Science,
Год журнала:
2023,
Номер
379(6628), С. 156 - 161
Опубликована: Янв. 12, 2023
Reproducing
ion
channel-based
neural
functions
with
artificial
fluidic
systems
has
long
been
an
aspirational
goal
for
both
neuromorphic
computing
and
biomedical
applications.
In
this
study,
were
successfully
accomplished
a
polyelectrolyte-confined
memristor
(PFM),
in
which
confined
polyelectrolyte-ion
interactions
contributed
to
hysteretic
transport,
resulting
memory
effects.
Various
electric
pulse
patterns
emulated
by
PFM
ultralow
energy
consumption.
The
property
of
enabled
the
mimicking
chemical-regulated
pulses.
More
importantly,
chemical-electric
signal
transduction
was
implemented
single
PFM.
With
its
structural
similarity
channels,
is
versatile
easily
interfaces
biological
systems,
paving
way
building
devices
advanced
introducing
rich
chemical
designs.
Advanced Functional Materials,
Год журнала:
2018,
Номер
29(5)
Опубликована: Дек. 4, 2018
Abstract
The
ability
of
high‐order
tuning
the
synaptic
plasticity
in
an
artificial
synapse
can
offer
significant
improvement
processing
time,
low‐power
recognition,
and
learning
capability
a
neuro‐inspired
computing
system.
Inspired
by
light‐assisted
dopamine‐facilitated
activity,
which
achieves
rapid
adaptation
lowering
threshold
plasticity,
two‐terminal
organolead
halide
perovskite
(OHP)‐based
photonic
is
fabricated
designed
modified
both
electrical
pulses
light
illumination.
Owing
to
accelerated
migration
iodine
vacancy
inherently
existing
coated
OHP
film
under
illumination,
device
exhibits
light‐tunable
functionalities
with
very
low
programming
inputs
(≈0.1
V).
It
also
demonstrated
that
long‐term
potentiation
decreases
weight
further
modulates
when
illuminates
device,
phenomenologically
analogous
dopamine‐assisted
process.
Notably,
exposure,
pattern
recognition
≈81.8%
accuracy
after
only
2000
phases
(60
000
=
one
epoch)
consumption
(4.82
nW/the
initial
update
for
potentiation),
≈2.6
×
10
3
times
lower
than
weights
are
updated
high
pulses.
Advanced Electronic Materials,
Год журнала:
2019,
Номер
5(9)
Опубликована: Авг. 14, 2019
Abstract
To
realize
highly
efficient
neuromorphic
computing
that
is
comparable
to
biological
counterparts,
bioinspired
systems,
consisting
of
biorealistic
artificial
synapses
and
neurons,
are
developed
with
memristive
devices
native
dynamics
resembling
neurons.
Tremendous
materials
have
been
successfully
used
emulate
diverse
functions
synapses,
as
well
in
the
last
decade.
Herein,
approaches
certain
synaptic
or
neuronal
introduced
state‐of‐art
experimental
demonstrations.
First,
working
principles
neurons
briefly
presented
provide
guidance
for
developing
Second,
recent
advances
development
homosynaptic
heterosynaptic
plasticity
disscussed.
In
particular,
important
learning
rules,
like
spiking‐timing‐dependent
Bienenstock–Cooper–Munro
elaborated
according
level
faithfulness
synapses.
Memristive
including
bioplausible
biophysical
described.
Finally,
challenges
perspectives
based
on
discussed.
Advanced Materials,
Год журнала:
2019,
Номер
31(21)
Опубликована: Апрель 8, 2019
All
external
sensory
stimuli
produce
a
spatiotemporal
pattern
of
action
potentials,
which
is
transmitted
to
the
biological
neural
system
be
processed.
The
relative
timing
synaptic
spikes
from
different
presynaptic
neurons
represents
features
stimuli.
A
fundamental
prerequisite
in
cortical
information
processing
discrimination
input
sequences.
Here,
capacitively
coupled
multiterminal
oxide-based
neuro-transistors
are
proposed
for
processing,
mimicking
dendritic
discriminability
experimental
results
demonstrate
that
such
neuromorphic
devices
can
act
as
compartments
computation.
Also,
an
example
sound
location
functionality
human
brain
also
emulated
by
constructing
simple
artificial
network
based
on
neuro-transistors.
Science,
Год журнала:
2023,
Номер
379(6628), С. 156 - 161
Опубликована: Янв. 12, 2023
Reproducing
ion
channel-based
neural
functions
with
artificial
fluidic
systems
has
long
been
an
aspirational
goal
for
both
neuromorphic
computing
and
biomedical
applications.
In
this
study,
were
successfully
accomplished
a
polyelectrolyte-confined
memristor
(PFM),
in
which
confined
polyelectrolyte-ion
interactions
contributed
to
hysteretic
transport,
resulting
memory
effects.
Various
electric
pulse
patterns
emulated
by
PFM
ultralow
energy
consumption.
The
property
of
enabled
the
mimicking
chemical-regulated
pulses.
More
importantly,
chemical-electric
signal
transduction
was
implemented
single
PFM.
With
its
structural
similarity
channels,
is
versatile
easily
interfaces
biological
systems,
paving
way
building
devices
advanced
introducing
rich
chemical
designs.