2003 Abstracts
Barnes
Burke
Chawla
Ellmore
Euston
Kawahara
Moser
Olson
Pennartz
Penner
Plummer
Poneta
Ramirez-Amaya
Rosi
Towers
Twining
Vazdarjanova
Yang
2005 Abstracts
2004 Abstracts
|
LACK OF SEQUENCE DISAMBIGUATION IN RAT MEDIAL PREFRONTAL CORTEX ENSEMBLES
D.R. Euston*; R.P. Roop; B.L. McNaughton
NSMA, Univ Arizona, Tucson, AZ, USA
Learning sequences with repeated elements requires differential
neural activity to disambiguate sequential contexts of the
repeated elements, but where in the brain the critical distinctions
are made is unknown. Rat hippocampal cells can disambiguate
sequential context, but hippocampal disambiguation is not necessary
for task performance (Bower et al., 2002). Another candidate
for sequence disambiguation is the medial prefrontal cortex
(mPFC), a region implicated in working memory. A rat was trained
to run to a sequence of goal locations around the perimeter
of a 1.5m arena using medial forebrain bundle stimulation as
reward. The sequence, 8 segments long, contained two repeated
segments (e.g., 5-2-7-1-5-2-3). During cued trials, a flashing
LED indicated the target. In contrast, non-cued trials forced
the rat to locate the target without the LED (i.e., from memory).
Cued and non-cued runs alternated in blocks of 3 complete sequences.
The rat acquired non-cued sequences within 3, 90-minute sessions
and made few errors even at locations requiring a context-dependent
choice of direction. Typically, a single session included 120
traversals of the full sequence. 40-80 cells per session were
simultaneously recorded from either the dorsal anterior cingulate
(ACd) or the prelimbic cortex (PL). Of 222 cells, 48% showed
elevated firing specific to the run beginning, the run midpoint,
or the destination approach. The majority also showed differential
firing on different physical segments of the sequence, but
firing rates and spatial distributions were essentially identical
on all repeated segments, even though the turns following these
repeated segments differed (i.e., there was no predictive activity
within the mPFC indicative of sequence disambiguation). This
was equally true of both cued trials and memory driven, non-cued
trials. It can be concluded that neither the PL or ACd is the
locus of the short-term memory processes required for sequence
performance.
Support Contributed By: JST CREST, NS020331 & MH01565
|