Journal of Neuroscience,
Год журнала:
2018,
Номер
38(37), С. 7976 - 7985
Опубликована: Авг. 6, 2018
Calcium
imaging
is
a
powerful
method
to
record
the
activity
of
neural
populations
in
many
species,
but
inferring
spike
times
from
calcium
signals
challenging
problem.
We
compared
multiple
approaches
using
datasets
with
ground
truth
electrophysiology
and
found
that
simple
non-negative
deconvolution
(NND)
outperformed
all
other
algorithms
on
out-of-sample
test
data.
introduce
novel
benchmark
applicable
recordings
without
electrophysiological
truth,
based
correlation
responses
two
stimulus
repeats,
used
this
show
unconstrained
NND
also
when
run
“zoomed
out”
∼10,000
cell
visual
cortex
mice
either
sex.
Finally,
we
NND-based
methods
match
performance
supervised
convolutional
networks
while
avoiding
some
biases
such
methods,
at
much
faster
running
times.
therefore
recommend
spikes
be
inferred
traces
because
its
simplicity,
efficiency,
accuracy.
SIGNIFICANCE
STATEMENT
The
experimental
currently
allows
for
largest
numbers
cells
simultaneously
two-photon
imaging.
However,
use
requires
neuronal
firing
correctly
large
resulting
datasets.
Previous
studies
have
claimed
complex
learning
outperform
task.
Unfortunately,
these
suffered
several
problems
biases.
When
repeated
analysis,
same
data
correcting
problems,
simpler
inference
perform
better.
Even
more
importantly,
can
artifactual
structure
into
trains,
which
turn
lead
erroneous
scientific
conclusions.
Of
evaluated,
an
extremely
performed
best
circumstances
tested,
was
run,
insensitive
parameter
choices,
making
incorrect
conclusions
less
likely.
Measuring
the
dynamics
of
neural
processing
across
time
scales
requires
following
spiking
thousands
individual
neurons
over
milliseconds
and
months.
To
address
this
need,
we
introduce
Neuropixels
2.0
probe
together
with
newly
designed
analysis
algorithms.
The
has
more
than
5000
sites
is
miniaturized
to
facilitate
chronic
implants
in
small
mammals
recording
during
unrestrained
behavior.
High-quality
recordings
long
were
reliably
obtained
mice
rats
six
laboratories.
Improved
site
density
arrangement
combined
created
data
methods
enable
automatic
post
hoc
correction
for
brain
movements,
allowing
from
same
2
These
probes
algorithms
stable
free
behavior,
even
animals
such
as
mice.
Annual Review of Neuroscience,
Год журнала:
2020,
Номер
43(1), С. 249 - 275
Опубликована: Июль 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.
Japanese Journal of Applied Physics,
Год журнала:
2020,
Номер
59(6), С. 060501 - 060501
Опубликована: Апрель 27, 2020
Understanding
the
fundamental
relationships
between
physics
and
its
information-processing
capability
has
been
an
active
research
topic
for
many
years.
Physical
reservoir
computing
is
a
recently
introduced
framework
that
allows
one
to
exploit
complex
dynamics
of
physical
systems
as
devices.
This
particularly
suited
edge
devices,
in
which
information
processing
incorporated
at
(e.g.,
into
sensors)
decentralized
manner
reduce
adaptation
delay
caused
by
data
transmission
overhead.
paper
aims
illustrate
potentials
using
examples
from
soft
robotics
provide
concise
overview
focusing
on
basic
motivations
introducing
it,
stem
number
fields,
including
machine
learning,
nonlinear
dynamical
systems,
biological
science,
materials
physics.
Flexible
behaviors
over
long
timescales
are
thought
to
engage
recurrent
neural
networks
in
deep
brain
regions,
which
experimentally
challenging
study.
In
insects,
circuit
dynamics
a
region
called
the
central
complex
(CX)
enable
directed
locomotion,
sleep,
and
context-
experience-dependent
spatial
navigation.
We
describe
first
complete
electron
microscopy-based
connectome
of