BMC Bioinformatics,
Год журнала:
2024,
Номер
25(1)
Опубликована: Дек. 18, 2024
Recent
developments
in
spatially
resolved
transcriptomics
(SRT)
enable
the
characterization
of
spatial
structures
for
different
tissues.
Many
decomposition
methods
have
been
proposed
to
depict
cellular
distribution
within
However,
existing
computational
struggle
balance
continuity
cell
with
preservation
cell-specific
characteristics.
To
address
this,
we
propose
Spall,
a
novel
network
that
integrates
scRNA-seq
data
SRT
accurately
infer
type
proportions.
Spall
introduced
GATv2
module,
featuring
flexible
dynamic
attention
mechanism
capture
relationships
between
spots.
This
improves
identification
patterns
analysis.
Additionally,
incorporates
skip
connections
loss
information,
thereby
enhancing
prediction
capability
rare
types.
Experimental
results
show
outperforms
state-of-the-art
reconstructing
on
multiple
datasets.
Notably,
reveals
tumor
heterogeneity
human
pancreatic
ductal
adenocarcinoma
samples
and
delineates
complex
tissue
structures,
such
as
laminar
organization
mouse
cerebral
cortex
cerebellum.
These
findings
highlight
ability
provide
reliable
low-dimensional
embeddings
downstream
analyses,
offering
new
opportunities
deciphering
structures.
The Journal of Physiology,
Год журнала:
2025,
Номер
unknown
Опубликована: Янв. 26, 2025
Abstract
The
cerebellum
is
critical
for
coordinating
movements
related
to
eating,
drinking
and
swallowing,
all
of
which
require
proper
control
the
tongue.
Cerebellar
Purkinje
cells
can
encode
tongue
movements,
but
it
unclear
how
their
simple
spikes
complex
induce
changes
in
shape
that
contribute
goal‐directed
movements.
To
study
these
relations,
we
recorded
stimulated
vermis
hemispheres
mice
during
spontaneous
licking
from
a
stationary
or
moving
water
spout.
We
found
rhythmic
with
both
spikes.
Increased
spike
firing
protrusion
induces
ipsiversive
bending
Unexpected
target
location
trigger
alter
subsequent
licks,
adjusting
trajectory.
Furthermore,
observed
increased
behavioural
state
at
start
end
bouts.
Using
machine
learning,
confirmed
alterations
cell
activity
accompany
licking,
different
often
exerting
heterogeneous
encoding
schemes.
Our
data
highlight
directional
movement
paramount
cerebellar
function
modulation
are
complementary
acquisition
execution
sensorimotor
coordination.
These
results
bring
us
closer
understanding
clinical
implications
disorders
swallowing.
image
Key
points
When
drinking,
make
directed
towards
source.
fire
rhythmically
tune
position
source
affects
direction
report
also
adjust
right
direction.
Purkinje
cell
(PC)
synapses
onto
cerebellar
nuclei
(CbN)
neurons
allow
signals
from
the
cortex
to
influence
rest
of
brain.
PCs
are
inhibitory
that
spontaneously
fire
at
high
rates,
and
many
PC
inputs
thought
converge
each
CbN
neuron
suppress
its
firing.
It
has
been
proposed
convey
information
using
a
rate
code,
synchrony
timing
or
both.
The
on
firing
was
primarily
examined
for
combined
effects
with
comparable
strengths,
individual
not
extensively
studied.
Here,
we
find
single
highly
variable
in
size,
dynamic
clamp
modeling
reveal
this
important
implications
PC-CbN
transmission.
Individual
regulate
both
Large
strongly
rates
transiently
eliminate
several
milliseconds.
Remarkably,
refractory
period
leads
brief
elevation
prior
suppression.
Thus,
suited
concurrently
codes
generate
precisely
timed
responses
neurons.
Either
synchronous
pauses
promote
rapid
time
scales
nonuniform
inputs,
but
less
effectively
than
uniform
inputs.
This
is
secondary
consequence
input
sizes
elevating
baseline
by
increasing
variability
conductance.
These
findings
may
generalize
other
brain
regions
synapse
sizes.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2025,
Номер
unknown
Опубликована: Янв. 29, 2025
Abstract
Identification
of
neuron
type
is
critical
to
understand
computation
in
neural
circuits
through
extracellular
recordings
awake,
behaving
animal
subjects.
Yet,
modern
recording
probes
have
limited
power
resolve
type.
Here,
we
leverage
the
well-characterized
architecture
cerebellar
circuit
perform
expert
identification
from
non-human
primates.
Using
deep-learning
classifiers
evaluate
information
contained
readily
accessible
features
for
identification.
Waveform,
discharge
statistics,
anatomical
layer,
and
functional
interactions
each
can
inform
labels
a
sizable
fraction
units.
Together,
as
inputs
classifier,
even
better.
Our
tools
methodologies,
validated
during
smooth
pursuit
eye
movements
floccular
complex
awake
monkeys,
guide
cerebellar-dependent
tasks
animals
across
species.
They
lay
groundwork
characterization
processing
cortex.
Impact
statement
To
how
brain
performs
computations
service
behavior,
develop
methods
link
activity
within
circuits.
show
derived
provide
complementary
disambiguate
identity
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Ноя. 14, 2024
When
a
neuron
modulates
its
firing
rate
during
movement,
we
tend
to
assume
that
it
is
contributing
control
of
movement.
However,
null
space
theory
makes
the
counter-intuitive
prediction
neurons
often
generate
spikes
not
cause
behavior,
but
prevent
effects
other
would
have
on
behavior.
What
missing
direct
way
test
this
in
brain.
Here,
found
marmoset
cerebellum,
spike-triggered
averaging
identified
potent
vector
unique
each
Purkinje
cell
(P-cell)
along
which
displaced
eyes.
Yet,
P-cells
were
active
just
for
saccades
their
vector,
all
saccades.
Simultaneous
recordings
revealed
two
different
produced
superposition
vectors.
The
result
was
population
activity
canceled
if
contributions
perpendicular
intended
But
why
one
producing
be
by
another
neuron?
To
answer
question,
recorded
from
mossy
fiber
inputs
and
interneurons
molecular
layer.
We
fibers
provided
copy
motor
commands
as
well
sensory
goal
then
transformed
these
so
predicted
when
movement
had
reached
should
stopped.
Because
output
faster
dynamics
than
present
individual
neurons,
cerebellum
placed
cells
subtractive
competition
with
such
downstream
spiking
partially
or
completely
eliminated
because
neuron.
Purkinje
cell
(PC)
dendrites
are
optimized
to
integrate
the
vast
cerebellar
input
array
and
drive
sole
cortical
output.
PCs
classically
seen
as
stereotypical
computational
units,
yet
mouse
morphologically
diverse
those
with
multi-branched
structure
can
receive
non-canonical
climbing
fiber
(CF)
multi-innervation
that
confers
independent
compartment-specific
signaling.
While
otherwise
uncharacterized,
human
universally
multi-branched.
Do
they
exceed
allometry
achieve
enhanced
integrative
capacities
relative
PCs?
To
answer
this,
we
used
several
comparative
histology
techniques
in
adult
analyze
cellular
morphology,
parallel
(PF)
CF
arrangement,
regional
PC
demographics.
Human
substantially
larger
than
previously
described;
allometric
constraint
by
thickness
largest
neuron
brain
6-7cm
total
dendritic
length.
Unlike
mouse,
ramify
horizontally
form
a
multi-compartment
motif
show
multiple
CFs.
spines
denser
(6.9
vs
4.9
spines/μm),
(∼0.36
0.29μm),
include
an
unreported
‘spine
cluster’
structure—features
may
be
congruent
PF
association
amplification
human-specific
adaptations.
By
extrapolation,
500,000
1
million
synaptic
inputs
compared
30-40,000
mouse.
Collectively,
morphology
arrangement
is
quantitatively
qualitatively
distinct
from
rodent.
Multi-branched
more
prevalent
posterior
lateral
cerebellum,
co-varying
functional
boundaries,
supporting
hypothesis
this
morphological
permits
expanded
multiplexing
subserve
task-dependent
needs
for
association.
Purkinje
cell
(PC)
dendrites
are
optimized
to
integrate
the
vast
cerebellar
input
array
and
drive
sole
cortical
output.
PCs
classically
seen
as
stereotypical
computational
units,
yet
mouse
morphologically
diverse
those
with
multi-branched
structure
can
receive
non-canonical
climbing
fiber
(CF)
multi-innervation
that
confers
independent
compartment-specific
signaling.
While
otherwise
uncharacterized,
human
universally
multi-branched.
Do
they
exceed
allometry
achieve
enhanced
integrative
capacities
relative
PCs?
To
answer
this,
we
used
several
comparative
histology
techniques
in
adult
analyze
cellular
morphology,
parallel
(PF)
CF
arrangement,
regional
PC
demographics.
Human
substantially
larger
than
previously
described;
allometric
constraint
by
thickness
largest
neuron
brain
6-7cm
total
dendritic
length.
Unlike
mouse,
ramify
horizontally
form
a
multi-compartment
motif
show
multiple
CFs.
spines
denser
(6.9
vs
4.9
spines/μm),
(∼0.36
0.29μm),
include
an
unreported
‘spine
cluster’
structure—features
may
be
congruent
PF
association
amplification
human-specific
adaptations.
By
extrapolation,
500,000
1
million
synaptic
inputs
compared
30-40,000
mouse.
Collectively,
morphology
arrangement
is
quantitatively
qualitatively
distinct
from
rodent.
Multi-branched
more
prevalent
posterior
lateral
cerebellum,
co-varying
functional
boundaries,
supporting
hypothesis
this
morphological
permits
expanded
multiplexing
subserve
task-dependent
needs
for
association.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2025,
Номер
unknown
Опубликована: Фев. 22, 2025
Abstract
We
exploit
identification
of
neuron
types
during
extracellular
recording
to
demonstrate
how
the
cerebellar
cortex’s
well-established
architecture
transforms
inputs
into
outputs.
During
smooth
pursuit
eye
movements,
floccular
complex
performs
distinct
input-output
transformations
temporal
dynamics
and
directional
response
properties.
The
responses
different
interneuron
localize
circuit
mechanisms
each
transformation.
Mossy
fibers
unipolar
brush
cells
emphasize
position
uniformly
across
cardinal
axes;
Purkinje
molecular
layer
interneurons
code
velocity
along
directionally
biased
Golgi
show
unmodulated
firing.
Differential
properties
transformation
last-order
cells.
pinpoint
site
granule
Specific
cell
population
allow
required
in
area
we
study
generalize
many
transformations,
providing
a
complete
framework
understand
computation.
Impact
statement
dissect
computations
performed
by
cerebellum
an
exemplar
sensory-motor
behavior,
taking
advantage
knowledge
architecture,
existence
discrete
types,
newfound
ability
identify
from
recordings.
Our
results
describe
contributions
major
computations,
needed
support
those
create
basis
set
enable
temporally-specific
motor
behavior
learning.
Abstract
Cerebellar
information
processing
is
critically
shaped
by
several
types
of
inhibitory
interneurons
forming
various
intra-cerebellar
feed-forward
and
feed-back
loops.
Evidence
gathered
over
the
past
decades
has
focused
interest
on
a
non-uniform
set
cortical
distinct
from
“classical”
Golgi,
basket
or
stellate
cells,
summarily
referred
to
as
PLIs
(for
Purkinje
cell
layer
interneurons).
Similarly,
cerebellar
nuclear
have
gained
increasing
attention.
Our
understanding
functions
these
cells
still
fragmentary.
For
humans,
we
lack
functional
data,
even
any
dependable
morphological
classification
for
cells.
Here,
I
used
publicly
available
single
based
gene
expression
data
compare
cortex
neurons
humans
mice.
Integration
revealed
transcriptomic
similarities
between
subsets
suggest
known
characteristics
may
be
helpful
devise
strategies
further
characterization
interneurons.
Comparison
human
murine
indicate
that
strongly
differ
genes
characterize
in
This
limits
their
utility
identify
classify
PLIs,
leaves
question
open
number
non-Golgi
resident
granule
layers
humans.