Working memory: all over the place?

F. Frank LeFever flefever at
Mon Mar 23 22:53:11 EST 1998

re Mishkin: see my post headed "Working Memory & LTM" or something
like that (I did not store it in LTM).  He is not part of the
"temporal" crowd but rather a pioneer who helped develop the concept of
TWO memory systems, one based on the hippocampus and the other on the
basal ganglia, BOTH of which can be said to underly longterm

There are always those who prefer very narrow operational definitions
to broader and more complex ones.  Indeed, they gave "operational
definition" a bad rep many years ago because in their hands this
smacked of reductionism or logical positivism and the like because they
thereby narrowly re-defined words which in common use had a broader

I preferred the model of a "theoretical construct" that Neal Miller
described: one SAMPLED from a population of independent variables and a
population of dependent variables to develop a system of relationships
which defined the theoretical construct.  (e.g. not just length of
deprivation OR just salt-loading as manipulations and not just rate of
running to water OR tolerance to quinine in the water as dependent
measures to define "thirst", but interrelations among these: IV # 1
with DV #1 but also IV #2 and DV #1, as well as IV #1 and DV #2, AND of
course IV #2 and DV #2).

The disparate operations you cite may yet lack this kind of coordinated
analysis, but the potential is there.  To the extent that many
different working memory operational definitions include dependence on
prefrontal functions, some coordinaion has already been achieved.

The importance of the more complex definition and of efforts to develop
it can be appreciated only by those aware of the clinical importance of
identifying and understanding aspects of working memory and its
failures in tasks better studied in humans and perhaps other primates)
than in rats and both more complex and more important than "delayed
nonmatching to sample".  (Very concrete example: People who would have
no problem with delayed matching or nonmatching to sample may have
enormous difficulty with, my favorite example, Trails B; and they can
tell you perfectly what they SHOULD have done.  I know, having tested a
quite a few such patients...)

Frank LeFever

In <6f70gd$1i8 at> mtaffe at (Mike Taffe) writes: 
>In article <6f4hup$rv9 at>,
>F. Frank LeFever <flefever at> wrote:
>>(bytheway: has no one besides me read anything but rat literature or
>>anything since c. 1988? I think even in the rat literature they have
>>profound importance of the prefrontal cortex of higher primates,
>>discussing "working memory" without a good knowledge of
>>primate--including human--neuropsychology is not playing with a full
>I don't think that this has much to do with the problem at hand which
would seem to lie in
>operationalization of the concept "working memory".  G-R and others
have done a nice job of showing
>involvement of the pre-frontal areas in memory.  Mishkin, Gaffan, Z-M
and their respective
>colleagues have done an elegant job of showing involvement of the
temporal structures in memory.
>Each and every paper has included at least one construct of "memory"
to explain the pattern of
>deficits exhibited by the subject primates.  The difference between
the two is that at least in the
>case of the temporal lobe memory wars, the participants have more or
less utilized the same task so
>that when they talk about the construct (episodic memory, short-term
recognition memory or what have
>you) you can be pretty much assured that they're talking about delayed
(non) matching to sample.  In
>contrast, the "working memory" crowd is all over the map.  spatial
delayed response tasks (grids a
>la bartus or voytko, Hamilton search (levin and bowman)), spatial
search (cantab folks), and now
>I've seen DNMS and delayed match to position represented as "working
memory" tasks.  Well, to my eye
>the requirements of all these tasks is varied to the extent that it is
difficult to
>get a handle on what the core concept and therefore target processes
ARE in the so-called "working"
>memory studies.  Don't get me wrong, I think that accurate performance
in traditional DNMS requires
>some different processes and presumably structures than say, delayed
spatial response to a limited
>matrix BUT this difference is blurred by the indiscretion with which
the term working memory is used
>in the literature.  so I now find it to be a useless term for
describing the difference that I might
>once have thought of as "working" aspects of memory.
>Michael Taffe, Ph.D        *  mtaffe at         
>Dept. of Neuropharmacology, The Scripps Research Institute
>(619) 784-7247		   *

More information about the Neur-sci mailing list