behavior than another. Demonstration that an area is responsible
for a function would require additional data from other types of studies,
such as those involving lesions, TMS, cryogenic inactivation, single-cell
recordings, etc. (Correlations with behavior allow stronger
inferences about functional roles since correlations indicate which areas
tend to be more strongly activated with better performance, but even
these do not meet the "necessary and sufficient"
criteria.)
This is not a trivial point. The entire brain may be involved to
varying degrees in most behaviors, a conclusion consistent with Karl
Lashley's studies in the 1950s as he searched for the locus of the engram
within the brain. Removal of any one part of the brain did not
eliminate recall in his memory tasks, but removal of larger portions of
the brain generally produced larger memory deficits. Nonetheless,
not all parts of the brain are equivalent for performing a task, and
neuroimaging does allow us to infer that certain parts of the brain are
more actively involved in certain functions (or processes) than
others.
Thus from neuroimaging studies we may infer that motor / premotor
cortices are more involved in motor planning than visual cortices, but
this does not allow us to infer that visual cortices are not involved in
motor planning behaviors. Indeed, movements are often guided by our
vision, and since recall of visual images have been shown to activate
various parts of visual cortices, it is quite possible that visual
cortices are also involved (albeit indirectly) in motor planning.
The fact that motor / premotor cortices are activated in studies
involving motor planning allows us to infer that the activity of these
areas is involved in computations relevant to motor planning, whereas
other neuroimaging studies allow us to infer that visual cortices are
normally involved in computations relevant to various aspects of visual
processing. The neuroimaging studies don't preclude other
functions, as demonstrated by activation of visual cortices by
somatosensory stimuli in the congenitally blind; nor do
neuroimaging studies indicate that motor / premotor cortices are the only
(or even primary) source of motor function, as evident from clinical
cases with stroke damage restricted to the pyramidal system.
So neuroimaging does not localize function. What it does do is
indicate that the activity of some areas are more involved in some types
of activities or functions than others, which allows inferences to be
made. As such, unthresholded maps represent a sort of probability
map of preferential activity, whereas a thresholded map specifies the
region where we can reliably expect such preferential activity. The
"better" map depends on what point you want to make, as
addressed better (and more succinctly!) by others in this
thread.
Doug Burman
Dept. of Communication Sciences & Disorders
Northwestern University
2240 Campus Drive
Frances Searle Building, Room 2-356
Evanston, IL 60208
phone 847-467-1549
fax 847-491-4975
email: [log in to unmask]