PLoS Biology,
Journal Year:
2022,
Volume and Issue:
20(12), P. e3001861 - e3001861
Published: Dec. 15, 2022
Recent
theoretical
models
suggest
that
deciding
about
actions
and
executing
them
are
not
implemented
by
completely
distinct
neural
mechanisms
but
instead
two
modes
of
an
integrated
dynamical
system.
Here,
we
investigate
this
proposal
examining
how
activity
unfolds
during
a
dynamic
decision-making
task
within
the
high-dimensional
space
defined
cells
in
monkey
dorsal
premotor
(PMd),
primary
motor
(M1),
dorsolateral
prefrontal
cortex
(dlPFC)
as
well
external
internal
segments
globus
pallidus
(GPe,
GPi).
Dimensionality
reduction
shows
four
strongest
components
functionally
interpretable,
reflecting
state
transition
between
deliberation
commitment,
transformation
sensory
evidence
into
choice,
baseline
slope
rising
urgency
to
decide.
Analysis
contribution
each
population
these
meaningful
differences
regions
no
clusters
region,
consistent
with
During
deliberation,
cortical
on
two-dimensional
“decision
manifold”
falls
off
manifold
at
moment
commitment
choice-dependent
trajectory
leading
movement
initiation.
The
structure
varies
regions:
In
PMd,
it
is
curved;
M1,
nearly
perfectly
flat;
dlPFC,
almost
entirely
confined
dimension.
contrast,
pallidal
primarily
urgency.
We
findings
reveal
functional
contributions
different
brain
system
governing
action
selection
execution.
Annual Review of Neuroscience,
Journal Year:
2020,
Volume and Issue:
43(1), P. 249 - 275
Published: July 8, 2020
Significant
experimental,
computational,
and
theoretical
work
has
identified
rich
structure
within
the
coordinated
activity
of
interconnected
neural
populations.
An
emerging
challenge
now
is
to
uncover
nature
associated
computations,
how
they
are
implemented,
what
role
play
in
driving
behavior.
We
term
this
computation
through
population
dynamics.
If
successful,
framework
will
reveal
general
motifs
quantitatively
describe
dynamics
implement
computations
necessary
for
goal-directed
Here,
we
start
with
a
mathematical
primer
on
dynamical
systems
theory
analytical
tools
apply
perspective
experimental
data.
Next,
highlight
some
recent
discoveries
resulting
from
successful
application
systems.
focus
studies
spanning
motor
control,
timing,
decision-making,
working
memory.
Finally,
briefly
discuss
promising
lines
investigation
future
directions
framework.
Current Opinion in Neurobiology,
Journal Year:
2021,
Volume and Issue:
70, P. 137 - 144
Published: Oct. 1, 2021
Advances
in
experimental
neuroscience
have
transformed
our
ability
to
explore
the
structure
and
function
of
neural
circuits.
At
same
time,
advances
machine
learning
unleashed
remarkable
computational
power
artificial
networks
(ANNs).
While
these
two
fields
different
tools
applications,
they
present
a
similar
challenge:
namely,
understanding
how
information
is
embedded
processed
through
high-dimensional
representations
solve
complex
tasks.
One
approach
addressing
this
challenge
utilize
mathematical
analyze
geometry
representations,
i.e.,
population
geometry.
We
review
examples
geometrical
approaches
providing
insight
into
biological
networks:
representation
untangling
perception,
geometric
theory
classification
capacity,
disentanglement
abstraction
cognitive
systems,
topological
underlying
maps,
dynamic
motor
dynamical
cognition.
Together,
findings
illustrate
an
exciting
trend
at
intersection
learning,
neuroscience,
geometry,
which
provides
useful
population-level
mechanistic
descriptor
task
implementation.
Importantly,
descriptions
are
applicable
across
sensory
modalities,
brain
regions,
network
architectures
timescales.
Thus,
has
potential
unify
networks,
bridging
gap
between
single
neurons,
populations
behavior.
Neuron,
Journal Year:
2022,
Volume and Issue:
110(7), P. 1258 - 1270.e11
Published: Jan. 31, 2022
How
do
neural
populations
code
for
multiple,
potentially
conflicting
tasks?
Here
we
used
computational
simulations
involving
networks
to
define
"lazy"
and
"rich"
coding
solutions
this
context-dependent
decision-making
problem,
which
trade
off
learning
speed
robustness.
During
lazy
the
input
dimensionality
is
expanded
by
random
projections
network
hidden
layer,
whereas
in
rich
units
acquire
structured
representations
that
privilege
relevant
over
irrelevant
features.
For
decision-making,
one
solution
project
task
onto
low-dimensional
orthogonal
manifolds.
Using
behavioral
testing
neuroimaging
humans
analysis
of
signals
from
macaque
prefrontal
cortex,
report
evidence
patterns
biological
brains
whose
geometry
are
consistent
with
regime.
