Nature Methods,
Год журнала:
2022,
Номер
20(2), С. 295 - 303
Опубликована: Дек. 30, 2022
Abstract
We
present
an
auxiliary
learning
task
for
the
problem
of
neuron
segmentation
in
electron
microscopy
volumes.
The
consists
prediction
local
shape
descriptors
(LSDs),
which
we
combine
with
conventional
voxel-wise
direct
neighbor
affinities
boundary
detection.
capture
statistics
about
to
be
segmented,
such
as
diameter,
elongation,
and
direction.
On
a
study
comparing
several
existing
methods
across
various
specimen,
imaging
techniques,
resolutions,
LSDs
consistently
increases
accuracy
affinity-based
over
range
metrics.
Furthermore,
addition
promotes
on
par
current
state
art
(flood-filling
networks),
while
being
two
orders
magnitudes
more
efficient—a
critical
requirement
processing
future
petabyte-sized
datasets.
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
Cell,
Год журнала:
2024,
Номер
187(10), С. 2574 - 2594.e23
Опубликована: Май 1, 2024
High-resolution
electron
microscopy
of
nervous
systems
has
enabled
the
reconstruction
synaptic
connectomes.
However,
we
do
not
know
sign
for
each
connection
(i.e.,
whether
a
is
excitatory
or
inhibitory),
which
implied
by
released
transmitter.
We
demonstrate
that
artificial
neural
networks
can
predict
transmitter
types
presynapses
from
micrographs:
network
trained
to
six
transmitters
(acetylcholine,
glutamate,
GABA,
serotonin,
dopamine,
octopamine)
achieves
an
accuracy
87%
individual
synapses,
94%
neurons,
and
91%
known
cell
across
D.
melanogaster
whole
brain.
visualize
ultrastructural
features
used
prediction,
discovering
subtle
but
significant
differences
between
phenotypes.
also
analyze
distributions
brain
find
neurons
develop
together
largely
express
only
one
fast-acting
GABA).
hope
our
publicly
available
predictions
act
as
accelerant
neuroscientific
hypothesis
generation
fly.
Neuron,
Год журнала:
2024,
Номер
112(15), С. 2581 - 2599.e23
Опубликована: Май 24, 2024
Anchoring
goals
to
spatial
representations
enables
flexible
navigation
but
is
challenging
in
novel
environments
when
both
must
be
acquired
simultaneously.
We
propose
a
framework
for
how
Drosophila
uses
internal
of
head
direction
(HD)
build
goal
upon
selective
thermal
reinforcement.
show
that
flies
use
stochastically
generated
fixations
and
directed
saccades
express
heading
preferences
an
operant
visual
learning
paradigm
HD
neurons
are
required
modify
these
based
on
used
symmetric
setting
expose
flies'
co-evolve
the
reliability
interacting
impacts
behavior.
Finally,
we
describe
rapid
new
headings
may
rest
behavioral
policy
whose
parameters
form
genetically
encoded
circuit
architecture.
Such
evolutionarily
structured
architectures,
which
enable
rapidly
adaptive
behavior
driven
by
representations,
relevant
across
species.
In
both
invertebrates
such
as
Drosophila
and
vertebrates
mouse
or
human,
the
brain
contains
most
diverse
population
of
cell
types
any
tissue.
It
is
generally
accepted
that
transcriptional
diversity
an
early
step
in
generating
neuronal
glial
diversity,
followed
by
establishment
a
unique
gene
expression
profile
determines
morphology,
connectivity,
function.
,
there
are
two
neural
stem
cells,
called
Type
1
(T1)
2
(T2)
neuroblasts.
contrast
to
T1
neuroblasts,
T2
neuroblasts
generate
intermediate
progenitors
(INPs)
expand
number
types.
The
T2-derived
neurons
contributes
large
portion
central
complex
(CX),
conserved
region
plays
role
sensorimotor
integration.
Recent
work
has
revealed
much
connectome
CX,
but
how
this
assembled
remains
unclear.
Mapping
derived
from
necessary
linking
assembly
adult
brain.
Here
we
perform
single
nuclei
RNA
sequencing
neuroblast-derived
glia.
We
identify
clusters
containing
all
known
classes
glia,
male/female
enriched,
161
neuron-specific
clusters.
map
neurotransmitter
neuropeptide
transcription
factor
combinatorial
codes
for
each
cluster
(presumptive
neuron
subtype).
This
directs
functional
studies
determine
whether
code
specifies
distinct
type
within
CX.
several
columnar
subtypes
(NPF+
AstA+)
closely
related
Our
data
support
hypothesis
represents
one
few
subtypes.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2020,
Номер
unknown
Опубликована: Июнь 13, 2020
Abstract
High-resolution
electron
microscopy
of
nervous
systems
enables
the
reconstruction
connectomes.
A
key
piece
missing
information
from
connectomes
is
synaptic
sign.
We
show
that
for
D.
melanogaster
,
artificial
neural
networks
can
predict
transmitter
type
released
at
synapses
micrographs
and
thus
add
putative
signs
to
connections.
Our
network
discriminates
between
six
transmitters
(acetylcholine,
glutamate,
GABA,
serotonin,
dopamine,
octopamine)
with
an
average
accuracy
87%/94%
synapses/entire
neurons.
developed
explainability
method
reveal
which
features
our
using
found
significant
ultrastructural
differences
classical
transmitters.
in
two
characterize
morphological
connection
properties
tens
thousands
neurons
classed
by
predicted
expression.
find
hemilineages
largely
express
only
one
fastacting
among
their
Furthermore,
we
different
may
differ
like
polarization
projection
targets.
