PubMed,
Год журнала:
2023,
Номер
30(10), С. 271 - 277
Опубликована: Окт. 1, 2023
Historically,
the
development
of
valid
and
reliable
methods
for
assessing
higher-order
cognitive
abilities
(e.g.,
rule
learning
transfer)
has
been
difficult
in
rodent
models.
To
date,
limited
evidence
supports
existence
higher
such
as
generation
complex
decision-making
mice,
rats,
rabbits.
this
end,
we
sought
to
develop
a
task
that
would
require
mice
learn
transfer
rule.
We
trained
visually
discriminate
series
images
(image
set,
six
total)
increasing
complexity
following
three
stages:
(1)
visual
target,
(2)
(ignore
any
new
around
target),
finally
(3)
apply
abstract
form
comparable
but
image
set.
evaluate
each
stage,
measured
days
(and
performance
by
day)
original
target
at
criterion,
get
back
criterion
when
set
were
altered
introduction
distractors
(rule
learning),
overall
experienced
versus
naïve
cohorts
tested
on
same
transfer).
Twenty-seven
wild-type
male
C57
using
Bussey-Saksida
touchscreen
operant
conditioning
boxes
(Lafayette
Instruments).
Two
black-white
sets
delivered
sequentially
(counterbalanced
order)
two
identical
mice.
Results
showed
all
able
effectively
their
initial
could
recall
it
>80
d
later.
also
found
quickly
"rule"
:
Ignore
continue
identify
embedded
more
images.
The
presence
was
supported
because
thresholds
regained
much
faster
than
introduced.
On
other
hand,
appeared
unable
stimuli.
This
is
discrimination
curves
no
better
an
(naïve)
matched
cohort
Overall
results
have
important
implications
phenotyping
research
particularly
modeling
disorders
Science,
Год журнала:
2023,
Номер
382(6669), С. 394 - 398
Опубликована: Окт. 26, 2023
The
nervous
system
coordinates
various
motivated
behaviors
such
as
feeding,
drinking,
and
escape
to
promote
survival
evolutionary
fitness.
Although
the
precise
behavioral
repertoires
required
for
distinct
are
diverse,
common
features
approach
or
avoidance
suggest
that
brain
substrates
a
wide
range
of
behaviors.
In
this
Review,
I
describe
framework
by
which
neural
circuits
specified
some
innate
drives
regulate
activity
ventral
tegmental
area
(VTA)
dopamine
neurons
reinforce
ongoing
planned
actions
fulfill
motivational
demands.
This
may
explain
why
signaling
from
VTA
is
ubiquitously
involved
in
many
types
diverse
volitional
actions,
well
how
sensory
interoceptive
cues
can
initiate
specific
goal-directed
actions.
Nature Communications,
Год журнала:
2024,
Номер
15(1)
Опубликована: Март 8, 2024
Abstract
Behavior
can
be
remarkably
consistent,
even
over
extended
time
periods,
yet
whether
this
is
reflected
in
stable
or
‘drifting’
neuronal
responses
to
task
features
remains
controversial.
Here,
we
find
a
persistently
active
ensemble
of
neurons
the
medial
prefrontal
cortex
(mPFC)
mice
that
reliably
maintains
trajectory-specific
tuning
several
weeks
while
performing
an
olfaction-guided
spatial
memory
task.
This
task-specific
reference
frame
stabilized
during
learning,
upon
which
repeatedly
show
little
representational
drift
and
maintain
their
across
long
pauses
exposure
repeated
changes
cue-target
location
pairings.
These
data
thus
suggest
‘core
ensemble’
forming
task-relevant
space
for
performance
consistent
behavior
periods
time.
Nature Communications,
Год журнала:
2024,
Номер
15(1)
Опубликована: Июль 2, 2024
Olfaction
is
influenced
by
contextual
factors,
past
experiences,
and
the
animal's
internal
state.
Whether
this
information
integrated
at
initial
stages
of
cortical
odour
processing
not
known,
nor
how
these
signals
may
influence
encoding.
Here
we
revealed
multiple
diverse
non-olfactory
responses
in
primary
olfactory
(piriform)
cortex
(PCx),
which
dynamically
enhance
PCx
discrimination
according
to
behavioural
demands.
We
performed
recordings
neurons
from
mice
trained
a
virtual
reality
task
associate
odours
with
visual
contexts
obtain
reward.
found
that
learning
shifts
activity
encoding
solely
regime
positional,
contextual,
associative
emerge
on
odour-responsive
become
mixed-selective.
The
modulation
was
dynamic,
improving
decoding
during
engagement
rewarded
contexts.
This
improvement
relied
acquired
mixed-selectivity,
demonstrating
integrating
extra-sensory
inputs
sensory
cortices
can
while
relevance
stimuli.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Янв. 3, 2024
Abstract
Machine
learning
research
has
achieved
large
performance
gains
on
a
wide
range
of
tasks
by
expanding
the
target
from
mean
rewards
to
entire
probability
distributions
—
an
approach
known
as
distributional
reinforcement
(RL)
1
.
The
mesolimbic
dopamine
system
is
thought
underlie
RL
in
mammalian
brain
updating
representation
value
striatum
2,3
,
but
little
about
whether,
where,
and
how
neurons
this
circuit
encode
information
higher-order
moments
reward
4
To
fill
gap,
we
used
high-density
probes
(Neuropixels)
acutely
record
striatal
activity
well-trained,
water-restricted
mice
performing
classical
conditioning
task
which
mean,
variance,
stimulus
identity
were
independently
manipulated.
In
contrast
traditional
accounts,
found
robust
evidence
for
abstract
encoding
variance
striatum.
Remarkably,
chronic
ablation
inputs
disorganized
these
representations
without
interfering
with
coding.
Two-photon
calcium
imaging
optogenetics
revealed
that
two
major
classes
medium
spiny
D1
D2
MSNs
contributed
code
preferentially
right
left
tails
distribution,
respectively.
We
synthesize
findings
into
new
model
harnesses
opponency
between
5–15
reap
computational
benefits
RL.
PLoS Biology,
Год журнала:
2024,
Номер
22(3), С. e3002536 - e3002536
Опубликована: Март 1, 2024
Associating
values
to
environmental
cues
is
a
critical
aspect
of
learning
from
experiences,
allowing
animals
predict
and
maximise
future
rewards.
Value-related
signals
in
the
brain
were
once
considered
property
higher
sensory
regions,
but
their
wide
distribution
across
many
regions
increasingly
recognised.
Here,
we
investigate
how
reward-related
begin
be
incorporated,
mechanistically,
at
earliest
stage
olfactory
processing,
namely,
bulb.
In
head-fixed
mice
performing
Go/No-Go
discrimination
closely
related
mixtures,
rewarded
odours
evoke
widespread
inhibition
one
class
output
neurons,
that
is,
mitral
cells
not
tufted
cells.
The
temporal
characteristics
this
suggest
it
odour-driven,
also
context-dependent
since
absent
during
pseudo-conditioning
pharmacological
silencing
piriform
cortex.
Further,
modulation
present
somata
apical
dendritic
tuft
cells,
suggesting
an
involvement
circuit
components
located
deep
Depth-resolved
imaging
granule
cell
gemmules
suggests
target
receive
extrinsic
drive.
Thus,
our
study
supports
notion
value-related
characteristic
processing
primary
area
narrows
down
possible
underlying
mechanisms
deeper
contact
perisomatically.
SignificanceMicroprism-mediated
calcium
imaging
of
deep
brain
structures
allows
for
reliable
tracking
thousands
cells
across
days,
a
marked
improvement
compared
with
industry-standard
approaches
utilizing
gradient-index
(GRIN)
lenses.AimWe
aim
to
develop
method
record,
track,
and
functionally
characterize
cortical
subcortical
neurons
using
an
endoscopic
microprism
benchmark
the
performance
against
traditional
GRIN
lenses.ApproachTo
improve
visualization
throughput
ensembles,
we
developed
characterized
protocol
implant
microprism.
We
performed
two-photon
through
microprisms
yield,
cell
quality,
optical
characteristics
lenses.ResultsWe
found
that
our
can
stably
isolate
track
active
days
weeks.
This
high-throughput
approach
facilitates
longitudinal
functional
characterization
large
groups
neural
ensembles
in
vivo
structures.
produces
high
fidelity
trackable
superior
resolution
quality
higher
spatial
precision
when
lenses.ConclusionsThe
technique
represents
significant
neuronal
ensemble
dynamics
within
tissues.