of 30
1
Transforming
representations
of movement
from body
-
to world
-
centric
space
Jenny Lu
1
, Elena
A.
Westeinde
1†
, Lydia Hamburg
2
,
Paul
M.
Dawson
1
,
Cheng Lyu
3
, Gaby Maimon
3
,
Shaul
Druckmann
2
, and Rachel I. Wilson
1
1
Department of Neurobiology
and Howard
Hughes Medical Institute
, Harvard M
edical School, Boston, MA,
USA.
2
Department of Neurobiology, Stanford University, Stanford, CA, USA.
3
Laboratory of Integrative Brain Function and Howard Hughes Medical Institute, The Rockefeller University,
New York, NY,
USA.
equal contributions
*e
-
mail:
rachel_wilson@hms.harvard.edu
When an animal moves through the world, its brain receives a stream of information about the body’s
translational movement. These incoming movement signals, relayed from sensory
organs or as copies of
motor commands, are referenced relative to the body. Ultimately, such body
-
centric movement signals
must be transformed into world
-
centric coordinates for navigation
1
. Here we show that this computation
occurs in the fan
-
shaped body in the
Drosophila
brain. We
identify
two cell types in the fan
-
shaped body,
PFNd and PFNv
2,3
,
that
conjunctively encode translational velocity signals and heading signals
in
walking flies
. Specifically, PFNd and PFNv neurons
form a Cartesian representation of body
-
centric
translational velocity
acquired from premotor brain regions
4,5
that is layered onto
a world
-
centric
heading
representation
inherited from upstream compass neurons
6
-
8
.
Then, w
e demonstrate that the next
network layer,
comprising
hΔB neurons
,
is wired
so as to
transform
the representation of
translational
velocity
from body
-
centric to
world
-
centric
coordinates
.
We show that this
transformation
is predicted by
a computational model derived directly from electron microscopy connectomic data
9
.
The model
illustrates
the key role of
a specific network motif,
whereby
the PFN neurons that synapse onto the same
hΔB
neuron have heading
-
tuning differences that offset the
differences in their preferred body
-
centric
directions of movement
. By integrating
a world
-
centric repre
sentation of
travel velocity over time, it
should be possible for the brain to form a
working memory
of the path traveled
through
the
environment
10
-
12
.
Natural locomotor behavior
combines
body
rotation
with
forward and
lateral
body translation
.
A
ll these
movements
contribute to the path traveled by the animal through the world
. For example,
when a
fly walks in a
curved path to the left, it
steps forward while rotating left and also
side
stepping
laterally
to the left
13,14
(Fig. 1a
,
Extended Data Fig. 1
)
.
T
he brain would underestimate the
path’s curvature unless it tracked
the body’s
lateral
movement
in addition to tracking its forward
movement
.
A walking insect can measure
its forward and lateral velocity
using proprioceptive signals and copies of internal
motor command signals
15
, as well as optic flow signals
16,17
. All these velocity signals arrive in body
-
centric
coordinates.
Ultimately,
however,
if the brain is to construct a memory of the fly’s
navigation path
10,18
,
it must
convert these
body
-
centric
signals
into
world
-
centric coordinates
1,19
.
F
or example, moving forward while facing
n
orth
should be
represented by the brain
’s navigation systems
as
equivalent to moving
right
while facing
w
est
(Fig
.
1a)
even if these two situations involve very different peripheral sensory signals and motor commands
.
If
this computation
is
performed in
the fly
brain,
it
would
likely occur downstream from the ellipsoid body
(EB)
of the central complex
.
The EB
contains a
topographic map of heading
which is
anchored to multimodal
environmental cues
6,20,21
in other words,
a world
-
centric heading representation
.
EB
projection neurons
(EPG
neurons
or “compass neurons”
) send
this
heading information
to
the
pro
tocerebral bridge (PB)
7,8
(Fig. 1b)
.
Indeed, a
recent study in the honeybee showed that
, in
some regions of the
central complex
downstream from
the PB, there are
incoming projections from neurons that encode
the direction of
optic flow
during flight
22
. This
result suggests
a possible fusion of translational velocity signals and heading signals in
the central complex
.
To
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
2
pursue this idea, we turned to the fly
Drosophila melanogaster
, where we can use transgenic driver lines
2
,3
to
target central complex neuron
s with well
-
defined synaptic connectivity
19
.
Neurons encoding heading and velocity
T
o
determine whether
heading information is integrated with velocity information
downstream from EPG
neurons
, we
used specific driver lines
3
to express
a
fast
cal
cium indicator
(
j
GCaMP
7f
)
23
in
various
cell types in
the PB
. We imaged these neurons
using
two
-
photon excitation microscopy
as the fly walked
freely on a
spherical treadmill
, surrounded by
a
visual virtual reality environment
24
(Fig. 1c)
. This environment was a
360°
panorama displaying a
prominent heading cue
namely,
a
bright
object
which rotated rightward when the fly
made a fictive leftward
turn
, and
vice versa
.
We
focused on
a
specific
PB cell type (PFNd
) downstream from compass neurons (Fig. 1d)
, because PFN
neurons are anatomically well
-
positioned to receive both heading information and translational velocity
information
22
.
We
found that
PFNd dendrites form two complete map
s
of heading
,
on
e on
each side of the PB
(
PFNd.
L
and PFNd.
R
,
Fig. 1
d
)
, resembling the two compl
ete maps in EPG axons (Fig. 1e,f
). A
s th
e fly rotates
clockwise, there is a
bump
of activity which moves
leftwa
rd
across the PFNd
.
R
population, and another bump
that moves
leftward
across the PFNd
.
L
population.
Overall, t
he position of the bump is
correlated with heading
,
as in EPG neurons
(
Extended Data
Fig.
2
).
Moreover
,
we
found
that
PFNd neurons
also
show dire
ction
-
selective responses to translational movement
.
Specifically
,
the
PFNd.
R
bump amplitude
increases
when the fly
translates
forward and right
, whereas
PFNd.
L
bump amplitude
increases
when the fly
translates
forward and left
(Fig. 1g
-
i
, Extended Data
Fig. 2
)
.
Thus,
PFNd
neurons
are sensitive to
both
forward and lateral velocity
.
These neurons are
somewhat
correlated with
rotation
as well
(Extended Data Fig.
2
)
, but this is
unsurprising
because lateral and rotational velocity are highly
correlated
durin
g
walking (Extended Data Fig. 1)
13,25
; moreover,
as we will see, the function of these neurons
is
most relevant to translational movement.
We
estimated
the preferred
translational
direction
of each population by fitting a linear function to
their
two
-
dimensional tuning profile
(Extended Data Fig.
2
); this
analysis
suggested
that PFNd.
R
and PFNd.
L
neurons
prefer
translation
at an angle of 31° and
-
31°, respectively
, relative
to heading (Fig. 1i
).
G
raded
velocity
increases in the preferred direction
produce graded increases in bump amplitude (Extended Data Fig.
2
). By
contrast, EPG neurons themselves are relatively insensitive to translational velocity
(Fig. 1g
).
Next, we used
whole
-
cell recordings
to
measure
the relative latency of
PFNd
activity and behavior
.
When we
cross
-
correlated
voltage
and
velocity
in the cell’s preferred t
ranslational velocity direction
, we found that
voltage
typically
lagged velocity by about
0.1 s
(
Fig. 1j)
.
This lag implies that PFNd
neurons are not directly
causal for these changes in velocity. In other words, PFNd neurons are measuring velocity rather than
commanding velocity.
These electrophysiological recordings also revealed
that effect of
velocity
(in the cell’s preferred
body
-
centric
direction) is to multiplicatively scale
the effect of heading
.
For
a given heading, the cell’s membrane potential
depends quite linearly on
velocity
(Fig. 1k)
. High
velocities
amplify the depolarizing effect of the pr
eferred
heading, as well as the hyperpolarizing effect of the
anti
-
preferred heading (Fig. 1k).
In
summary
, PFNd neurons
support
a
conjunctive
code for several navigational variables.
T
he spatial phase of
the activity bump
s
(i.e. the position of the bump
along the PB)
encodes heading
, which is a world
-
centric
variable
.
Meanwhile, t
he amplitude of
each
bump
scales linearly with
translational velocity
in
a
specific
direction
, with
the right and left PFNd populations tuned to different body
-
centric directions
.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
3
Velocity coding via graded
release from
inhibition
To understand the origin
of the
conjunctive
code in
PFNd
neurons
, we identified the inputs to
PFNd neurons
in
the hemibrain connectome
9
(Extended
Data
Fig.
3
)
.
We then used specific driver lines to express
j
GCaMP
7f
in
the
cell
types that provide major
unilateral
input
to PFNd
(Fig. 1d
, Extended Data Fig. 3
)
to understand
how the
left and right PFNd popul
ations acquire their translation
al
direction
selectivity
.
We
fou
nd
strong
direction
-
selective
translational
velocity
signals
in two
cell types, SpsP and LNO2.
These two cell types are output
projection neurons from little
-
studied brain regions that have be
en implicated
as premotor regions associated
with locomotion
(the SPS and LAL, respectively)
4,5
.
Notably,
we found t
hat SpsP and LNO2
neurons
are
anti
-
correlated with forward velocity
(Fig. 2a
-
f
)
, rather
than being positively correlated as PFNd neurons are.
To determine whether
SpsP and LNO2 neurons
might be
inhibitory, we
reconstructed
examples of
these
neurons
in a
se
cond
large
-
scale
EM dataset
26
which allows us to
leverage
machine learning
to
predict
the
neurotransmi
tters
associated with
each cell
27
.
We found that SpsP and
LNO2 output synapses were
mainly
scored
as glutamatergic
.
G
lut
a
mate
is
often
an inhibitory neurotransmitter
in the
Drosophila
brain
28
, and indeed
we confirmed that optogenetic activation of SpsP neurons
produces PFNd
neuron hyperpolarization
, with
the characteristic pharmacological signature of monosynaptic inhibition via
glutamate
-
gated chloride
(GluCl
)
channels (Fig.
2g)
.
We also confirmed that a split
-
Gal4 hemidriver that
reports vesicular glutamate transporter expression
29
can drive expression
in
LNO2 neurons (Extended Data Fig.
4
). Thus, both types of premotor projection neurons
are likely inhibitory
.
This means that, as
forward velocity
increases, PFNd neurons
receive graded disinhibition
from
premotor regions
.
Recall that
PFNd neurons
are sensitive to lateral velocity
.
We found that
SpsP
and
LNO2
neurons
are also
sensitive
to lateral velocity
(Fig. 2a
-
f
, Extended Data Fig.
5
)
; moreover,
their lateral direction selectivity can
explain
the
observed
lateral
direction selectivi
ty of
PFNd neurons
(Fig. 2h)
.
Specifically
, when the fly
walks
to
the
right
,
SpsP
and LNO2
neurons inhibit PFNd.
L
, while
also
disinhibiting
PFNd.
R
.
Thus, these premotor
projection neurons can account for
both the
lateral
and forward velocity tuning of PFNd
neurons.
PFNd neurons also receive a major input from
IbSpsP
neurons
(
Extended Data Fig.
5
)
.
However, we
found that
IbSpsP
neurons are not very selective for the direction of translational motion (Extended Data Fig. 5)
.
Thus,
they are
unlikely to account for
direction
-
selective
velocity tuning in PFNd neurons
.
To summarize, PFNd neurons
in each brain hemisphere
inherit
a
map of heading direction from the compass
system.
These maps are under the inhibitory control of premotor projectio
n neurons
which encode
both
the
speed and
direction of body movement
.
Lateral
movement
releases
the ipsilateral
PFNd
heading
map
from
inhibition,
while
increasing
inhibit
ion
of
the contralateral map.
By contrast
, forward movement disinhibits both
map
s
.
In
this way, body
-
centric
velocity information is
layered
onto
the
world
-
centric
map of heading.
A complete
Cartesian
system
for velocity
Next, w
e investigated a
nother cell type, PFNv,
with
a
morphology
nearly identical
to PFNd neurons (Fig. 3a).
Like PFNd neuron
s
, PFNv neurons are downstream from compass neurons (Fig. 1d
, Extended Data Fig. 3
).
Accordingly, we found that
PFNv neurons
also
form a heading map in each
half
of the PB
(Fig. 3b), with a
bump of activity whose
position correlates with heading (
Extended Data Figure
6
).
Notably
,
we found that
PFNv bump amplitude
encode
s
translational velocity along
direction
s
that are opposite
to those
encoded
by
PFNd neurons
.
Whereas PFNd neurons are excited by forward and ipsi
lateral velocity,
PFNv neurons
prefer
backward and contralateral velocity (Fig. 3
c
-
e
).
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
4
Because
PFNd and PFNv neurons
have
such different
translational direction
selectivity
, we might expect them
to receive
different
inputs from
the brain’s
premotor centers (SPS and LAL
4,5
)
.
The
hemibrain
connectome
9
confirms this
prediction
:
SpsP and LNO2 neurons
make many synapses onto
PFNd neurons
, but they
make
almost no synapses onto PFNv neurons
.
Rather, PFNv neurons are downstream from
other
LAL projection
neuron types (LNO1 a
nd LNO3; Extended Data Fig.
3)
which
contribute relatively little to PFNd input.
LNO1
neuron activity was not
strongly
correlated w
ith body velocity under our exper
i
mental conditions
(Extended
Data Fig.
7
)
, so it seems more likely that LNO3 neurons are the major drivers of PFNv responses, although we
were not able to obtain a selective driver for LNO3 neurons that would allow us to te
st this idea.
In summary, PFNd and PFNv neurons
together constitute a
complete
set of Cartesian
axes
for
encoding
translational velocity (Fig. 3e). This representation is
in
body
-
centric coordinates
.
Thus, w
hen the fly rotates,
the PFN velocity
-
vector rep
resentation rotates
along with the
fly’s
heading
(Fig. 3e).
By
conjointly encoding
the
full
360°
of
body
-
centric
translational velocity directions and the full 360
°
of heading directions, PFNd and
PFNv neurons are poised to transform velocity representations from a body
-
centric coordinate system to a
world
-
centric coordinate system.
Integrating opponent populations
Next, we examined what happens downstream from
PFNd and PFNv neurons
. Both of these cell types
project
to the fan
-
shaped body (FB)
, where their axonal projections preserve the rough topography of the heading map
2
.
The hemibrain connectome
9
reveals that
PFNd and PFNv
projections
from the same PB glomerulus are
precisely overlaid in the FB
,
and
they
converge onto the same cell type, called hΔB (Fig. 3g).
This
projection pattern
seems puzzling
.
PF
Nd and PFNv neurons from the same PB glomerulus have
opposing
velocity preferences
(Fig. 3e)
. Thus, the
convergence
of these projections
would
seem to
cancel the directional
information provided by
each
PFN
population.
The resolution to this puzzle
lies in the subcellular
targeting
of these synaptic connections
.
Each
hΔB
neuron
spans
half
of
the horizontal extent of the FB
, with a dendrite at one
pole
and axon terminals at the other
(Fig.
3g)
.
In analyzing the connectome
, we discovered
that
PFNv neurons
synapse
selectively
onto
hΔB
dendrites,
whereas
PFNd neurons
synapse onto
both
hΔB
axon terminals
and dendrites, with a preference for axon
terminals
(Fig. 3h)
.
If
we take the perspective of
an individual hΔB neuron,
we s
ee that its
PFNd and PFNv inputs from the left PB
will have opposite heading tuning
(Fig. 3i, left)
. Similarly, PFNd and PFNv inputs from the right PB will also
have opposite heading tuning
(Fig. 3i, right)
. If PFNd neurons
exclusively
targeted hΔB axon te
rminals (as
schematized in Fig. 3i), then the heading difference would be
exactly
180°
, but b
ecause PFNd neurons also
target hΔB dendrites, the heading difference is closer to 150°
(Fig. 3j).
Recall that PFNd and PFNv neurons
from the PB on the same hemisp
here have
nearly
oppo
site
preferred directions of
translational velocity, which
roughly matches this heading difference (Fig. 3e).
To summarize
,
PFNd and PFNv inputs to the same hΔB neuron have nearly opposite heading tuning,
and also
nearly opposite translational velocity tuning. As a result,
their
translational velocity
signals
should
combine
constructively, rather than destructively.
The key poi
nt is that PFNd and PFNv neurons that wire together (onto
the same hΔB neuron) have heading
-
tuning differences that offset their translation
al
direction
tuning
differences
.
This is a notable example of a geometrical computation in the brain which relies on
the precise
geometry of subcellular synaptic targeting.
Interhemispheric integration
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
5
The
re is yet another important feature of the
PFN projection pattern: projections from the right and left copies
of the PB heading map are not aligned in the FB
2
. Rather, there is a systematic shift
,
so t
hat
projections from the
right PB are shifted leftward, while projections from the left PB are shifted rightward (Fig. 3k). Thus,
when
we
analyze
d
the hemibrain connectome from the perspective of
an
individual
hΔB
neuron
, we
found
that PFNd
inputs
from the right and left
have preferred headings that differ by
~45°
, on average
(Fig. 3j).
Note that this
heading
-
tuning difference
is opposite to
the translation
al
direction tuning
difference in the two PFNd
populations
(Fig. 3k)
.
This matched shift allo
ws the hΔB neuron to constructively combine information from
the left and right
hemispheres
.
The same matched shift occurs for PFNv neurons, although here
the
re is a somewhat larger
heading tuning
difference
between left and
right inputs
to an individual
hΔB neuron
(Fig. 3j)
.
Accordingly, there is also a
somewhat larger difference in the preferred
translational
directions of r
ight and left
PFNv populations
(as
compared with
the difference in the right
-
left
PFNd populations
,
Fig. 3e).
Again, the
key point i
s that
the PFN
neurons that wire together have heading
-
tuning differences that offset their translation
al
direction
tuning
differences
.
A body
-
to
-
world coordinate transformation
To see the functional consequence of this anatomy, consider again the perspe
ctive of an individual hΔB neuron
(Fig. 3l). Its inputs
have diverse preferred headings, and diverse body
-
centric direction preferences. What these
inputs have in common is their preferred world
-
centric travel direction (Fig. 3l). Thus, the
identity of the active
hΔB
neurons should encode the direction of world
-
centric travel. T
ravel in a given direction should activate the
same hΔB neurons, regardless of whether the fly is walking right or left, forwards or backwards.
In other words,
the bra
in’s translational
velocity
-
vector
should be
transformed from body
-
centric coordinates to world
-
centric
coordinates.
To investigate this idea, we implemented a computational model
of the network
, comprising 40 PFNd, 20
PFNv
, and 19
hΔB
neurons,
identical
to the cell numb
ers
in the hemibrain connectome
9
.
For simplicity, we
directly modeled the activity of PFN neurons as a function of heading and body
-
centric
translational
velocity
,
rather than modeling the network inputs to PFN neurons.
In the model, e
ach PFN neuron is
cosine
-
tuned to
heading
.
T
he amplitude of
its
heading response
is
multiplicatively
scaled by
the
nonnegative
component of
the
fly’s translation
al
velocity
in that
cell’s preferred body
-
centric direction (Fig. 4
a
).
This
follows
what we see
in
PFN membrane
voltage
data
, where velocity multiplicatively scales the cell’s heading tuning
(Fig. 1k).
PFN
hΔB
connectivity is
taken directly from the
connectome
,
with
weights
proportional
to the number of
synapses
per
connection
(Fig. 4b)
.
This
direct
approach is a departure from previous
studies
where connections
weights
from connectome data
were computationally optimized to achieve a particular modeli
ng result
30,31
;
here,
connections were simply taken directly from data
, with no optimization
.
Finally, each hΔB neuron in the
model simply sums its
PFN inputs
.
Notably,
in this model,
the
position of the
hΔB
activity bump
tracks the fly’s world
-
cent
ric travel
direction
(Fig.
4c) and
the amplitude of the bump scales with travel speed, although the gain of this relationship is highest
when the fly is walking straight forward (Fig. 4d). The model’s travel direction encoding i
s remarkably g
ood:
f
or a given world
-
centric travel direction, the position of the hΔB bump
is
essentially always the same,
regardless of whether the fly
’s
heading
direction
matches it travel
direction
(Fig. 4e).
Travel direction
encoding
is
disrupted
if w
e permute
the connectivity matrix to remove the 45° phase shift in
hΔB inputs from left and
right PFN populations
(Fig. 4b
,e
).
It
is
even more disrupted
if
we remove either PFNv or PFNd
neurons from
the model
(Fig. 4e)
. Although PFNv
neurons contribute onl
y
a
relatively
small
number of synapses
(Fig.
3
h),
they are essential
because
their velocity tuning opposes that of PFNd neurons
. Unless both opponent
populations are functional, they cannot push the hΔB bump to the correct position.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
6
Finally, we tested the predictions of this model by imaging
j
GCaMP7f in hΔB neurons. Although hΔB dendrites
a
nd axons are intercalated
, calcium signals should be dominated by
axons
, given the concentration of calcium
channels at presynaptic terminals. Thus
, we
would expect to see
a moving bump of activity whose amplitude
scales with travel speed. This is indeed what we observed (Fig. 4f,g).
The
bump position
generally
followed the
visual heading cue, which is
expected
whenever the fly is
walking
fairly straight
.
Notably, however, the hΔB
bump
shifted away
from the cue when the fly’s lateral speed was high relative to its forward velocity (Fig. 4f)
i.e., when travel direction deviated from heading (Fig. 4h)
.
The observed
bump shift
was
relatively
small, which
is what we would expect
, given the limited
temporal
resolution
of calcium imaging and the fact that
travel
-
heading deviations are very transient during normal walking
(Fig. 4f). Importantly, however,
we saw
a
quantitatively similar
shift in
e
very hΔB experiment
,
and we
did not observe
this
type of shift
in PFNd
experiments
(Fig. 4i
-
k
, Extended
Data
Fig. 8
)
or
in
PFNv experiments (Extended Data Fig.
8
)
.
In summary,
while
the
PFN bump
position
faithfully tracks
the fly’s
heading
direction
,
the
hΔB bump
position
is
sensitive to the
intermittent
deviations between travel and heading
that
occur during
normal walking
.
This
result
supports the idea
that hΔB neurons form a map of travel direction, rather than heading direction.
Meanwhile, the amplitud
e of hΔB
activity
encodes travel
speed
.
Together, these results imply
that hΔB neurons
constitute a vector
-
velocity representation of travel direction and speed
in world
-
centric coordinates.
Discussion
The transformation of velocity signals into world
-
centric coordinates is a prerequisite for computing a world
-
centric path during navigation.
Drosophila
and other
insects can use
path integration
to navigate back to a
familiar site
in the absence of
spatia
l position
cues
10
-
12
.
Path integration
requires the brain to measure
not only
direction
but also
d
istance
(or speed).
A
simple model proposed twenty
-
five years ago by Wi
ttmann and
Schwegler
32
proposed that the compass system in the insect brain
33
is the source of direction information
;
for
ward
speed
then
multiplicatively scale
s
the output of the compass system, and this representation
is
integrated over time
to produce a vectorial representation of the distance traveled in each heading direction
.
A
limitation of this model is that it
assume
s
all
translational
velocity
is straight ahead
, with no lateral component
.
This problem is solved in a model recently proposed by Stone, Webb, and colleagues
22
.
Inspired by
experimental work in honeybees, this mode
l begins with a Cartesian coordinate system for translational velocity
specifically, one cell tuned to
forward
-
right
velocity, and another cell tuned to forward left velocity
.
In
the
model
, each of these
two
neurons
projects
to a separate downstream
“int
egrator”
population
which also
receives
a complete heading map
. Heading signals are subtracted
from
translational velocity
signals
, and the result
is
summed
over time
in each of the
se
two
integrator populations
.
Note that, in the Stone
-
Webb model, t
here is no
explicit representation of world
-
centric travel velocity.
Rather,
this model
separately
stores
path components
along two orthogonal axes
of translation
.
It was predicted
that PFN neurons (called CPU4 in bees)
correspond
to the integrator cells o
f the model
.
A limitation of this model is that it only works when there
is
no backward
component to the
agent’s
movement
.
H
ere
we show
that PFN neurons do indeed combine heading and translational velocity signals
.
However,
we
show there are
actually four
populations
of PFN neurons
that collectively tile the full 360° of velocity space
in a
Cartesian coordinate system
.
Contrary to the prevailing model,
we
found no evidence
that PFN
d or PFNv
neurons
integrate velocity information over time to yield a distan
ce calculation
.
Next
, we show that this network
is wired to perform
a computation which has not previously been attributed to
the insect brain: it
computes an explicit representation of the body’s velocity in world
-
centric coordinates
. This
computation
oc
curs at the next layer of the network, in hΔB neurons. The identity of the active neurons in the
hΔB population represents world
-
centric travel direction. This means that, when the
agent
rotates, the hΔB
coordinate system remains stable rather than rotatin
g.
Meanwhile
, the amplitude of hΔB activity
covaries with
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
7
travel
speed.
A
neural representation of speed is important to accurate navigation, and indeed there is behavioral
evidence that
path integration in
walking
16,17
and flying
34
insects
is sensitive to speed cues
.
Moreover, o
ur model shows
that the
crucial element
in
this computation
is the precise
targeting
of PFNd and
PFNv output synapses. In this wiring pattern, PFN neurons converge onto the same hΔB neuron if the
difference in their translational
-
direction preference is offset by a shift in their heading
-
direction
-
preference. We
suggest that an analogous wiring pattern might be found in the vertebrate brain, in the arrangement of inputs to
world
-
centric velocity
-
vector cells
35
.
More gen
erally, our work provides the first detailed mechanistic description of a body
-
to
-
world vectorial
coordinate transformation in any species. This is potentially relevant to the many different vectorial codes in
mammalian
navigation
system
s
1
.
Some of these vectorial codes operate in body
-
centric coordinates
35
-
38
,
whereas others operat
e in world
-
centric coordinates
35,39
-
42
. It has been proposed that the outputs of body
-
centric
vector cells are combined to produce world
-
centric vector cells in downstream network layers
43,44
. This woul
d
seem to be a crucial element of the neurobiology of navigation. Our results show that this does in fact occur
and indeed how it occurs
in an insect brain
.
A parallel study reaches many of the same conclusions
45
.
Ultimately,
path position representations
must be
compared to internal
spatial
goals
.
Then, they must be
transposed back into a body
-
centric reference frame for steering control
25
.
By identifying
wiring
patterns in the
conn
ectome
19
, exploring these patterns in computational models, and testing these models through p
hysiology
experiments, it should be possible to achieve
an
understand
ing of
all these computations at an algorithmic
and
biophysical level
.
Author contributions:
Experiments were designed by J.L. and R.I.W. with input from E.W.
and S.D.
J.L.
performed i
maging
experiments and analyses
.
E.W. performed
electrophysiological
experiments and analyses
.
L.H. and S.D. designed and implemented the computational model with input from J.L. and R.I.W.
P.D.
performed MCFO experiments.
C.L. and G.M. provided the
hΔB
sp
lit
-
Gal4 line
prior to publication
.
Acknowledgements:
This study benefited enormously from the public release of the hemibrain connectome in
January 2020 by the FlyEM Team at Janelia. Isabel Haber and Anna Li annotated neurons and synapses in
the
full ad
ult female brain dataset (
FAFB
)
26
, and Nils Eckstein and Alexander Bates generated neurotransmitter
pr
edictions based on those data, using algorithms they designed and implemented with Jan Funke and Greg
Jefferis. Will Dickson and Michael Dickinson
shared
modified FicTrac software, panels hardware support, and
foam sphere machining help.
Tanya Wolff, Gerry
Rubin, and Vivek Jayaraman provided fly stocks.
We thank
Michael Dickinson and Amir Behbahani for helpful discussions, and members of the Wilson lab for comments
on the manuscript. This work was supported by the Harvard Medical School Neurobiology Imaging
Facility
(NINDS P30
#NS072030
), the HMS Research Computing Group O2 cluster, and the HMS Research
Instrumentation Core Facility. This study was supported by F30 DC017698 (to J.L.),
T32
GM007753
, and U19
NS104655. R.I.W. and G.M. are HHMI
Investigators
.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
8
References
1.
Bicanski, A. & Burgess, N. Neuronal vector coding in spatial cognition.
Nat Rev Neurosci
21
, 453
-
470
(2020).
2.
Wolff, T., Iyer, N. A. & Rubin, G. M. Neuroarchitecture and neuroanatomy of the
Drosophila
central
complex: A GAL4
-
based dissection of protocerebral bridge neurons and circuits.
J. Comp. Neurol.
523
,
997
-
1037 (2015).
3.
Wolff, T. & Rubin, G. M. Neuroarchit
ecture of the Drosophila central complex: A catalog of nodulus
and asymmetrical body neurons and a revision of the protocerebral bridge catalog.
J Comp Neurol
526
,
2585
-
2611 (2018).
4.
Namiki, S. & Kanzaki, R. Comparative neuroanatomy of the lateral access
ory lobe in the insect brain.
Front. Physiol.
7
(2016).
5.
Namiki, S., Dickinson, M. H., Wong, A. M., Korff, W. & Card, G. M. The functional organization of
descending sensory
-
motor pathways in
Drosophila
.
eLife
7
, e34272 (2018).
6.
Seelig, J. D. & Jayaram
an, V. Neural dynamics for landmark orientation and angular path integration.
Nature
521
, 186
-
191 (2015).
7.
Green, J., Adachi, A., Shah, K. K., Hirokawa, J. D., Magani, P. S. & Maimon, G. A neural circuit
architecture for angular integration in
Drosophila
.
Nature
546
, 101
-
106 (2017).
8.
Turner
-
Evans, D., Wegener, S., Rouault, H., Franconville, R., Wolff, T., Seelig, J. D., Druckmann, S. &
Jayaraman, V. Angular velocity integration in a fly heading circuit.
eLife
6
(2017).
9.
Scheffer, L. K., Xu, C. S., Jan
uszewski, M., Lu, Z., Takemura, S. Y., Hayworth, K. J., Huang, G. B.,
Shinomiya, K., Maitlin
-
Shepard, J., Berg, S.
et al.
A connectome and analysis of the adult Drosophila
central brain.
eLife
9
(2020).
10.
Kim, I. S. & Dickinson, M. H. Idiothetic path int
egration in the fruit fly
Drosophila melanogaster
.
Curr.
Biol.
27
, 2227
-
2238 e2223 (2017).
11.
Müller, M. & Wehner, R. Path integration in desert ants, Cataglyphis fortis.
Proc Natl Acad Sci U S A
85
, 5287
-
5290 (1988).
12.
Esch, H. & Burns, J. Distance est
imation by foraging honeybees.
J Exp Biol
199
, 155
-
162 (1996).
13.
DeAngelis, B. D., Zavatone
-
Veth, J. A. & Clark, D. A. The manifold structure of limb coordination in
walking
Drosophila
.
eLife
8
(2019).
14.
Geurten, B. R., Jahde, P., Corthals, K. & Gopfert, M. C. Saccadic body turns in walking Drosophila.
Front Behav Neurosci
8
, 365 (2014).
15.
Wittlinger, M., Wehner, R. & Wolf, H. The ant odometer: stepping on stilts and stumps.
Science
312
,
1965
-
1967 (2006)
.
16.
Ronacher, B. D. & Wehner, R. Desert ants
Cataglyphis fortis
use self
-
induced optic flow to measure
distance travelled.
Journal of Comparative Physiology A
177
, 21
-
27 (1995).
17.
Schöne, H. Optokinetic speed control and estimation of travel distance i
n walking honeybees.
Journal of
Comparative Physiology A
179
, 587
-
592 (1996).
18.
Heinze, S., Narendra, A. & Cheung, A. Principles of insect path integration.
Curr. Biol.
28
, R1043
-
R1058 (2018).
19.
Hulse, B. K., Haberkern, H., Franconville, R., Turner
-
Eva
ns, D. B., Takemura, S., Wolff, T., Noorman,
M., Dreher, M., Dan, C., Parekh, R., Hermundstad, A. M., Rubin, G. M. & Jayaraman, V. A connectome
of the Drosophila central complex reveals network motifs suitable for flexible navigation and context
-
dependent
action selection.
bioRxiv
, doi:10.1101/2020.1112.1108.413955 (2020).
20.
Okubo, T. S., Patella, P., D'Alessandro, I. & Wilson, R. I. A neural network for wind
-
guided compass
navigation.
Neuron
107
, 924
-
940 e918 (2020).
21.
Hardcastle, B. J., Omoto, J. J.,
Kandimalla, P., Nguyen, B.
-
C. M., Keleş, M. J., N.K., B., Hartenstein, V.
& Frye, M. A. A visual pathway for skylight polarization processing in Drosophila.
bioRxiv
,
doi.org/10.1101/2020.1109.1110.291955 (2020).
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
9
22.
Stone, T., Webb, B., Adden, A., Weddig,
N. B., Honkanen, A., Templin, R., Wcislo, W., Scimeca, L.,
Warrant, E. & Heinze, S. An anatomically constrained model for path integration in the bee brain.
Curr.
Biol.
27
, 3069
-
3085.e3011 (2017).
23.
Dana, H., Sun, Y., Mohar, B., Hulse, B. K., Kerlin, A.
M., Hasseman, J. P., Tsegaye, G., Tsang, A.,
Wong, A., Patel, R.
et al.
High
-
performance calcium sensors for imaging activity in neuronal
populations and microcompartments.
Nat Methods
16
, 649
-
657 (2019).
24.
Reiser, M. B. & Dickinson, M. H. A modular disp
lay system for insect behavioral neuroscience.
J.
Neurosci. Methods
167
, 127
-
139 (2008).
25.
Rayshubskiy, A., Holtz, S. L., D'Alessandro, I., Li, A. A., Vanderbeck, Q. X., Haber, I. S., Gibb, P. W.
& Wilson, R. I. Neural circuit mechanisms for steering con
trol in walking Drosophila.
bioRxiv
2020.04.04.024703; doi.org/10.1101/2020.04.04.024703
(2020).
26.
Zheng, Z., Lauritzen, J. S., Perlman, E., Robinson, C. G., Nichols, M., Milkie, D., Torrens, O., Price, J.,
Fisher, C. B., Sharifi, N.
et al.
A complete el
ectron microscopy volume of the brain of adult
Drosophila
melanogaster.
Cell
174
, 730
-
743 (2018).
27.
Eckstein, N., Bates, A. S., Du, M., Hartenstein, V., Jefferis, G. S. X. E. & Funke, J. Neurotransmitter
classification from electron microscopy images at synaptic sites in Drosophila.
bioRxiv
,
doi.org/10.1101/2020.1106.1112.148775 (2020).
28.
Liu, W. W. &
Wilson, R. I. Glutamate is an inhibitory neurotransmitter in the Drosophila olfactory
system.
Proc Natl Acad Sci U S A
110
, 10294
-
10299 (2013).
29.
Lacin, H., Chen, H. M., Long, X., Singer, R. H., Lee, T. & Truman, J. W. Neurotransmitter identity is
acquir
ed in a lineage
-
restricted manner in the Drosophila CNS.
eLife
8
(2019).
30.
Eschbach, C., Fushiki, A., Winding, M., Schneider
-
Mizell, C. M., Shao, M., Arruda, R., Eichler, K.,
Valdes
-
Aleman, J., Ohyama, T., Thum, A. S.
et al.
Recurrent architecture for ad
aptive regulation of
learning in the insect brain.
Nat Neurosci
23
, 544
-
555 (2020).
31.
Wanner, A. A. & Friedrich, R. W. Whitening of odor representations by the wiring diagram of the
olfactory bulb.
Nat Neurosci
23
, 433
-
442 (2020).
32.
Wittman, T. & Schwe
gler, H.
Biol. Cybern.
73
, 569
-
575 (1995).
33.
Homberg, U., Heinze, S., Pfeiffer, K., Kinoshita, M. & el Jundi, B. Central neural coding of sky
polarization in insects.
Philos Trans R Soc Lond B Biol Sci
366
, 680
-
687 (2011).
34.
Srinivasan, M., Zhang, S.,
Lehrer, M. & Collett, T. Honeybee navigation en route to the goal: visual
flight control and odometry.
J Exp Biol
199
, 237
-
244 (1996).
35.
Vinepinsky, E., Cohen, L., Perchik, S., Ben
-
Shahar, O., Donchin, O. & Segev, R. Representation of
edges, head direction, and swimming kinematics in the brain of freely
-
navigating fish.
Sci. Rep.
10
,
14762 (2020).
36.
Wang, C., Chen, X., Lee, H., Deshmukh,
S. S., Yoganarasimha, D., Savelli, F. & Knierim, J. J.
Egocentric coding of external items in the lateral entorhinal cortex.
Science
362
, 945
-
949 (2018).
37.
Hinman, J. R., Chapman, G. W. & Hasselmo, M. E. Neuronal representation of environmental
boundarie
s in egocentric coordinates.
Nat Commun
10
, 2772 (2019).
38.
Alexander, A. S., Carstensen, L. C., Hinman, J. R., Raudies, F., Chapman, G. W. & Hasselmo, M. E.
Egocentric boundary vector tuning of the retrosplenial cortex.
Sci Adv
6
, eaaz2322 (2020).
39.
So
lstad, T., Boccara, C. N., Kropff, E., Moser, M. B. & Moser, E. I. Representation of geometric
borders in the entorhinal cortex.
Science
322
, 1865
-
1868 (2008).
40.
Savelli, F., Yoganarasimha, D. & Knierim, J. J. Influence of boundary removal on the spatial
representations of the medial entorhinal cortex.
Hippocampus
18
, 1270
-
1282 (2008).
41.
Deshmukh, S. S. & Knierim, J. J. Influence of local objects on hippocampal representations: Landmark
vectors and memory.
Hippocampus
23
, 253
-
267 (2013).
42.
Lever, C.,
Burton, S., Jeewajee, A., O'Keefe, J. & Burgess, N. Boundary vector cells in the subiculum of
the hippocampal formation.
J Neurosci
29
, 9771
-
9777 (2009).
43.
Byrne, P., Becker, S. & Burgess, N. Remembering the past and imagining the future: a neural model
of
spatial memory and imagery.
Psychol Rev
114
, 340
-
375 (2007).
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
10
44.
Bicanski, A. & Burgess, N. A neural
-
level model of spatial memory and imagery.
eLife
7
(2018).
45.
Lyu, C., Abbott, L. F. & Maimon, G. A neuronal circuit for vector computation builds an a
llocentric
traveling
-
direction signal in the Drosophila fan
-
shaped body.
bioRxiv
, see companion manuscript
(2020).
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
11
Online Methods
Fly husbandry and genotypes:
Unless otherwise specified, flies were raised on cornmeal
-
molasses food
(Archon Scientific) in an incubator on a 12
-
hour:12
-
hour light:dark cycle at 25
°C
at 50
-
70% relative humidity.
Flies for the experiments in Figure 2g and E
xtended Figure 5c were cultured on Nutri
-
Fly GF German Food
(Genessee Scientific) with
0
.1%
Te
gosept
(
p
-
h
ydroxy
-
benzoic acid
, Genessee Scientific),
80 mM propionic
acid
(Sigma
-
Aldrich),
and 0.6 mM all
trans
-
retinal (
ATR;
Sigma
-
Aldrich). Vials containing
AT
R
food were
shielded from light with aluminum foil to prevent phot
oconversion of ATR
. The no
-
ATR control flies for
Extended Figure 5c were kept on cornmeal
-
molasses food. All experiments used flies with at least one wild
-
type copy of the
white
gene. Genoty
pes of fly stocks used in each figure are as follows:
Figure 1
PFNd calcium imaging:
w/+; P{R16D01
-
p65.AD}attP40/+; P{R15E01
-
Gal4.DBD}attP2/PBac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
EPG calcium imaging:
w/+; +; P{GMR60D05
-
GAL4}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
PFNd whole
-
cell recording:
w/+; P{R16D01
-
p65.AD}attP40/
P{20XUAS
-
IVS
-
mCD8::GFP}attP40
; P{R15E01
-
Gal4.DBD}attP2/+
Figure 2
SpsP calcium imaging:
w/+; P{VT019012
-
p65.AD}attP40/+; P{R72C10
-
Gal4.DBD}attP2/
P
B
ac{20XUAS
-
IVS
-
jGCaMP
7f}VK00005
LNO2 calcium imaging:
+;
Mi{Trojan
-
p65AD.2}Vglut[MI04979
-
Tp65AD.2]
/+; P{VT008681
-
Gal4.DBD}attP2/
P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
SpsP optogenetic activation
with
PFNd whole
-
cell recording:
w/+; P{GMR16D01
-
lexA}attP40/
P{VT019012
-
p65.AD}attP40;
P{
13xLexAop2
-
IVS
-
pmyr::GFP
}
VK00005,
P{
20xUAS
-
CsChrimson
-
mCherry
-
trafficked
}
su(Hw)attP1
/
P{R72C10
-
Gal4.DBD}attP2
Figure 3:
PFNv calcium imaging:
w/+; P{R22G07
-
p65.AD}attP40/+; P{VT063307
-
Gal4.DBD}attP2/P
B
ac
{20XUAS
-
IVS
-
jGCaMP7f}VK00005
Figure 4:
hΔB calcium imaging:
+; P{R72B05
-
p65.AD}attP40/+; P{VT055827
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
PFNd calcium imaging:
w/+; P{R16D01
-
p65.AD}attP40/+; P{R15E01
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK000
05
Extended Figure 2:
PFNd calcium imaging:
w/+; P{R16D01
-
p65.AD}attP40/+; P{R15E01
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
PFNd whole
-
cell recording:
w/+; P{R16D01
-
p65.AD}attP40/
P{20XUAS
-
IVS
-
mCD8::GFP}attP40
;
P{R15E01
-
Gal4.DBD}attP2/+
Extended Figure
4
:
GFP expression pattern:
+;
Mi{Trojan
-
p65AD.2}Vglut[MI04979
-
Tp65AD.2]
/
P{20XUAS
-
IVS
-
mCD8::GFP}attP40
; P{VT008681
-
Gal4.DBD}attP2/+
MultiColor Flip Out (MCFO):
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
12
+
/ w[1118],
P{R57C10
-
FLPL}su(Hw)attP8
;
Mi{Trojan
-
p65A
D.2}Vglut[MI04979
-
Tp65AD.2]
/+;
P{VT008681
-
Gal4.DBD}attP2/Pbac{10xUAS(FRT.stop)myr::smGdP
-
HA}VK00005,
P{10xUAS(FRT.stop)myr::smGdP
-
V5
-
THS
-
10xUAS(FRT.stop)myr::smGdP
-
FLAG}su(Hw)attP1
Extended Figure
5
:
SpsP optogenetic activation
with
PFNd whole
-
cell rec
ording:
w/+; P{GMR16D01
-
lexA}attP40/
P{VT019012
-
p65.AD}attP40; P{
13xLexAop2
-
IVS
-
pmyr::GFP
}
VK00005,
P{
20xUAS
-
CsChrimson
-
mCherry
-
trafficked
}
su(Hw)attP1
/
P{R72C10
-
Gal4.DBD}attP2
Empty split
-
Gal4
optogenetic activation
control with
PFNd whole
-
cell
recording:
w/+; P{GMR16D01
-
lexA}attP40/
P{p65.AD.Uw}attP40
; P{
13xLexAop2
-
IVS
-
pmyr::GFP
}
VK00005,
P{
20xUAS
-
CsChrimson
-
mCherry
-
trafficked
}
su(Hw)attP1
/
P{GAL4.DBD.Uw}attP2
IbSpsP calcium imaging:
w/+; P{R47G08
-
p65.AD}attP40/+; P{VT012791
-
Gal4.DBD}attP2/
P
B
ac{2
0XUAS
-
IVS
-
jGCaMP7f}VK00005
Extended Figure
6
:
PFNv calcium imaging:
w/+; P{R22G07
-
p65.AD}attP40/+; P{VT063307
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
Extended Figure 7:
LNO
1
calcium imaging:
+;
P{VT020742
-
p65.AD}attP40
/+;
P
{
VT017270
-
GAL4.DBD}attP2
/
P
B
ac{20XUAS
-
IVS
-
jGCaMP7s}VK00005
Extended Figure 8:
hΔB calcium imaging:
+; P{R72B05
-
p65.AD}attP40/+; P{VT055827
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
PFNd calcium imaging:
w/+;
P{R16D01
-
p65.AD}attP40/+; P{R15E01
-
Gal4.DBD}attP2/P
B
ac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
PFNv calcium imaging:
w/+; P{R22G07
-
p65.AD}attP40/+; P{VT063307
-
Gal4.DBD}attP2/PBac{20XUAS
-
IVS
-
jGCaMP7f}VK00005
Origins of transgenic stocks
:
The following stocks were obt
ained from the Bloomington Drosophila Stock Center (BDSC) and published as
follows:
P{GMR60D05
-
GAL4}attP2
(
BDSC 39247)
46
,
P{GMR16D01
-
lexA}attP40 (BDSC 52503)
46
,
P{R72B05
-
p65.AD}attP40 (BDSC
70939)
46
, P{VT055827
-
Gal4.DBD}attP2 (BDSC 71851)
47
,
P{VT008681
-
Gal4.DBD}attP2 (BDSC 73701)
47
, Mi{Trojan
-
p65AD.2}Vglut[MI04979
-
Tp65AD.2] (BDSC 82986)
29
,
Pbac{20XUAS
-
IVS
-
jGCaMP7f}VK00005 (BDSC 79031)
23
, and P{
p65.AD.Uw}attP40;
P{GAL4.DBD.Uw}attP2 (BDSC 79603)
48
.
P{20XUAS
-
IVS
-
mCD8::GFP}attP40 was a gift from B. Pfeiffer and G. Rubin and was
described previously
46
.
The split
-
Gal4 line targeting PFNd neurons was ss00078 (P{R16D01
-
p65.AD}attP40; P{R15E01
-
Gal4.DBD}attP2). The
split
-
Gal4 line targeting SpsP neurons was ss52267 (P{VT019012
-
p65.AD}attP40;
P{R72C10
-
Gal4.DBD}attP2). The split
-
Gal4 line targeting IbSpsP neurons was ss04778 (P{R47G08
-
p65.AD}attP40; P{VT012791
-
Gal4.DBD}attP2). The split
-
Gal4 line targeting PFNv neuron
s was ss52628
(P{R22G07
-
p65.AD}attP40;P{VT063307
-
Gal4.DBD}attP2). The split
-
Gal4 line targeting LNO1 neurons was
ss47398 (
P{VT020742
-
p65.AD}attP40; P
{
VT017270
-
GAL4.DBD}attP2
). These lines were obtained from the
Janelia Research Campus FlyBank and have been
described
previously
3
.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
13
We constructed a split
-
Gal4 line to target LNO2 neurons that incorporates the Vglut
AD
transgene
29
. This split
-
Gal4 line is
Mi{Trojan
-
p65AD.2}Vglut[MI04
979
-
Tp65AD.2]
; P{VT008681
-
Gal4.DBD}attP2. We validated
the expression of this line using immunohistochemical anti
-
GFP staining, and also using Multi
-
Color
-
Flip
-
Out
(MCFO) to visualize single
-
cell morphologies. On occasion, this split line labels a cell typ
e innervating nodulus
subunit 3 (NO3); MCFO results suggest that this is a separate
cell type from LNO2 and does not innervate NO2
(Extended Data Fig. 4)
.
The recombinant
chromosome
P
{
13xLexAop2
-
IVS
-
pmyr::GFP
}
VK00005,
P{
20xUAS
-
CsChrimson
-
mCherry
-
trafficke
d
}
su(Hw)attP1
was a gift from Vivek Jayaraman’s lab.
MCFO experiments used
w[1118],
P{R57C10
-
FLPL}su(Hw)attP8
; +; Pbac{10xUAS(FRT.stop)myr::smGdP
-
HA}VK00005,
P{10xUAS(FRT.stop)myr::smGdP
-
V5
-
THS
-
10xUAS(FRT.stop)myr::smGdP
-
FLAG}su(Hw)attP1 (BDSC
64087
)
49
.
We constructed a split
-
Gal4 line to target
hΔB
neuron
s. This split
-
Gal4 line is
+; P{R72B05
-
p65.AD}attP40;
P{VT055827
-
Gal4.DBD}attP2.
Fly prep
a
ration and dissection
For calcium imaging experiments, we used female flies 20
-
50 hours post
-
eclosion and food
-
deprived (providing
only a Kimwipe with water) for at
least 12 hours prior to the experiment. No circadian restriction was imposed
for the time of experiments. For optogenetic activation experiments in Fig. 2g and Extended Figure 5c, we used
female flies 1
-
5 days post
-
eclosion. Flies were kept on Nutri
-
Fly G
F German Food with 0.6 mM ATR
. For all
other electrophysiology experiments, we used female flies 24
-
48 hours old
; 5/7 flies included in our dataset
were food
-
deprived for 12
-
24 hours
. No circadian restriction was imposed for the time of experiments.
Flies
were briefly cold anesthetized prior to dissection. For calcium imaging experiments and electrophysiology
experiments during walking behavior, we secured the fly in an inverted pyramidal platform CNC
-
machined
from black Delrin (Autotiv, Protolabs) with th
e head pitched forward so that the posterior surface of the head
was more accessible to the microscope objective. For electrophysiology experiments with optogenetic
activation, we used a photochemically
-
etched, flat stainless
-
steel shim stock platform (Etc
hit), and the head was
oriented normally (dorsal
-
side up). The wings were removed, and the fly head and thorax were secured to the
holder using UV
-
curable glue (
Loctite AA 3972) and cured
with
ultraviolet light (LED
-
200, Electro
-
Lite Co
).
To remove large b
rain movements, the proboscis was glued using UV
-
curable glue. The extracellular saline
composition was:
103
mM NaCl, 3 mM KCl, 5 mM TES, 8 mM trehalose, 10 mM glucose, 26 mM NaHCO
3
, 1
mM NaH
2
PO
4
, 1.5
mM CaCl
2
, and 4 mM MgCl
2
(osmolarity 270
-
275 mOsm)
. The
saline was
bubbled with
95% O
2
and 5% CO
2
to reach
a final pH of
~
7.3
. A window was opened in the head cuticle, and trachea and fat
were removed to expose the brain. To further reduce brain movement, muscle 16 was inactivated by gently
tugging or clipping
the esophagus posteriorly, or by clipping the muscle anteriorly. For electrophysiology
experiments, the perineural sheath was removed with fine forceps over the brain region of interest. For all
electrophysiology experiments, saline was continuously super
fused over the brain; for calcium imaging, saline
was superfused prior to experiments.
Two
-
photon calcium imaging
We used a galvo
-
galvo
-
resonant two
-
photon microscope (Thorlabs Bergamo II, Vidrio
RMR Scanner) with a
fast piezoelectric objective scanner (
Physik Instrument P725
)
and a
20×/1.0
NA objective
(
XLUMPLFLN20XW
, Olympus)
for volumetric imaging. We used a C
hameleon Vision
-
S Ti
-
Sapphire
femtosecond laser tuned to 940 nm for two
-
photon GCaMP e
xcitation. Emission was collected on GaAsP PMT
detectors (Hamamatsu) through a 525
-
nm bandpass filter (Thorlabs). We used
Scan
I
mage 2018
software
50
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
14
(Vidrio Technologies) to control the microscope, and imaging data were collected in ScanImage using
National
Instruments
PXIe
-
6341 hardware.
The imaging region for all
experiments was 256x128 pixels, with 12 slices in the z
-
axis for each volume (3
-
5
μm
per slice) resulting in a ~10 Hz volumetric scanning rate. For PFNd and PFNv imaging experiments, we
imaged the PB. For SpsP and IbSpsP imaging experiments, we imaged the
PB. For LNO2 imaging experiments,
we imaged the NO
.
For hΔB imaging experiments, we imaged the FB.
Patch
-
clamp recordings
Thick
-
wall filamented borosilicate glass (
OD 1.5, ID 0.86 mm, Sutter
) pipettes with
a resistance range of 9
-
12
M
Ω
were pulled using
a P
-
97 Sutter puller. Pipettes were filled with an internal solution
51
consisting of
140
mM
KOH, 140
mM aspartic acid, 1
mM KCl, 10
mM HEPE
S, 1 mM EGTA, 4 mM MgATP, 0.5 mM Na
3
GTP, and
13
mM
b
iocytin hydrazide
,
filtered twice through a 0.22
-
μm
PVDF
filter
. To visualize the
cells for recording,
we used a FLIR camera (Chameleon3 CM
-
U3
-
13Y3C) mounted on an upright compound microscope (Olympus
BX51WI) with a 40× water immersion objective (LUMPlanFLN 40XW, Olympus). We used a 100
-
W Hg arc
lamp (Olympus, U
-
LH100HG) and an eGFP lon
g
-
pass filter to detect GFP fluorescence. For optogenetics
experiments, the brain was illuminated from below using bright field transmitted light through the microscope
condenser to identify cell bodies for recording, which was then turned off prior to opt
ogenetic stimulus delivery.
For walking experiments, the fly was illuminated from below using a fiber optic coupled LED (M740F2,
Thorlabs) coupled to a ferrule
-
terminated patch cable (200
-
μM core
, 0.22 n.a.
, Thorlabs
) attached to a fiber
optic cannula (200
-
μM core, 0.22 n.a.
, Thorlabs
). The cannula was glued to the ventral side of the holder and
positioned approximately 135° from the front of the fly so as to be unobtrusive to the fly’s visual field.
Throughout the experiment, saline
bubbled with 95% O
2
/ 5
% CO
2
was superfused over the fly using a gravity
pump at a rate of 2 mL/min. Whole cell recordings were performed using an Axopatch 200B amplifier with a
CV
-
203BU headstage (Molecular Devices)
. Data were low
-
pass filtered at 5 kHz and acquired on a
NiDAQ
PCIe
-
6363 card
(National Instruments)
at 20
kHz
. The l
iquid junction
potential
was corrected
by subtracting 13
mV from recorded voltages
52
.
Spherical treadmill and locomotion measurement
For calcium imaging experiments, flies were positioned on a 9
-
mm ball made from foam (FR
-
4615, General
Plastics). The ball was painted with a bla
ck pattern using model paint (Vallejo Black Model Color Paint). The
spherical treadmill consisted of this ball floating on air in a concave hemispherical depression on a plenum 3
-
D
printed from clear acrylic (Autotiv). Medical
-
grade breathing air was flowe
d through a hole at the bottom of the
depression. The ball was illuminated with a
round
-
board 36 infrared LED lamp (SODIAL)
. Ball movement was
tracked using a video
camera (CM3
-
U3
-
13Y3M
-
CS, FLIR) fitted with a macro zoom lens (
Tamron 23FM08L
8
-
mm
1:1.4 lens
). The camera faced the ball from the right side of the fly at a 90° angle. We removed one panel
of the visual panorama to accommodate the camera view of the ball. The camera frame rate was 50 Hz.
Machine vision software (FicTrac v2.0) was used t
o track the position of the ball
53
. We modified FicTrac to
output computed ball position parameters in real time through the Redis pub
lish/subscribe messaging paradigm.
We wrote custom Python software to read in FicTrac outputs from Redis and to produce analog voltage signals
through a Phidget analog output device (Phidget Analog 4
-
Output 1002_0B). The forward axis ball
displacement, yaw
axis ball displacement, gain
-
modified forward ball displacement (not used for experiments in
this study), and gain
-
modified yaw ball displacement were output through the Phidget analog device. For
closed
-
loop experiments, the gain
-
modified yaw ball displa
cement voltage signal was used to update the
azimuthal position of the visual cues displayed by the visual panorama. All voltage analog signals were
digitized and acquired using
NiDAQ PCI
-
6341 (National Instruments) at 4 kHz
.
The forward axis, lateral axis
,
and yaw axis ball movements from FicTrac, along with their timestamps, were recorded by the custom Python
software and saved to a HDF5 file for each experiment.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint
15
For electrophysiology experiments,
the following parameters were altered. The ball was illum
inated using a 780
nm mounted LED source (M780L3, Thorlabs). The ball’s movement was tracked using
GS3
-
U3
-
41C6NIR
video camera (FLIR) fitted with an InfiniStix 94
-
mm 0.5× macro zoom lens. One panel 180° behind the fly was
removed to accommodate the camera
view of the ball and the light source. FicTrac v2.1 was used to track the
position of the ball in real time
53
. We recorded the forward,
side, and yaw displacement of the ball via a
NiDAQ
PCIe
-
6363 card at 20
kHz
. Via built
-
in serial communication support, we used a custom Python script to output
FicTrac parameters to a Phidget analog output device (Phidget Analog 4
-
Output 1002_0B).
Visua
l panorama
and visual stimuli
To display visual stimuli, we used a circular panorama built from modular square (8
×
8 pixel) LED panels
24
.
The circular arena was 12 panels in circumference and 2 panels tall. For calcium imaging experiments, we
removed one panel 90° to the right of the fly; the bottom panel at that azimuth remained to display stimuli. For
electrophysiology experiments, we rem
oved one panel 180° behind the fly. In all experiments, the modular
panels contained blue LEDs with peak blue (470 nm) emission; blue LEDs were chosen to reduce overlap with
the GCaMP emission spectrum. For calcium imaging experiments, four layers of filte
rs were added in front of
the LED arena (Rosco, R381) to further reduce overlap in spectra. A final diffuser layer was placed in front of
the filters (SXF
-
0600, Snow White Light Diffuser, Decorative Films). For electrophysiology experiments, only
the diffu
ser layer was used.
The visual stimulus displayed was a bright 2
-
pixel
-
wide vertical bar. The bar’s height was the full 2
-
panel
height of the area (except for 75
-
105° to the right of the fly, when the bar was 1 full panel in height). The
azimuth position
of the bar was controlled during closed
-
loop experiments via the voltage signal from the
Phidget device, which was used to convert FicTrac outputs to an analog voltage signal. For calcium imaging
experiments, a 0.8× yaw gain was used; this meant that for a
given yaw displacement of the ball, the visual cue
displacement was 0.8× the ball’s yaw displacement. For electrophysiology experiments, a 1× yaw gain was
used.
The visual panorama provided no information about body velocity. It therefore seems likely th
at the velocity
signals we observed were due to proprioceptive feedback and/or motor efference copy. However, it is possible
that flies could see the pattern on the spherical treadmill moving as they walked, and thus there may also be a
contribution from v
entral optic flow signals.
Experimental
trial
structure
during calcium imaging
For calcium imaging experiments, prior to data collection, all flies were allowed to walk for 5 min in darkness
and then at least 10 min in closed loop with the visual cue. Fo
r calcium imaging experiments, data were
collected in two 300
-
s trials in closed loop with a bright bar; there was a 5
-
s interval of darkness between trials.
For electrophysiology experiments, flies were given at least 10 minutes of walking in closed loop
with the
visual cue prior to data collection. Each electrophysiology experiment consisted of 3 continuous 200
-
s closed
loop trials with a 1
-
s inter
-
trial interval in darkness.
O
ptogenetic
stimuli and pharmacology
Optogenetic stimuli were delivered using a
Hg lamp and an
ET
-
Cy5 long
-
pass filter (590
-
650nm, Chroma)
,
with a power of ~10 mW/mm
2
. A shutter
(Uniblitz Electronic)
was used to control the light pulse duration.
Light pulses (10
ms) were delivered at 4
-
s inter
-
pulse intervals, in three sessions of 15
0 pulses each. In the first
session, the extracellular saline contained
1
μM TTX
(554412, EMD Biosciences)
. In the second session,
1μM
p
icrotoxin (CAS 124
-
87
-
8, Sigma Aldrich)
was added. In the third session, picrotoxin was increased to
1
00
μM
.
In no
-
ATR c
ontrol experiments, the light pulse was 50 ms long.
.
CC-BY-NC-ND 4.0 International license
available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint
this version posted December 23, 2020.
;
https://doi.org/10.1101/2020.12.22.424001
doi:
bioRxiv preprint