The function of auditory neurons in cricket phonotaxis i. influence of hyperpolarization of identified neurons on sound localization


Schildberger, K.; Hoerner, M.

Journal of Comparative Physiology A Sensory Neural and Behavioral Physiology 163(5): 621-632

1988


In order to examine the role of particular identified auditory neurons of the cricket, Gryllus bimaculatus, in orientation to a sound source, a method has been developed by which intracellular recordings can be made while the animal walks on an air-suspended sphere, which is rotated by the leg movements (Fig. 1). The angular velocities of sphere rotation were found to depend on the direction of incident sound, on its intensity and frequency and on the temporal pattern of the sound stimulus (Figs. 2, 3). While the cricket was walking, auditory neurons discharged extra action potentials, not correlated with the sound stimulus, and the neuronal response to the sound itself was reduced (Figs. 4, 5). Suppressing the spike activity by hyperpolarization of a local neuron in the prothoracic ganglion (ON1) reduced in some animals the tendency to turn toward the sound source on the side of the ear that excites the ON1 (Figs. 6-8). Hyperpolarization of a neuron that ascends from the prothoracic ganglion into the brain (AN1), while sound was presented to the ear that excites this neuron, caused all animals to reverse direction; that is, they turned away form the sound source and from the side of the inactivated AN1 (Figs. 9, 10). Hyperpolarization of another ascending neuron (AN2) caused a reduction in turning velocity in half of the animals; but this effect occurred only with high sound pressure levels, and the direction of walking was not reversed (Figs. 11, 12). From the influences on turning tendency observed in these experiments, it appears that the paired AN1s (and possibly the AN2s at high intensities) may provide inputs to a central compoarator that dictates turning tendency in phonotaxis.

J
Comp
Physiol
A
(1988)
163:621-631
Journal
of
sory,
Comparative
Sen
and
Physiology
A
t'ylatr:
©
Springer-Verlag
1988
The
function
of
auditory
neurons
in
cricket
phonotaxis
I.
Influence
of
hyperpolarization
of
identified
neurons
on
sound
localization
Klaus
Schildberger
i
and
Michael
Horner
2
1
Max-Planck-Institut
fiir
Verhaltensphysiologie,
D-8131
Seewiesen,
Federal
Republic
of
Germany
2
L
Zoologisches
Institut
der
Universitat,
Abteilung
fiir
Zellbiologie,
D-3400
Gottingen,
Federal
Republic
of
Germany
Accepted
March
25,
1988
Summary.
In
order
to
examine
the
role
of
particu-
lar
identified
auditory
neurons
of
the
cricket,
Gryl-
lus
bimaculatus,
in
orientation
to
a
sound
source,
a
method
has
been
developed
by
which
intracellu-
lar
recordings
can
be
made
while
the
animal
walks
on
an
air-suspended
sphere,
which
is
rotated
by
the
leg
movements
(Fig.
1).
The
angular
velocities
of
sphere
rotation
were
found
to
depend
on
the
direction
of
incident
sound,
on
its
intensity
and
frequency
and
on
the
temporal
pattern
of
the
sound
stimulus
(Figs.
2,
3).
While
the
cricket
was
walking,
auditory
neu-
rons
discharged
extra
action
potentials,
not
corre-
lated
with
the
sound
stimulus,
and
the
neuronal
response
to
the
sound
itself
was
reduced
(Figs.
4,
5).
Suppressing
the
spike
activity
by
hyperpolari-
zation
of
a
local
neuron
in
the
prothoracic
gangli-
on
(ON1)
reduced
in
some
animals
the
tendency
to
turn
toward
the
sound
source
on
the
side
of
the
ear
that
excites
the
ON1
(Figs.
6-8).
Hyperpo-
larization
of
a
neuron
that
ascends
from
the
pro-
thoracic
ganglion
into
the
brain
(AN1),
while
sound
was
presented
to
the
ear
that
excites
this
neuron,
caused
all
animals
to
reverse
direction;
that
is,
they
turned
away
from
the
sound
source
and
from
the
side
of
the
inactivated
AN1
(Figs.
9,
10).
Hyperpolarization
of
another
ascending
neu-
ron
(AN2)
caused
a
reduction
in
turning
velocity
in
half
of
the
animals;
but
this
effect
occurred
only
with
high
sound
pressure
levels,
and
the
direction
of
walking
was
not
reversed
(Figs.
11,
12).
From
the
influences
on
turning
tendency
ob-
served
in
these
experiments,
it
appears
that
the
paired
AN1s
(and
possibly
the
AN2s
at
high
inten-
sities)
may
provide
inputs
to
a
central
comparator
that
dictates
turning
tendency
in
phonotaxis.
Introduction
Analyses
of
the
neuronal
mechanisms
underlying
behavior
are
usually
based
on
correlations
between
certain
kinds
of
behavior
and
the
properties
of
neu-
rons
and
their
connections.
For
causal
analysis,
however,
the
behavior
should
be
modified
by
the
direct
manipulation
of
the
physiological
activity
of
identified
neurons.
Orthopteran
insects
have
re-
cently
been
used
for
such
studies;
the
activity
of
identified
neurons
has
been
recorded
during
flight,
walking
and
stridulation
(Robertson
and
Pearson
1982;
Nolen
and
Hoy
1984;
Godden
and
Graham
1984;
Kien
1983;
Hedwig
1986).
Here
we
extend
this
approach
to
the
auditory
system
of
crickets
during
phonotactic
walking.
Female
crickets
give
positive
phonotactic
re-
sponses
to
the
calling
song
of
conspecific
males;
that
is,
they
track
the
song
when
free
and
show
turning
tendency
when
tethered.
This
behavior
can
be
measured
quantitatively
to
obtain
information
about
sound
localizing
ability
and
its
limits,
and
about
the
female
cricket's
criteria
for
recognition
of
the
conspecific
signal
(Weber
et
al.
1981;
Thor-
son
et
al.
1982;
Schmitz
et
al.
1982;
Stout
et
al.
1983;
Huber
and
Thorson
1985).
Central
processing
of
the
conspecific
song
ini-
tially
occurs
in
the
prothoracic
ganglion;
several
auditory
neurons
with
cell
bodies
located
here
have
been
identified
by
physiological
and
morphological
criteria
(Casaday
and
Hoy
1977;
Popov
et
al.
1978;
Wohlers
and
Huber
1978,
1982;
Moiseff
and
Hoy
1983;
Stout
et
al.
1985).
All
the
auditory
neurons
that
have
been
identi-
fied
are
present
as
mirror-image
pairs
with
cell
bodies
in
opposite
halves
of
the
ganglion,
and
each
member
of
the
pair
is
affected
differently
by
the
two
ears.
In
the
search
for
mechanisms
of
sound
localization,
interactions
involving
the
members
of
!Th
Rotation
0 0 0
0
0
0
0 0
0
1
3
0
RS
Sound
Gener
Tope
622
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
these
pairs
either
directly
or
by
way
of
higher
level
comparisons
have
been
among
the
chief
candidates.
Because
the
results
of
the
experiments
presented
here
will
be
interpreted
in
terms
of
such
interactions,
a
brief
introduction
to
our
terminolo-
gy
is
in
order.
The
omega
neuron
ON1
has
a
den-
dritic
field
on
the
same
side
as
its
cell
body
and
an
axon
that
crosses
to
the
other
side
of
the
gangli-
on;
it
is
excited
by
the
ear
on
the
same
side
as
its
cell
body
and
dendrites
and
inhibited
(indirect-
ly)
by
the
ear
on
the
other
side
(Wohlers
and
Huber
1982).
The
ascending
neurons
AN1
and
AN2,
which
have
their
cell
bodies
on
one
side
of
the
ganglion
and
their
dendritic
fields
and
axons
on
the
other,
are
excited
by
the
ear
on
the
same
side
as
the
dendrites;
the
ear
on
the
opposite
side
has
no
effect
on
AN1
and
inhibits
(or,
in
some
cases,
weakly
excites)
AN2.
Obviously,
the
'laterality'
of
these
neurons
is
ambiguous.
In
the
original
de-
scription
(Wohlers
and
Huber
1982)
the
`side'
of
a
neuron
was
determined
by
the
position
of
its
cell
body;
the
left
neuron
was
the
one
with
its
cell
body
on
the
left.
But
because
neurons
of
these
three
types
are
excited
by
the
ear
on
the
same
side
as
their
dendrites,
regardless
of
the
position
of
the
cell
body,
for
the
present
purposes
it
is
more
useful
to
adopt
a
different
convention.
Here
we
regard
a
sound
source
as
ipsilateral
to
a
neuron
if
the
neuron
is
excited
by
the
ear
on
that
side
that
is,
if
the
sound
source
is
ipsilat-
eral
to
the
dendrites
of
the
neuron.
The
directional
characteristic
of
the
ears
is
reflected
in
the
activity
of
these
central
neurons
and
further
enhanced
by
contralateral
inhibition
(Wiese
and
Eilts
1985;
Kleindienst
et
al.
1981;
Selverston
et
al.
1985).
The
omega
neuron
ON1
is
thought
to
have
the
special
function
of
binaural
contrast
enhancement.
In
this
paper
we
present
a
method
that
enables
intracellular
recording
from
these
neurons
while
the
animal
is
engaged
in
phonotactic
walking
(Fig.
1).
Using
it,
we
can
observe
whether
the
known
response
characteristics
of
the
neurons
are
altered
during
the
behavioral
responses
to
sound.
When
recording
from
the
integrating
segment
it
is
also
possible
to
inactivate
the
cells
reversibly
by
injecting
negative
current
and
to
check
for
asso-
ciated
changes
in
walking.
With
such
measures
of
the
influence
of
a
particular
neuron
on
motor
per-
formance,
one
can
determine
the
manner
and
de-
gree
of
involvement
of
various
neurons
in
the
be-
havior.
Materials
and
methods
Females
of
the
gryllid
species
G.
bimaculatus,
raised
in
the
labo-
ratory,
were
acoustically
isolated
in
the
last
instar
and
for
2—
Fig.
1.
Diagram
of
the
experimental
design
for
intracellular
re-
cording
from
neurons
in
the
prothoracic
ganglia
during
walk-
ing.
The
animal
is
fixed
to
a
holder
but
can
turn
an
air-sus-
pended
styrofoam
sphere
by
its
walking
movements.
The
in-
tended
translational
(Tra)
and
rotational
(Rot)
components
are
recorded
separately.
Two
speakers
are
mounted
50°
to
the
left
(LS)
and
right
(RS)
of
the
long
axis
of
the
animal
(not
drawn
to
scale)
4
weeks
after
the
imaginal
moult,
until
they
were
taken
for
the
experiment.
The
cricket
is
attached
to
a
holder
so
that
the head,
thorax
and
abdomen
are
immobilized;
the
legs,
which
can
move
freely,
are
in
contact
with
a
hollow
styrofoam
sphere
(diameter
12
cm,
weight
5
g;
see
also
Dahmen
1980)
supported
in
an
airstream.
As
the
animal
carries
out
walking
movements,
it
turns
the
sphere
in
a
direction
opposite
to
the
intended
walk-
ing
direction
(Fig.
1).
The
rotation
of
the
sphere
is
measured
with
a
camera
nearly
identical
to
the
one
used
in
the
walking
compensator
(see
Kramer
1975;
Weber
et
al.
1981).
Reflecting
dots
(diameter
2
mm)
are
glued
to
the
sphere
at
intervals
smaller
than
the
diameter
of
the
camera's
field
of
view
(30
mm).
The
camera
monitors
the
motion
of
the
dot
nearest
to
the
lower
left
edge
of
this
measurement
area,
and
extracts
the
x
and
y
components
of
the
motion.
Because
the
measurement
area
is
situated
on
the
equator
of
the
sphere
and
in
front
of
the
animal,
these
data
specify
the
translational
and
rotational
components
of
the
walk.
When
one
point
drifts
out
of
the
measurement
field,
the
monitor
switches
to
another;
these
discontinuities
are
elimi-
nated
automatically
by
the
data
processing
system.
The
measurement
system
operates
with
a
temporal
resolu-
tion
of
10
ms,
a
spatial
resolution
of
0.3
mm,
and
a
linear
range
of
velocities
from
0.3
to
about
100
cm/s.
To
find
the
total
path
distance
represented
by
the
sphere's
rotation,
all
movements
in
each
100
ms
segment
were
added
up
separately
for
the
two
axes
of
rotation,
and
from
these
sums
the
total
path
length,
the
velocities
and
their
distributions
were
calculated;
parametric
statistical
tests
(t-test,
F-test)
were
applied
to
compare
velocity
distributions.
Two
loudspeakers
were
situated
30
cm
away
from
the
ani-
mal,
50°
to
the
right
and
left
of
the
long
axis
of
the
body.
The
frequency
characteristics
of
the
speakers
were
equalized
over
the
range
of
3-20
kHz.
The
sound
field
was
not
homoge-
neous
because
of
the
presence
of
sound-reflecting
objects,
but
the
signals
near
5
kHz
could
be
equalized
in
the
vicinity
of
the
ears.
Echos
in
this
frequency
range
were
more
than
30
dB
weaker
than
the
signal.
The
sound
stimulus
was
an
artificial
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
623
tra
ns
la
tion
im
l
calling
song
(chirps
of
4
syllables
20
ms
in
duration,
separated
by
20
ms
pauses;
chirp
repetition
interval
500
ms),
at
a
carrier
frequency
that
could
be
varied
between
2
and
20
kHz
and
an
intensity
between
45
and
90
dB
SPL
(RMS).
The
dorsal
surface
of
the
animal
was
opened
and
the
sides
of
the
thorax
pulled
slightly
apart.
The
prothoracic
ganglion
was
exposed
by
removal
of
the
dorsal
musculature
and
the
gut;
the
lateral
and
ventral
tracheae
(including
the
acoustic
trachea)
remained
intact,
as
did
the
nerves
from
the
ganglion
and
the
connectives.
The
ganglion
was
stabilized
mechanically
by
two
spoons,
one
below
serving
as
a
platform
and
the
other
above
with
a
hole
for
the
electrode
insertion.
A
glass
microelectrode
filled
with
3%
Lucifer
Yellow
was
inserted
into
the
ganglion
from
the
dorsal
surface
and
penetrated
neurons
in
the
auditory
neu-
ropil.
Acoustic
stimulation
revealed
the
type
of
the
neuron
pen-
etrated;
e.g.
receptor
cells
and
interneurons
were
distinguished
by
their
latency,
threshold
curves
were
used
to
discriminate
between
low-frequency
(AN1,
ON1)
and
high-frequency
neu-
rons
(AN2),
and
the
two
neuron
types
AN1
and
ON1
clearly
copied
the
temporal
pattern
of
the
song
more
precisely
than
other
neurons.
For
confirmation
of
the
neuron
type
all
cells
were
marked
intracellularly
with
Lucifer
Yellow;
after
the
ex-
periment
the
ganglia
were
processed
histologically
by
the
con-
ventional
procedures
and
photographed
as
a
whole
mount.
It
was
possible
to
record
intracellularly
for
as
long
as
one
hour.
An
experiment
was
accepted
for
evaluation
only
if
the
recording
was
stable
for
at
least
15
min,
the
animal
exhibited
an
unambiguous
phonotactic
response
during
recording,
and
the
injection
of
1.5
to
3
nA
of
negative
current
caused
a
clear
reduction
in
spike
activity.
For
the
latter
reason
the
recording
site
was
always
the
integrating
segment
where
graded
potentials
and
spikes
are
observable.
The
neuronal
activity
was
recorded
simultaneously
with
the
sphere
movement
and
evaluated
with
the
aid
of
a
computer.
Results
Behavior
Upon
presentation
of
the
calling
song
from
one
speaker
the
crickets,
with
the
body
immobilized,
made
leg
movements
that
would
have
tended
to
turn
them
toward
the
side
of
the
active
speaker.
Correspondingly,
the
free
sphere
counterrotated.
When
the
song
was
switched
to
the
other
speaker,
they
reversed
their
turning
tendency
after
1-
2
chirps.
This
reversal
can
occur
with
no
pause
in
walking,
especially
if
the
sound
is
loud.
In
their
phonotactic
response
these
animals
are
as
selective
for
the
temporal
pattern
of
the
song
as
freely
walk-
ing
animals
(see
Thorson
et
al.
1982).
Songs
with
syllable
intervals
equal
to
or
less
than
20
ms,
or
equal
to
or
greater
than
60
ms,
elicited
slower
walking
or
no
response
at
all
(Fig.
2).
The
mounting
and
dissection
procedures
can
affect
an
animal's
behavior;
in
some
cases
phono-
taxis
was
eliminated
altogether
and
in
others
the
turning
tendency
was
extremely
asymmetrical.
Such
experiments
have
been
excluded.
In
all
the
experiments
described
below,
during
the
recording
2
1
2
rotation
Irn)
xxw
4111101111111111.
.1-1111-11HIHIF
-01
-
10
-
11111
-
10
20
4
0
60
100
Fig.
2.
Reconstruction
of
the
path
of
a
cricket
walking
in
the
presence
of
various
sounds.
The
translational
(tra)
and
rota-
tional
(rot)
components
are
plotted
additively
on
the
ordinate
and
abscissa,
respectively.
At
each
starting
point
the
sound
is
coming
from
the
left
speaker.
Every
30
s
thereafter
it
was
switched
from
one
speaker
to
the
other
(arrows).
The
temporal
pattern
of
the
sound
stimulus
is
shown
below
the
abscissa
corre-
sponding
to
the
respective
walking
trace
(syllable
repetition
in-
tervals
are
10,
20,
40,
60
and
100
ms).
When
the
pattern
is
attractive
(40
ms
syllable
repetition
interval),
the
animal
system-
atically
attempts
to
turn
toward
the
active
speaker,
changing
its
turning
direction
at
each
speaker
switch.
The
two
middle
traces,
showing
the
phonotactic
behavior
before
preparation
for
recording
and
during
recording
from
a
neuron,
do
not
differ
significantly.
Intensity
(80
dB
SPL)
and
timing
of
speaker
switches
were
identical
in
all
6
tests
of
interneuronal
activity
the
cricket's
walking
be-
havior
met
the
criteria
for
phonotaxis
applied
here:
significant
continuous
turning
tendency
toward
the
active
speaker
for
at
least
2
min
and
change
of
turning
tendency
following
change
of
speaker.
Most
of
the
animals
occasionally
interrupted
their
phonotactic
walking
for
a
few
seconds,
al-
though
some
walked
continuously
for
over
20
min.
The
duration
and
frequency
of
the
pauses
were
like
those
found
for
tracking
behavior
on
the
walk-
ing
compensator
(Weber
et
al.
1981;
Schmitz
et
al.
1982).
If
the
pauses
lasted
longer
than
10
s,
phono-
taxis
was
considered
to
have
stopped;
after
such
a
long
pause
the
turning
tendency
no
longer
de-
pended
on
speaker
position
and
was
not
altered
by
a
change
in
direction
of
the
sound
source.
Crickets
oriented
less
readily,
the
longer
an
experi-
ment
lasted.
Even
animals
that
were
initially
re-
sponsive
stopped
responding
within
an
hour
after
preparation
for
the
experiment
was
completed.
The
translatiorial
velocities
varied
(both
among
animals
and
in
a
given
animal)
between
1
and
10
cm/s.
The
rotational
velocities
averaged
27
deg/
s
at
80
dB,
but
occasionally
velocities
as
high
as
100
deg/s
were
observed.
When
sound
was
pre-
sented
on
one
side,
the
velocity
of
turning
toward
during
recording
intact
are
5
chirps
111111111_1
1
II
I
1111
o
-
1 I
I 1
mV
translational
velocity
5
-
cm
/s
0.
rotational
velocity
40
-
deg
Is
0-
1
624
K.
Schildberger
and
M.
Homer:
Function
of
cricket
auditory
neurons
90dB
80
dB
70dB
left
ON1
10
time
[Si
active
speaker
DLL
C
RL
60dB
RS
LS
0
S
120-
0
0
0
40
C
-
-50
0
50
deg
/s
Fig.
3.
Distribution
of
angular
velocities
during
stimulation
with
calling
song
at
different
intensities
from
the
right
(RS)
or
left
(LS)
speaker.
Positive
values
indicate
turning
to
the
right,
negative
to
the
left
(as
in
all
following
figures).
Under
each
stimulus
condition
220
chirps
were
presented
(
=110
s).
The
angular
velocities
were
measured
while
recording
from
the
left
ON1.
The
mean
angular
velocity
increased
between
60
and
80
dB
and
decreased
again
at
90
dB
that
side
varied
greatly
even
though
the
sound
in-
tensity
was
constant;
furthermore,
even
during
clear
phonotaxis,
turns
away
from
the
speaker
oc-
curred
(Fig.
3).
In
general,
turning
velocity
de-
pended
on
sound
intensity.
Sound
intensities
below
55
dB
evoked
no
significant
turning
tendencies
to-
ward
the
side
of
the
active
speaker.
In
the
range
of
60-80
dB,
the
mean
turning
velocity
of
most
animals
rose
to
25-35
deg/s.
With
higher
intensities
there
was
no
further
increase
in
angular
velocity,
and
those
of
many
animals
decreased
(Fig.
3).
Neuronal
responses
As
long
as
the
crickets
were
not
walking,
the
re-
sponses
of
the
auditory
neurons
in
the
prothoracic
ganglion
to
conspecific
calling
song
were
no
differ-
ent
from
those
recorded
in
a
totally
immobilized
animal.
For
instance,
the
threshold
of
the
omega
Fig.
4.
Intracellular
recording
from
the
left
ON1
during
walking
in
the
presence
of
calling
song
(70
dB).
First
trace
(from
top):
intracellular
recording
of
ON1
activity;
2nd
trace:
marks
indi-
cating
the
sound
stimulus;
3rd
trace:
translational
velocity,
in
which
positive
values
indicate
forward
walking
and
negative
values
backward
(as
in
all
the
following
figures);
4th
trace:
angular
velocity
(positive
to
the
right,
negative
to
the
left);
letters
at
the
bottom
indicate
the
active
speaker.
Data
in
Figs.
4-6
come
from
the
same
preparation.
During
walking
the
auditory
response
of
the
neuron
is
no
longer
clearly
apparent
neuron
ON1
to
tones
at
5
kHz
was
about
40-45
dB
for
ipsilateral
presentation
and
10-15
dB
higher
for
contralateral
presentation.
Auditory
protho-
racic
neurons
have
little
or
no
'
spontaneous
'
activ-
ity,
and
encode
the
song
pattern
in
the
manner
characteristic
of
their
type
(see
below).
When
the
crickets
were
walking,
two
additional
effects
ap-
peared.
First,
action
potentials
uncorrelated
with
the
sound
stimulus
were
discharged
and,
second,
the
responses
to
the
calling
song
developed
gaps
that
could
become
so
large
that
the
response
disap-
peared
altogether,
especially
for
contralateral
stim-
ulation
(Figs.
4,
5).
However,
if
the
intensity
of
the
song
was
higher
than
75
dB,
there
was
always
a
clear
response
even
during
walking.
The
origin
and
possible
significance
of
these
effects
are
con-
sidered
in
a
subsequent
paper
(Schildberger
et
al.
1988).
Inactivation
of
ON1
The
omega
neuron
ON1
is
located
in
the
protho-
racic
ganglion.
It
is
tuned
to
the
carrier
frequency
of
the
conspecific
calling
song
and
copies
the
tem-
translational
velocity
left
ON
1
C.M/S
1
5-
rotational
velocity
40-
deg
I
s
0
rotation
2
4
time
lminI
active
speaker
LS
LSH
LS
RS
LS
LSH
I
I
AP
20
I
I
U
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
625
cm
rotational
velocity
deg/s
20
-
0-
rotation
cm
7
no.
of
action
potentials
N
10-
15
-
dot
display
of
action
potentials
ms
no.
of
action
potentials
1
4
0
active
speaker
100-
g
I
.
[mild
I
iiiIIIIIII
20
time
[51
RS
LS
Fig.
5.
Intracellular
recording
from
the
left
ON1
during
walking
in
the
presence
of
calling
song
(80
dB).
Top
trace:
angular
ve-
locity;
2nd
trace:
rotational
component
(integrated
over
time),
with
upward
deflection
indicating
an
intended
right
turn
and
downward
a
left
turn
(as
in
all
following
figures);
3rd
trace:
number
of
action
potentials
in
the
first
200
ms
after
the
begin-
ning
of
each
chirp
(as
in
all
following
figures);
4th
trace:
neuro-
nal
discharge,
in
which
each
dot
represents
a
spike,
timing
of
the
spikes
within
a
500
ms
cycle
(chirp
interval)
on
the
ordinate,
the
successive
chirps
on
the
abscissa;
the
active
speaker
is
indi-
cated
at
the
bottom.
This
30
s
sequence
is
part
of
the
longer
sequence
shown
in
Fig.
6
(loudspeaker
switch
after
the
4th
min).
With
contralateral
sound
representation
(RS),
so
that
the
ear
that
excites
this
neuron
was
less
strongly
stimulated
than
the
other
ear,
the
neuronal
response
to
the
sound
during
walking
was
more
strongly
affected
than
with
ipsilateral
presentation
poral
pattern
of
that
song
(Wohlers
and
Huber
1982).
It
receives
excitatory
input
from
the
ear
ipsi-
lateral
to
the
cell
body
and
is
inhibited
by
activity
from
the
contralateral
ear.
This
inhibition
is
me-
diated
by
its
mirror-image
partner,
the
ON1
on
the
opposite
side
(Selverston
et
al.
1985).
Because
of
these
properties,
the
difference
in
level
of
excitation
at
the
two
ears
is
increased
at
Fig.
6.
Intracellular
recording
from
the
left
ON1
during
walking
in
the
presence
of
calling
song
(80
dB).
Top
trace:
translational
velocity;
2nd
trace:
angular
velocity;
3rd
trace: rotational
com-
ponent,
4th
trace:
number
of
action
potentials;
the
active
speaker
(RS
or
LS)
and
the
injection
of
hyperpolarizing
current
(H) are
indicated
at
the
bottom.
During
injection
of
negative
current
the
turning
direction
changed
more
often
than
when
the
cell
was
not
hyperpolarized
the
level
of
the
ON1
pair.
The
question
was
wheth-
er
the
ON1
activity
has
a
measurable
influence
on
the
behavior.
To
test
this,
an
electrode
was
inserted
into
an
ON1
on
one
side
and
the
animal
was
induced
to
walk
(Figs.
6-8),
tending
to
turn
toward
whichever
of
the
speakers
was
broadcasting
the
calling
song.
The
ON1
cell
was
distinctly
more
strongly
excited
by
ipsilateral
than
by
contralateral
sound.
During
hyperpolarization,the
number
of
action
potentials
per
chirp
was
reduced,
for
sound
from
either
side
(Fig.
7,
left).
The
average
turning
velocity
in
re-
sponse
to
ipsilateral
sound
was
also
reduced
be-
cause,
although
turning
toward
the
speaker
side
continued,
there
was
also
a
strong
tendency
to
turn
626
K.
Schildberger
and
M.
Homer:
Function
of
cricket
auditory
neurons
N
left
ON1
RSH
N
rotational
velocity
RSH
20
80
n
10
Fig.
7.
Left:
distribution
of
the
number
of
action
potentials
per
chirp
discharged
by
the
left
ON1
in
a
LSH
walking
animal;
RSH
right
speaker
active,
neuron
hyperpolarized;
RS
right
speaker
active;
LSH
left
speaker
active,
neuron
hyperpolarized,
LS
left
speaker
active;
under
each
condition
240
chirps
(5
kHz,
80
dB)
were
presented.
Right:
distributions
of
angular
velocities
measured
at
the
same
times
as
the
spikes
on
the
left.
Hyperpolarization
of
ON1
with
the
sound
ipsilateral
(LSH)
reduced
the
mean
turning
velocity
and
increased
the
scatter;
hyperpolarization
during
contralateral
sound
presentation
had
0
50
deg
I
s
no
significant
effect
RS
LSH
LS
7
AP/chirp
-50
LS
RS
away
from
it
(Figs.
6,
7,
right).
These
effects
of
hyperpolarization
were
reversible
and
reproduc-
ible.
The
turning
velocity
also
depended
on
sound
intensity,
but
the
effect
of
hyperpolarization
on
turning
velocity
during
ipsilateral
sound
presenta-
tion
occurred
at
all
the
intensities
tested.
On
the
other
hand,
hyperpolarization
of
ON1
did
not
change
the
turning
velocity
during
contralateral
sound
presentation,
at
any
intensity
(Fig.
8).
The
example
in
Fig.
8
demonstrates
that
these
effects
were
observable
even
in
animals
that
did
not
re-
spond
symmetrically
to
sound
from
the
left
and
right.
Without
sound
presentation,
the
mean
turn-
ing
tendency
of
this
particular
animal
was
11
deg/s
to
the
right.
Thus,
hyperpolarization
during
ipsilat-
eral
sound
stimulation
reduced
the
turning
ten-
dency
to
about
the
value
of
the
no-sound
condi-
tion.
Not
all
animals
exhibited
significant
effects
of
hyperpolarizing
ON1
(Table
1).
Inactivation
of
AN1
The
cell
body
of
the
neuron
AN1
is
in
the
protho-
racic
ganglion;
contralateral
to
it,
the
axon
ascends
to
the
brain
(Wohlers
and
Huber
1982;
Schild-
berger
1984).
It
has
a
low
threshold
to
tones
at
the
carrier
frequency
of
the
conspecific
calling
song,
35-40
dB,
and
copies
the
pattern
of
the
song
in
its
discharge.
Like
the
other
identified
auditory
interneurons,
the
AN1
has
a
mirror-image
partner
cell.
These
cells
are
excited
by
the
ear
ipsilateral
to
the
ascending
axon
(as
previously,
a
sound
source
is
considered
to
be
ipsilateral
to
an
AN1
when
it
is
on
the
same
side
as
the
ear
that
excites
the
neuron).
The
properties
of
the
AN1
suggest
that
it
sends
to
the
brain
information
required
for
sound
localization
and
pattern
recognition.
Hyperpolarization
of
the
ipsilateral
ANI
dur-
ing
presentation
of
calling
song
caused
the
animal
to
reverse
its
turning
tendency;
having
previously
turned
toward
the
side
of
the
active
speaker,
during
AN1
hyperpolarization
the
cricket
turned
toward
the
other
side
(compare
LS
and
LSH
in
Fig.
9).
This
effect
was
reversible
and
reproducible.
In
the
example
of
Fig.
10,
turning
velocity
increased
with
sound
intensity
up
to
about
75
dB
and
then
re-
mained
constant;
between
55
and
75
dB
hyperpo-
larization
produced
a
reversal
of
direction.
In
this
intensity
range
the
turning
velocity
under
condi-
tions
of
ipsilateral
sound
presentation
plus
hyper-
polarization
was
indistinguishable
from
that
dur-
ing
contralateral
sound
presentation
without
hy-
perpolarization.
At
80
dB
the
turning
velocities
during
hyperpolarization
were
significantly
differ-
627
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
AP/Chirp
left
ON
1
cm
/s
0-
-
—0
LS
5-
,
.---•
LSH
RS
••-•-•
RSH
translational
velocity
left
AN
1
I
rotational
velocity
10
-
.0
.
..11;"""
--//
degls
rotational
velocity
degls
40.
01,f
rotation
cm
25
0
'°-
/
60
80
dB
Fig.
8.
Top:
relation
between
discharge
of
the
left
ON1
and
the
intensity
of
the
song,
the sound
direction,
and
the
injection
of
hyperpolarizing
current.
Abbreviations
as
in
Fig.
7.
Under
each
condition
120
chirps
were
presented
(
=
60
s).
Bottom:
an-
gular
velocities
measured
at
the
same
times
as
the
spikes;
at
each
intensity,
the
differences
in
angular
velocity
between
LS
and
RS,
LS
and
LSH
and
RS
and
LSH
are
significant
(I-test,
P<
0.001);
asterisk
marks
angular
velocity
when
no
sound
is
presented;
vertical
bars
give
standard
deviation
of
the
mean
of
the
respective
data
points
Table
1.
Effects
of
hyperpolarization
of
neurons
on
phonotaxis
Neuron
ON1
AN1
AN2
Other
No.
recorded
during
walking
25
5
20
8
No.
recorded
during
phonotaxis
8
4
8
4
No.
in
which
hyperpol.
decreases
angular
velocity
3
1
4
0
No.
in
which
hyperpol.
causes
reversal
of
turning
direction
0
4
0 0
ent
from
those
during
sound
presentation
from
ei-
ther
direction
without current
injection.
These
effects
on
locomotion
were
closely
corre-
lated
with
the
effects
of
hyperpolarization
on
the
neuronal
discharge.
That
is,
sound
stimuli
between
15
no.
of
action
potentials
AP
20-M
time
!mini
active
speaker
LS
LSH
LS
LSH
Fig.
9.
Intracellular
recording
from
the
left
AN1
during
phono-
tactic
response
to
calling
song.
Top
trace:
translational
veloci-
ty;
2nd
trace:
angular
velocity;
3rd
trace:
rotational
compo-
nent;
4th
trace:
number
of
action
potentials;
stimulus
condi-
tions
indicated
at
bottom.
Sound
intensity
65
dB
during
the
first
LS/RS
sequence,
75
dB
during
the
others
55
and
75
dB
elicited
significantly
less
spike
activi-
ty
when
presented
ipsilaterally
during
current
in-
jection
than
when
presented
contralaterally
with-
out
current
injection
(compare
LSH
and
RS
in
the
top
graph
of
Fig.
10),
but
at
80
dB
this
difference
disappeared.
On
the
assumption
that
the
members
of
the
AN1
pair
have
mirror-image
directional
characteristics,
it
would
follow
that
the
animal
al-
ways
turns
toward
the
side
of
the
more
strongly
excited
AN1.
All
the
animals
tested
(see
Table
1)
reversed
direction
when
one
AN1
was
hyperpolar-
ized,
as
long
as
the
neuronal
response
to
ipsilateral
sound
during
hyperpolarization
was
smaller
than
the
response
to
contralateral
sound.
The
other
case
LS
RS
LS
RS
628
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
AP/Chirp
left
AN
1
translational
velocity
left
AN2
cm/s
5
/
V
A-
o---o
LS
ce
.
LSH
a
.....
-DRS
0
60
80
dB
Fig.
10.
Top:
relation
between
the
responses
of
the
left
AN1
in
a
walking
animal
and
the
sound
intensity,
sound
direction,
and
current
injection.
Hyperpolarization
of
AN1
during
ipsilat-
eral
sound
presentation
caused
a
reversal
of
turning
direction;
under
each
condition
120
chirps
were
presented;
at
each
intensi-
ty
differences
between
all
the
data
for
the
different
conditions
are
statistically
significant
(t-test,
P<
0.01)
except
for
LSH
and
RS
at
55
and
80
dB.
Bottom:
angular
velocities
measured
at
the
same
times
as
the
neuronal
responses;
values
for
LS
at
each
intensity
are
statistically
different
from
those
for
the
other
conditions (t-test,
P
<0.001);
except
for
the
values
at
80
dB
the
data
for
LSH
and
RS
do
not
differ
significantly;
asterisk
marks
angular
velocity
when
no
sound
is
presented
noted
as
a
decrease
rather
than
a
reversal
reflects
the
response
in
Fig.
10
at
80
dB.
Inactivation
of
AN2
AN2
is
a
plurisegmental
neuron
similar
to
AN1.
It
originates
in
the
prothoracic
ganglion
and
termi-
nates
in
the
brain,
but
its
tuning
differs
from
that
of
AN1.
AN2
responds
to
a
broad
band
of
fre-
quencies
above
10
kHz,
as
well
as
to
the
calling
song
at
intensities
of
50-60
dB
or
higher.
It
does
not
copy
the
song
pattern
as
accurately
as
AN1.
AN2
is
excited
by
the
ear
ipsilateral
to
its
axon
and
inhibited,
or
in
some
cases
weakly
excited,
by
the
other
ear
(Wohlers
and
Huber
1982).
The
inhi-
bition
is
mediated
by
the
contralateral
ON1
(Sel-
rotational
velocity
deg
ls
40-
J'A
rotation
cm
15-
no.
of
action
potentials
AP
20-
Klq
ArAkIII..
2
time
Imini
active
speaker
LS
LSH
RS
RSH
LS
LSH
LS
Fig.
11.
Intracellular
recording
from
the
left
AN2
during
phon-
otactic
responses
to
calling
song.
Top
trace:
translational
veloc-
ity;
2nd
trace:
angular
velocity;
3rd
trace:
rotational
compo-
nent;
4th
trace:
number
of
action
potentials
per
chirp;
stimulus
conditions
indicated
at
the
bottom.
Sound
intensity
90
dB
for
the
first
LS
sequence
and
80
dB
for
the
other
sequences
verston
et
al.
1985).
Although
the
main
role
of
AN2
is
thought
to
lie
in
the
processing
of
high-
frequency
sound,
a
function
in
positive
phonotaxis
cannot
be
ruled
out
a
priori.
Hyperpolarization
of
AN2
during
ipsilateral
sound
presentation
reduced
the
mean
turning
ve-
locity
(Fig.
11);
as
in
the
case
of
ON1
hyperpolari-
zation,
there
was
an
increased
tendency
to
turn
away
from
the
speaker.
In
the
example
of
Fig.
12,
when
AN2
was
not
hyperpolarized
the
turning
ve-
locity
rose
with
increasing
intensity
up
to
80
dB
and
then
fell
off.
The
effect
of
hyperpolarization
in
this
high-intensity
range
depended
on
the
direc-
tion
of
the
sound
source.
With
ipsilateral
presenta-
tion
of
sound
at
80
dB
or
more,
hyperpolarization
of
AN2
decreased
the
tendency
to
turn
toward
the
degls
rotational
velocity
25-
0
_
15
K.
Schildberger
and
M.
Homer:
Function
of
cricket
auditory
neurons
629
left
AN2
ond
type
of
omega
neuron
(ON2),
a
descending
neuron
(DN1)
and
a
neuron
with
ascending
and
descending
axon
(TNI).
So
far
no
change
in
turn-
ing
tendency
has
been
found
to
result
from
hyper-
polarization
of
any
of
these
cells
(see
Table
1).
AP/Chirp
o--
-0
LS
LSH
RS
RSH
10
/
/
/
cr
rotational
velocity
0
0-
60
dB
Fig.
12.
Top:
relation
between
the
response
of
the
left
AN2
in
a
walking
animal
and
the
sound
intensity,
sound
direction
and
current
injection.
Under
each
condition
120
chirps
were
presented;
there
are
significant
differences
between
all
groups
at
80
and
90
dB
(t-test,
P<
0.001).
Bottom:
angular
velocities
measured
at
the
same
times
as
the
neuronal
activity;
with
ipsi-
lateral
presentation
at
over
70
dB
hyperpolarization
reduced
turning
velocities,
and
with
contralateral
presentation
it
raised
turning
velocity.
Differences
between
LS
and
RS
are
significant
at
every
intensity,
as
are
those
between
LS
and
LSH
at
80
and
90
dB
and
between
RS
and
RSH
at
90
dB
(t-test,
P<
0.001)
speaker
if
the
neuronal
response
was
smaller
than
the
response
to
contralateral
sound
without
hyper-
polarization.
During
contralateral
presentation
at
high
intensities,
hyperpolarization
increased
the
tendency
to
turn
toward
the
speaker,
by
a
signifi-
cant
amount
at
90
dB.
In
either
case,
reduction
of
the
spike
rate
of
AN2
during
ipsilateral
stimula-
tion
below
the
rate
during
contralateral
one
af-
fected
turning
tendency
only
at
high
sound
intensi-
ties
and
never
caused
a
complete
reversal
of
direc-
tion
such
as
was
produced
by
inactivating
AN1.
Furthermore,
hyperpolarization
of
AN2
affected
the
behavior
in
only
4
of
8
animals
tested.
Inactivation
of
other
auditory
neurons
Auditory
neurons
other
than
the
above
have
also
been
identified
in
the
prothoracic
ganglion
a
sec-
Discussion
In
the
experimental
situation
described
here,
crick-
ets
exhibit
phonotactic
behavior.
As
a
measure
of
this
phonotactic
performance,
we
have
recorded
the
direction
and
magnitude
(in
terms
of
angular
velocity)
of
the
turning
tendency.
The
angular
ve-
locities
we
have
observed
are
in
the
same
range
as
those
measured
in
the
same
cricket
species
with
other
methods
(Stabel
and
Wendler
1986).
Never-
theless,
certain
limitations
of
this
method
should
be
noted.
For
intracellular
recording
the
cricket's
body
must
be
suspended
from
a
rigid
support,
so
that
the
forces
acting
on
the
legs
are
not
necessarily
the
same
as
those
that
act
when
the
animal
is
sup-
porting
its
own
weight.
Correct
loading
of
the
joints
is
very
important
for
properly
coordinated
walking
(see
Graham
1985).
Therefore
the
system
was
carefully
adjusted
so
that
the
positions
of
body
and
legs
matched
those
of
the
freely
walking
ani-
mal
as
closely
as
possible.
Under
these
conditions
walking
was
initiated
spontaneously,
the
normal
tripod
gait
was
used,
walking
was
accompanied
by
the
typical
movements
of
antennae
and
palps,
and
the
duration
and
frequency
of
the
pauses
in
walking
corresponded
to
those
in
free-moving
ani-
mals.
The
weight
of
the
sphere
was
about
3/2
the
average
body
weight
of
an
adult
female,
so
that
the
translational
inertia
corresponded
approxi-
mately
to
that
for
a
freely
walking
cricket
(Dahmen
1980;
Weber
et
al.
1981).
The
rotational
moment
of
inertia,
however,
is
considerably
higher
than
in
free
walking,
which
can
affect
the
animal's
gait
during
attempted
turns.
Asymmetric
turning
tendencies
were
also
fre-
quently
observed,
but
were
not
due
to
the
appara-
tus
because
different
animals
were
asymmetric
in
different
directions.
Animals
that
did
not
clearly
turn
toward
an
active
speaker
were
rejected.
Although
the
data
base
was
limited
for
the
rea-
sons
given
above,
the
reproducibility
and
revers-
ibility
of
the
effects
observed
justify
certain
conclu-
sions.
Hyperpolarization of
a
neuron
diminishes
its
response
to
sound.
If
there
is
a
comparator
that
evaluates
the
difference
in
excitation
of
the
left
and
right
cells
of
a
pair,
when
one
cell
is
hyperpolarized
the
comparator
will
derive
an
erroneous
estimate
of
sound-source
position
that
could
produce
an
-
79
degls
1
25-
630
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
altered
turning
tendency.
In
the
case
of
AN1
such
an
alteration
was
in
fact
observed;
the
direction
of
the
turning
tendency
was
reversed
as
long
as
the
discharge
of
the
hyperpolarized
cell
in
response
to
ipsilateral
sound
was
less
than
the
response
to
contralateral
sound
without
hyperpolarization.
Given
that
the
members
of
the
AN1
pair
have
mir-
ror-image
directional
characteristics,
it
follows
that
the
animal
will
turn
toward
the
side
of
the
more
strongly
excited
AN1.
It
remains
unclear
whether
sound
localization
is
possible
when
the
AN1
neu-
rons
on
both
sides
are
inoperative.
Therefore
the
demonstration
of
necessity
of
the
AN1
pair
for
phonotaxis
is
not
yet
complete.
The
turning
tendency
can
also
be
influenced
by
inactivation
of
another
ascending
cell,
AN2.
But
here
the
influence
of
hyperpolarization
with
ipsilateral
sound
does
not
suffice
for
a
complete
reversal
of
walking
direction,
as
it
does
in
AN1.
In
view
of
the
higher
threshold
of
AN2
to
5
kHz
stimuli,
one
would
expect
an
effect
of
hyperpolari-
zation
to
become
apparent
only
at
higher
sound
intensities.
Interindividual
differences
in
threshold
of
the
AN2
could
also
explain
why
hyperpolariza-
tion
of
the
AN2
was
not
effective
in
all
animals.
An
effect
of
hyperpolarization
on
turning
velocity
was
observed
chiefly
in
animals
in
which
AN2
had
a
low
threshold
in
the
5
kHz
region.
In
Teleogryllus
an
HF-neuron
(INT-1,
probab-
ly
homologous
to
AN2
in
Gryllus)
has
been
shown
to
influence
the
flight
behavior
(Nolen
and
Hoy
1984).
Hyperpolarization
of
INT-1
abolishes
activ-
ity
of
contralateral
abdominal
flexion
muscles,
an
action
that
is
thought
to
diminish
the
negative
phonotaxis
of
these
animals
in
response
to
stimula-
tion
with
ultrasound.
By
contrast,
hyperpolariza-
tion
of
AN2
in
walking
crickets
diminishes
positive
phonotaxis.
These
disparate
findings
are
not
neces-
sarily
contradictory.
The
sound
frequencies
used
in
the
two
experiments
are
different,
as
was
the
behavioral
context.
Furthermore,
it
has
not
been
definitely
established
that
the
two
neurons
are
in-
deed
homologous
and
have
identical
functions.
Attempts
have
also
been
made
to
examine
the
influence
of
auditory
neurons
on
phonotaxis
in
Acheta
(Atkins
et
al.
1984;
Stout
et
al.
1985).
These
authors
report
specific
impairments
of
orien-
tation
after
selective
killing
of
individual
auditory
neurons.
However,
the
results
of
these
studies
are
not
comparable
with
those
described
here.
On
one
hand,
it
is
unclear
whether
the
physiological
conse-
quences
of
selective
cell
killing
are
comparable
to
those
of
reversible
reduction
of
activity
by
hyper-
polarization;
on
the
other,
the
methods
by
which
behavior
is
measured
in
the
two
cases
are
very
different.
In
our
experiments
effects
on
behavior
were
assayed
continuously
in
an
open-loop
situa-
tion
at
the
same
time
that
the
activity
of
single
neurons
was
monitored.
The
authors
cited
above
analyzed
runs
in
an
arena.
The
effects
on
phono-
taxis
after
cell
killing
that
they
reported
were
based
on
the
description
of
a
single
run,
in
one
particular
stimulus
configuration,
as
compared
with
another
single
run
of
the
same
animal
before
the
cell
had
been
killed.
The
significance
of
the
differences
between
the
pre-
and
post-killing
behavioral
observations
ap-
pears
questionable
for
the
following
reasons
:
(i)
Single
runs
in
an
arena
are
necessarily
quite
brief,
because
of
the
short
distance
between
the
starting
point
and
the
speaker
(in
this
case,
56
cm).
Given
a
mean
velocity
of
3
cm/s
(Stout
et
al.
1976)
the
runs
described
in
these
papers
would
each
have
lasted
about
20-30
s.
(ii)
The
runs
of
intact
ani-
mals,
as
well
as
those
in
some
control
experiments,
are
by
no
means
always
directed
straight
toward
the
speaker
(see
Atkins
et
al.
1984,
Figs.
3,
4,
6,
7;
Stout
et
al.
1983,
Fig.
2).
In
relatively
short
sin-
gle
runs
this
fact
can
make
the
interpretation
of
orientation
performance
difficult
or
even
question-
able.
(iii)
The
paths
travelled
by
the
same
individ-
ual
in
successive
test
series,
even
when
the
animal
is
intact,
vary
greatly
although
the
stimulus
para-
meters
are
identical
(see
Stout
et
al.
1976,
Fig.
3).
It
is
not
clear
why,
in
our
experiments,
hyper-
polarization
of
AN2
or
of
ON1
in
some
cases
does
not
produce
a
measurable
change
in
behavior.
One
possibility
is
that
individuals
differ
in
the
relative
influence
of
these
two
neurons
on
the
central
com-
parator.
Another
explanation
could
lie
in
the
re-
stricted
range
of
variation
of
the
stimulus
parame-
ters.
The
auditory
stimulus
was
always
simulated
calling
song,
and
the
sound
was
incident
from
only
two
directions.
Therefore
we
cannot
exclude
the
possibility
that
neurons
other
than
AN1
might
par-
ticipate
more
actively
in
other
stimulus
configura-
tions.
Inactivation
of
an
AN1
did
not
abolish
phono-
taxis
but
induced
only
a
directional
error.
The
con-
specific
song
pattern
was
recognized
as
well
as
be-
fore.
The
implication
may
be
that
there
are
other
(as
yet
unidentified)
neurons
on
the
same
side
as
the
inactivated
neuron
that
transmit
the
conspe-
cific
pattern
to
the
brain.
On
the
other
hand,
it
might
be
sufficient
for
the
recognition
process
if
the
patterned
signal
reaches
the
brain
on
only
one
side
as
transection
experiments
of
cervical
connec-
tives
indicate
for
crickets
and
grasshoppers
(Weber
pers.
comm.;
Regen
1926;
Ronacher
1986).
In
any
case,
the
neuron
pair
AN1
sends
to
the
brain
signif-
K.
Schildberger
and
M.
Horner:
Function
of
cricket
auditory
neurons
631
icant,
if
not
absolutely
necessary
information
for
sound
localization
and
pattern
recognition.
It
is
still
entirely
unclear
whether
or
how
these
two
as-
pects
of
the
calling
song
are
processed
indepen-
dently
of
one
another
in
the
brain,
or
how
the
brain
triggers
and
controls
phonotactic
walking.
Acknowledgements.
We
thank
M.A.
Biederman-Thorson
for
the
translation
into
English,
J.
Thorson
and
F.
Huber
for
critical
comments,
H.U.
Kleindienst
for
providing
computer
programs,
M.L.
Obermayer
and
S.
Schmaderer
for
technical
assistance.
References
Atkins
GS,
Ligman
F,
Burghardt
F,
Stout
FJ
(1984)
Changes
in
phonotaxis
by
the
female
cricket
Acheta
domesticus
after
killing
identified
acoustic
interneurons.
J
Comp
Physiol
A
154:795-804
Casaday
GB,
Hoy
RR
(1977)
Auditory
interneurons
in
the
cricket
Teleogryllus
oceanicus:
physiological
and
anatomical
properties.
J
Comp
Physiol
121
:1-13
Dahmen
HJ
(1980)
A
simple
apparatus
to
investigate
the
orien-
tation
of
walking
insects.
Experientia
36:685-687
Godden
DH,
Graham
D
(1984)
A
preparation
of
the
stick
insect
Carausius
morosus
for
recording
intracellularly
from
identified
neurones
during
walking.
Physiol
Entomol
9:275-286
Graham
D
(1985)
Pattern
and
control
of
walking
in
insects.
Adv
Insect
Physiol
18:31-140
Hedwig
B
(1986)
On
the
role
in
stridulation
of
plurisegmental
interneurons
of
the
acridid
grasshopper
Omocestus
viridulus
L.
J
Comp
Physiol
A
158:429-444
Huber
F,
Thorson
J
(1985)
Cricket
auditory
communication.
Sci
Am
253:
60-68
Kien
J
(1983)
The
initiation
and
maintenance
of
walking
in
the
locust:
an
alternative
to
the
command
concept.
Proc
R
Soc
Lond
B
219:137-174
Kleindienst
HU,
Koch
UT,
Wohlers
DW
(1981)
Analysis
of
the
cricket
auditory
system
by
acoustic
stimulation
using
a
closed
sound
field.
J
Comp
Physiol
141:283-296
Kramer
E
(1975)
Orientation
of
the
male
silkmoth
to
the
sex
attractant
Bombykol.
In:
Denton
DA,
Coghlan
IP
(eds)
Olfaction
and
taste.
Academic
Press,
New
York,
pp
329-355
Moiseff
A,
Hoy
RR
(1983)
Sensitivity
to
ultrasound
in
an
iden-
tified
auditory
interneuron
in
the
cricket:
a
possible
neural
link
to
phonotactic
behavior.
J
Comp
Physiol
152:155-167
Nolen
TG,
Hoy
RR
(1984)
Initiation
of
behavior
by
single
neurons:
the
role
of
behavioral
context.
Science
226:992-994
Popov
AV,
Markovich
AM,
Andjan
AS
(1978)
Auditory
inter-
neurons
in
the
prothoracic
ganglion
of
the
cricket
Gryllus
bimaculatus.
J
Comp
Physiol
126:183-192
Regen
J
(1926)
Uber
die
Beeinflussung
der
Stridulation
von
Thamnotrizon
apterus
durch
kiinstlich
erzeugte
Tone
und
verschiedenartige
Gerausche.
Sitz
Ber
Akad
Wiss
Wien
Math
Nat
Kl
135:329-368
Robertson
RM,
Pearson
KG
(1982)
A
preparation
for
the
in-
tracellular
analysis
of
neuronal
activity
during
flight
in
the
locust.
J
Comp
Physiol
146:311-320
Ronacher
B
(1986)
Auslosung
von
Drehung
und
Gesangsant-
wort
der
Mannchen
von
Chorthippus
biguttulus
durch
den
Weibchengesang
:
Eingrenzung
der
notwendigen
Nerven-
bahnen
durch
Ausschaltexperimente.
In:
Elsner
N,
Rath-
mayer
W
(eds)
Sensomotorik.
Identifizierte
Neurone.
Thieme,
Stuttgart,
p
143
Schildberger
K
(1984)
Temporal
selectivity
of
identified
audito-
ry
neurons
in
the
cricket
brain.
J
Comp
Physiol
A
154:171-185
Schildberger
K,
Milde
JJ,
Homer
M
(1988)
The
function
of
auditory
neurons
in
cricket
phonotaxis
II.
Modulation
of
auditory
responses
during
locomotion.
J
Comp
Physiol
A
163:633-640
Schmitz
B,
Scharstein
H,
Wendler
G
(1982)
Phonotaxis
in
Gryl-
lus
campestris
L.
J
Comp
Physiol
148:431-444
Selverston
AI,
Kleindienst
HU,
Huber
F
(1985)
Synaptic
con-
nectivity
between
cricket
auditory
interneurons
as
studied
by
selective
photoinactivation.
J
Neurosci
5:1283-1292
Stabel
J,
Wendler
G
(1986)
Akustische
Interneurone
und
das
Vorzeichen
der
Phonotaxis.
In:
Elsner
N,
Rathmayer
W
(eds)
Sensomotorik.
Identifizierte
Neurone.
Thieme,
Stutt-
gart,
p
145
Stout
J,
Gerard
G,
Hasso
S
(1976)
Sexual
responsiveness
me-
diated
by
the
corpora
allata
and
its
relationship
to
phono-
taxis
in
the
female
cricket.
J
Comp
Physiol
108:1-9
Stout
JF,
DeHaan
CH,
McGhee
RW
(1983)
Attractiveness
of
the
male
Acheta
domesticus
calling
song
to
females.
I.
De-
pendence
on
each
of
the
calling
song
features.
J
Comp
Phys-
iol
153:
509-521
Stout
JF,
Atkins
G,
Burghardt
F
(1985)
The
characterization
and
possible
importance
for
phonotaxis
of
L-shaped
ascend-
ing
acoustic
interneurons
in
the
cricket.
In:
Kalmring
K,
Elsner
N
(eds)
Acoustic
and
vibrational
communication
in
insects.
Parey,
Berlin
Hamburg,
pp
89-100
Thorson
J,
Weber
T,
Huber
F
(1982)
Auditory
behavior
of
the
cricket.
J
Comp
Physiol
146:361-378
Weber
T,
Thorson
J,
Huber
F
(1981)
Auditory
behavior
of
the
cricket.
J
Comp
Physiol
141:215-232
Wiese
K,
Eilts
K
(1985)
Functional
potential
of
recurrent
lateral
inhibition
in
cricket
audition.
In:
Kalmring
K,
Elsner
N
(eds)
Acoustic
and
vibrational
communication
in
insects.
Parey,
Berlin
Hamburg,
pp
33-40
Wohlers
DW,
Huber
F
(1978)
Intracellular
recording
and
stain-
ing
of
cricket
auditory
interneurons.
J
Comp
Physiol
127
:
11-28
Wohlers
DW,
Huber
F
(1982)
Processing
of
sound
signals
by
six
types
of
neurons
in
the
prothoracic
ganglion
of
the
cricket
Gryllus
campestris
L.
J
Comp
Physiol
146:161-173