We
consider
the
problem
of
olfactory
searches
in
a
turbulent
environment.
focus
on
agents
that
respond
solely
to
odor
stimuli,
with
no
access
spatial
perception
nor
prior
information
about
odor.
ask
whether
navigation
target
can
be
learned
robustly
within
sequential
decision
making
framework.
develop
reinforcement
learning
algorithm
using
small
set
interpretable
states
and
train
it
realistic
cues.
By
introducing
temporal
memory,
we
demonstrate
two
salient
features
traces,
discretized
few
states,
are
sufficient
learn
plume.
Performance
is
dictated
by
sparse
nature
odors.
An
optimal
memory
exists
which
ignores
blanks
plume
activates
recovery
strategy
outside
obtain
best
performance
letting
their
show
mostly
casting
cross
wind,
similar
behavior
observed
flying
insects.
The
robust
substantial
changes
plumes,
suggesting
minor
parameter
tuning
may
adapt
different
environments.
Nature Communications,
Год журнала:
2022,
Номер
13(1)
Опубликована: Авг. 8, 2022
Abstract
To
navigate
towards
a
food
source,
animals
frequently
combine
odor
cues
about
source
identity
with
wind
direction
location.
Where
and
how
these
two
are
integrated
to
support
navigation
is
unclear.
Here
we
describe
pathway
the
Drosophila
fan-shaped
body
that
encodes
attractive
promotes
upwind
navigation.
We
show
neurons
throughout
this
encode
odor,
but
not
direction.
Using
connectomics,
identify
local
called
h∆C
receive
input
from
previously
described
pathway.
exhibit
odor-gated,
direction-tuned
activity,
sparse
activation
of
in
reproducible
direction,
activity
required
for
persistent
orientation
during
odor.
Based
on
connectome
data,
develop
computational
model
showing
can
promote
goal
such
as
an
source.
Our
results
suggest
processed
by
separate
pathways
within
goal-directed
Journal of Neuroscience,
Год журнала:
2018,
Номер
38(44), С. 9383 - 9389
Опубликована: Окт. 31, 2018
Localizing
the
sources
of
stimuli
is
essential.
Most
organisms
cannot
eat,
mate,
or
escape
without
knowing
where
relevant
originate.
For
many,
if
not
most,
animals,
olfaction
plays
an
essential
role
in
search.
While
microorganismal
chemotaxis
relatively
well
understood,
larger
animals
algorithms
and
mechanisms
olfactory
search
remain
mysterious.
In
this
symposium,
we
will
present
recent
advances
our
understanding
flies
rodents.
Despite
their
different
sizes
behaviors,
both
species
must
solve
similar
problems,
including
meeting
challenges
turbulent
airflow,
sampling
environment
to
optimize
information,
incorporating
odor
information
into
broader
navigational
systems.
Neuron,
Год журнала:
2019,
Номер
104(3), С. 544 - 558.e6
Опубликована: Авг. 27, 2019
In
pursuit
of
food,
hungry
animals
mobilize
significant
energy
resources
and
overcome
exhaustion
fear.
How
need
motivation
control
the
decision
to
continue
or
change
behavior
is
not
understood.
Using
a
single
fly
treadmill,
we
show
that
flies
persistently
track
food
odor
increase
their
effort
over
repeated
trials
in
absence
reward
suggesting
dominates
negative
experience.
We
further
tracking
regulated
by
two
mushroom
body
output
neurons
(MBONs)
connecting
MB
lateral
horn.
These
MBONs,
together
with
dopaminergic
Dop1R2
signaling,
behavioral
persistence.
Conversely,
an
octopaminergic
neuron,
VPM4,
which
directly
innervates
one
acts
as
brake
on
feeding
olfaction.
Together,
our
data
suggest
function
for
internal
state-dependent
expression
can
be
suppressed
external
inputs
conveying
competing
drive.
Neuron,
Год журнала:
2020,
Номер
107(5), С. 924 - 940.e18
Опубликована: Июль 17, 2020
Spatial
maps
in
the
brain
are
most
accurate
when
they
linked
to
external
sensory
cues.
Here,
we
show
that
compass
Drosophila
is
direction
of
wind.
Shifting
wind
rightward
rotates
as
if
fly
were
turning
leftward,
and
vice
versa.
We
describe
mechanisms
several
computations
integrate
information
into
compass.
First,
an
intensity-invariant
representation
computed
by
comparing
left-right
mechanosensory
signals.
Then,
signals
reformatted
reduce
coding
biases
inherent
peripheral
mechanics,
cues
brought
same
circular
coordinate
system
represents
visual
self-motion
Because
incorporates
both
cues,
it
should
enable
navigation
under
conditions
where
no
single
cue
consistently
reliable.
These
results
how
local
can
be
transformed
a
global,
multimodal,
abstract
space.
How
insects
navigate
complex
odor
plumes,
where
the
location
and
timing
of
packets
are
uncertain,
remains
unclear.
Here
we
imaged
plumes
simultaneously
with
freely-walking
flies,
quantifying
how
behavior
is
shaped
by
encounters
individual
packets.
We
found
that
navigation
was
stochastic
did
not
rely
on
continuous
modulation
speed
or
orientation.
Instead,
flies
turned
stochastically
stereotyped
saccades,
whose
direction
biased
upwind
prior
encounters,
while
magnitude
rate
saccades
remained
constant.
Further,
used
to
modulate
transition
rates
between
walks
stops.
In
more
regular
environments,
continuously
orientation,
even
though
can
still
occur
randomly
due
animal
motion.
find
in
less
predictable
random
both
space
time,
walking
encounter
timing.
Annual Review of Condensed Matter Physics,
Год журнала:
2021,
Номер
13(1), С. 191 - 213
Опубликована: Ноя. 9, 2021
Fluid
turbulence
is
a
double-edged
sword
for
the
navigation
of
macroscopic
animals,
such
as
birds,
insects,
and
rodents.
On
one
hand,
enables
pheromone
communication
among
mates
possibility
locating
food
by
their
odors
from
long
distances.
Molecular
diffusion
would
indeed
be
unable
to
spread
over
relevant
distances
in
natural
conditions.
other
turbulent
flows
are
hard
predict,
learning
effective
maneuvers
navigate
them
challenging,
we
discuss
this
review.
We
first
provide
summary
olfactory
organs
that
sense
airborne
or
surface-bound
odors,
well
computational
tasks
animals
face
when
extracting
information
useful
an
signal.
A
compendium
dynamics
transport
emphasizes
those
aspects
directly
impact
animals’
behavior.
The
state
art
on
navigational
strategies
discussed,
followed
concluding
section
dedicated
future
challenges
field.
For
many
organisms,
searching
for
relevant
targets
such
as
food
or
mates
entails
active,
strategic
sampling
of
the
environment.
Finding
odorous
may
be
most
ancient
search
problem
that
motile
organisms
evolved
to
solve.
While
chemosensory
navigation
has
been
well
characterized
in
microorganisms
and
invertebrates,
spatial
olfaction
vertebrates
is
poorly
understood.
We
have
established
an
olfactory
assay
which
freely
moving
mice
navigate
noisy
concentration
gradients
airborne
odor.
Mice
solve
this
task
using
gradient
cues
do
not
require
stereo
performance.
During
performance,
respiration
nose
movement
are
synchronized
with
tens
milliseconds
precision.
This
synchrony
present
during
trials
largely
absent
inter-trial
intervals,
suggesting
sniff-synchronized
a
behavioral
state
rather
than
simply
constant
accompaniment
fast
breathing.
To
reveal
spatiotemporal
structure
these
active
sensing
movements,
we
used
machine
learning
methods
parse
motion
trajectories
into
elementary
motifs.
Motifs
fall
two
clusters,
correspond
investigation
approach
states.
Investigation
motifs
lock
precisely
sniffing,
individual
preferentially
occur
at
specific
phases
sniff
cycle.
The
allocentric
indicates
advantage
both
sides
sharpest
part
odor
gradient,
consistent
serial-sniff
strategy
sensing.
work
clarifies
sensorimotor
strategies
mouse
guides
ongoing
underlying
neural
mechanisms.