Psychological Review,
Journal Year:
2023,
Volume and Issue:
131(2), P. 349 - 372
Published: Sept. 5, 2023
When
faced
with
distraction,
we
can
focus
more
on
goal-relevant
information
(targets)
or
less
goal-conflicting
(distractors).How
people
use
cognitive
control
to
distribute
attention
across
targets
and
distractors
remains
unclear.We
address
this
question
by
developing
a
novel
Parametric
Attentional
Control
Task
(PACT)
that
'tag'
participants'
sensitivity
target
distractor
information.We
these
precise
measures
of
develop
process
model
explain
how
participant
towards
distractors.Across
three
experiments,
find
participants
met
the
demands
task
independently
controlling
their
processing
information,
exhibiting
distinct
adaptations
manipulations
incentives
conflict.Whereas
preferentially
led
enhancement,
conflict
previous
trial
suppression.These
drivers
altered
early
in
trial,
promptly
followed
reactive
reconfiguration
task-appropriate
feature
sensitivity.To
provide
process-level
account
empirical
findings,
neural
network
evidence
accumulation
attractor
dynamics
over
weights
reconfigures
processing.These
results
computational
provides
new
insights
into
multivariate
attentional
signals
are
optimized
achieve
goals.
Cell Reports,
Journal Year:
2021,
Volume and Issue:
35(9), P. 109090 - 109090
Published: June 1, 2021
The
neural
mechanisms
that
generate
an
extensible
library
of
motor
motifs
and
flexibly
string
them
into
arbitrary
sequences
are
unclear.
We
developed
a
model
in
which
inhibitory
basal
ganglia
output
neurons
project
to
thalamic
units
themselves
bidirectionally
connected
recurrent
cortical
network.
the
patterns
as
silencing
some
while
leaving
others
disinhibited
free
interact
with
cortex
during
specific
motifs.
show
small
number
can
control
dynamics
noise-robust
way.
Additionally,
single
"preparatory"
thalamocortical
network
produce
fast
support
rapid
transitions
between
any
pair
learned
If
associated
each
sequence
component
segregated,
many
outputs
be
without
interference
then
combined
orders
for
flexible
production
long
complex
sequences.
Deleted Journal,
Journal Year:
2023,
Volume and Issue:
20(1), P. 1 - 18
Published: Jan. 10, 2023
Abstract
Traditional
joint-link
robots
have
been
widely
used
in
production
lines
because
of
their
high
precision
for
single
tasks.
With
the
development
manufacturing
and
service
industries,
requirement
comprehensive
performance
robotics
is
growing.
Numerous
types
bio-inspired
investigated
to
realize
human-like
motion
control
manipulation.
A
study
route
from
inner
mechanisms
external
structures
proposed
imitate
humans
animals
better.
this
idea,
a
brain-inspired
intelligent
robotic
system
constructed
that
contains
visual
cognition,
decision-making,
control,
musculoskeletal
structures.
This
paper
reviews
cutting-edge
research
systems.
Two
software
systems
corresponding
hardware
are
established,
aiming
at
verification
applications
next-generation
robots.
Annual Review of Neuroscience,
Journal Year:
2023,
Volume and Issue:
46(1), P. 233 - 258
Published: March 27, 2023
Flexible
behavior
requires
the
creation,
updating,
and
expression
of
memories
to
depend
on
context.
While
neural
underpinnings
each
these
processes
have
been
intensively
studied,
recent
advances
in
computational
modeling
revealed
a
key
challenge
context-dependent
learning
that
had
largely
ignored
previously:
Under
naturalistic
conditions,
context
is
typically
uncertain,
necessitating
contextual
inference.
We
review
theoretical
approach
formalizing
face
uncertainty
core
computations
it
requires.
show
how
this
begins
organize
large
body
disparate
experimental
observations,
from
multiple
levels
brain
organization
(including
circuits,
systems,
behavior)
regions
(most
prominently
prefrontal
cortex,
hippocampus,
motor
cortices),
into
coherent
framework.
argue
inference
may
also
be
understanding
continual
brain.
This
theory-driven
perspective
places
as
component
learning.
Trends in Cognitive Sciences,
Journal Year:
2024,
Volume and Issue:
28(7), P. 614 - 627
Published: April 4, 2024
Working
memory
(WM)
is
a
fundamental
aspect
of
cognition.
WM
maintenance
classically
thought
to
rely
on
stable
patterns
neural
activities.
However,
recent
evidence
shows
that
population
activities
during
undergo
dynamic
variations
before
settling
into
pattern.
Although
this
has
been
difficult
explain
theoretically,
network
models
optimized
for
typically
also
exhibit
such
dynamics.
Here,
we
examine
versus
coding
in
data,
classical
models,
and
task-optimized
networks.
We
review
principled
mathematical
reasons
why
do
not,
while
naturally
coding.
suggest
an
update
our
understanding
maintenance,
which
computational
feature
rather
than
epiphenomenon.
Recent
studies
have
identified
rotational
dynamics
in
motor
cortex
(MC),
which
many
assume
arise
from
intrinsic
connections
MC.
However,
behavioral
and
neurophysiological
suggest
that
MC
behaves
like
a
feedback
controller
where
continuous
sensory
interactions
with
other
brain
areas
contribute
substantially
to
processing.
We
investigated
these
apparently
conflicting
theories
by
building
recurrent
neural
networks
controlled
model
arm
received
the
limb.
Networks
were
trained
counteract
perturbations
limb
reach
toward
spatial
targets.
Network
activities
signals
network
exhibited
structure
even
when
removed.
Furthermore,
recordings
monkeys
performing
similar
tasks
also
not
only
but
somatosensory
cortex.
Our
results
argue
may
reflect
throughout
voluntary
system
involved
online
control
of
actions.
Learned
movements
can
be
skillfully
performed
at
different
paces.
What
neural
strategies
produce
this
flexibility?
Can
they
predicted
and
understood
by
network
modeling?
We
trained
monkeys
to
perform
a
cycling
task
speeds,
artificial
recurrent
networks
generate
the
empirical
muscle-activity
patterns.
Network
solutions
reflected
principle
that
smooth
well-behaved
dynamics
require
low
trajectory
tangling.
had
consistent
form,
which
yielded
quantitative
qualitative
predictions.
To
evaluate
predictions,
we
analyzed
motor
cortex
activity
recorded
during
same
task.
Responses
supported
hypothesis
dominant
signals
reflect
not
muscle
activity,
but
network-level
for
generating
activity.
Single-neuron
responses
were
better
accounted
than
Similarly,
population
trajectories
shared
their
organization
with
trajectories,
solutions.
Thus,
cortical
could
based
on
need
via
allow
smooth,
robust
control
over
movement
speed.
