Working memory and LTM

F. Frank LeFever flefever at
Mon Mar 23 22:18:30 EST 1998

Yes, I think you are quite right in thinking that new items in working
memory have a potential for becoming long-term memories; but it may
also be that there is a process working to PREVENT this in some (most?)
cases.  For many tasks, it would be disadvantageous to have one's
memory cluttered up with  old items which might interfere with new
items (interference being one of the mechanisms of forgetting, in old,
old memory theories).

There is also the sense of "working WITH memory", which 
Fuster's demonstations of impairing specific aspects of task
performances by cooling EITHER a region of parietal cortex OR its
corresponding prefrontal area seem to examplify: i.e., briefly
retrieving and using items which are ALREADY in long-term memory.

Nonetheless, the possibility of dual paths for these items remains.
I am reminded of a conversation with Mishkin a few years ago, after one
of his presntations on "two memory sysems"--i.e., the hippocampal
system (perhaps "declarative" or "episodic"  or--choose your favorite
term) and the "basal ganglial" (or "procedural" or "habit" or--?)
system.  He pointed out that although it is important to try to
separate these out by anatomical and procedural "dissections", it is
likely that both are involved to some extent, concurrently, in every
"real-life" example of learning.

Which in terms is reminiscent of the idea (developed by--Rescorla? 
many years ago) that such real-life examples always involved both
classical (so-called "Pavlovian") and instrumental (so-called
"Skinnerian"--or perhaps "Thorndikean"?) conditioning...

re HM: I hesitated to mention him, embarassed that I did not recall
details of his case well enough to use him confidently as an example,
but inasmuch as you have suggested it--yes, I concur, he probably had
fair preservation of working memory.  I believe he was reported (have
not had a chance to refresh my memory) to have done fairly well on the
WAIS (probably Wechsler-Bellevue initially), some subtests of which
have "working memory" demands.

Indeed, according to some analyses, reading or comprehending oral
presentations of complex sentences requires working memory.  A very
early example, by Lashley I believe was: "Righting with the left hand,
the man--" (I forget the complete sentence); only by holding the SOUND
of the first word in working memory until one hears the word "canoe"
can one interpret that sound as "righting" instead of "writing" (one's
likely first "working" assumption).

F. LeFever
New York Neuropsychology Group

In <35167CC8.4FA054F3 at> marcello spinella
<optimism32 at> writes: 
>It  makes sense that something held in working memory will be encoded
into LTM to
>some degree as well, since it probably is salient. My only question to
this would
>be: can someone without a hippocampus (e.g. H.M.) still perform
working memory
>tasks?  I think the answer is yes, he can hold information on-line for
the moment.
>Unless he continues to rehearse it or is distrated, the info can be
maintained, but
>shortly after it rapidly decays.  Perhaps the hippocampal role in WM
is to carry
>the contents of WM in to LTM.
>Just a hypothesis
>Uwilwantme wrote:
>> Have we forgotten that you cannot completely localize function to
ONE area of
>> the brain?
>>  Also important... the intricate connections of the hippomcampus AND
>> prefrontal cortex.  If you block NMDA receptors of the hippocampus,
you disrupt
>> working memory (yes, Frank, in an animal model...but think about PCP
>> Highfield, Nixon, & Amsel, 1996; Behavioral Neuroscience).   NMDA
receptors are
>> involved in a putative molecular model of learning and memory, LTP. 
>> when you interfere with LTP, you interfere with memory. 
Interestingly, LTP
>> occurs primarily in the hippocampus and most is NMDA receptor
dependent.   I
>> don't think you can rule out the involvement of the hippocampus with
>> molecular  information out there.

More information about the Neur-sci mailing list