Recognition of temporal, patterns by identified auditory neurons in the cricket brain


Schildberger, K.

Acoustic and Vibrational Communication in Insects: 41-49

1985


41
RECOGNITION
OF
TEMPORAL
PATTERNS
BY
IDENTIFIED
AUDITORY
NEURONS
IN
THE
CRICKET
BRAIN
KLAUS
SCHILDBERGER
ABSTRACT
Three
types
of
auditory
interneurons
have
been
identified
in
the
cricket
brain
by
intracellular
recording
and
staining:
ascending
neurons
and
two
types
of
local
brain
cells.
The
tuning
curves
revealed
broad
band
neurons
and
those
which
are
tuned
to
the
carrier
frequency
of
the
conspecific
calling
song
in
each
type
of
brain
cells.
Ascending
neurons
with
the
best
frequency
at
5
kHz
copy
different
time
patterns
of
acoustic
stimuli
very
well,
while
a
precise
copying
is
lost
in
brain
cells.
On
the
other
hand,
ascending
cells
respond
to
nearly
all
time
patterns,
while
specific
types
of
brain
cells
respond
only
to
those
patterns
which
are
effec-
tive
in
phonotactic
behaviour.
So
there
might
exist
specific
detectors
for
behaviour
relevant
acoustic
signals
in
the
brain
of
the
cricket.
INTRODUCTION
Female
crickets
are
attracted
by
the
species-specific
calling
song
of
conspecific
males.
Two
parameters
of
the
song
are
particularly
important
in
eliciting
and
maintaining
the
phonotactic
response
of
the
females:
spectral
compositon
and
temporal
pattern
(POPOV
&
SHUVALOV,
1977;
THORSON
et
al.,
1982).
Whereas
the
auditory
system
of
the
cricket
exhibits
tuning
to
certain
frequencies
at
the
level
of
the
primary
nerve
fibres
(NOCKE,
1972;
ESCH
et
al.,
1980;
HUTCHINGS
&
LEWIS,
1981),
no
specific
temporal
filtering
is
decernible
at
this
level.
In
the
prothoracic
ganglion
further
processing
of
the
auditory
information
from
a
number
of
local,
ascending
and
descending
neurons
occurs
(POPOV
&
MARKOVICH,
1982;
WOHLERS
&
HUBER,
1978,
1982;
BOYD
et
al.,
1984).
However,
neither
the
ascending
nor
the
descending
neurons
in
the
prothorax
are
specifically
tuned
to
phonotactically
effective
temporal
patterns.
A
main
goal
in
understanding
the
mechanisms
underlying
phonotaxis
is
to
find
out
whether
the
brain
contains
identifiable
neurons
with
specific
temporal
filter
properties
that
in
fact
match
the
behaviourally
demonstrated
recognition
para-
meters
of
the
conspecific
song.
In
the
experiments
described
here,
both
the
anatomical
and
the
physiological
properties
of
individual
auditory
neurons
in
the
cricket
brain
were
examined
by
intracellular
recording
and
staining.
In
particular,
the
pattern-dependent
responses
of
the
brain
neurons
were
compared
with
those
of
ascending
neurons.
Kalmring/F.Isner
(eds.),
Acoustic
and
Vibrational
Communication
in
Insects
C
1985
Paul
Parey
(
a
?I
f
(
a
b
BNC
2
AN
°8r-
*
b
BNC1
)r*,1
c
d
42
SCHILDBERGER
ANATOMY
The
two
prothoracic
neurons
ANI
and
AN2
(see:
WOHLERS
&
HUBER,
1982)
ascend
to
the
brain.
Here,
the
AN1
axon
projects
to
a
ventral
posterior
region
at
the
border
of
the
proto-
and
deutocerebrum;
the
axon
continues
anterior
and
dorsally
to
end
in
a
main
projection
field
that
is
lateral
to
the
alpha-lobe.
AN2
passes
through
the
brain
in
a
similar
direction
and
ends
with
its
main
arborization
field
in
a
region
that
is
located
similarly
to
that
of
ANI
(Fig.
1).
GN
Fig.
1.
Auditory
neurons
in
the
brain.
Left:
low-frequency
(5
kHz)
tuned
auditory
neurons;
reconstructions
from
serial
sections
of
Lucifer
Yellow
stained
cells;
frontal
view;
GN
neurons
are
from
Acheta
domesticus,
all
other
neurons
from
Gryllus
bimaculatus;
abbreviations
(as
in
all
other
figures):
AN:
ascending
neuron,
BNC1:
brain
neuron
class
1,
BNC2:
brain
neuron
class
2,
GN:
extrinsic
glomerular
neuron,
DN:
descending
brain
neuron.
Right:
high-frequency
and
broad-band
tuned
auditory
neurons.
Auditory
brain
neurons
43
In
addition
to
these
ascending
cells,
a
number
of
auditory
interneurons
were
found
that
lie
entirely
within
the
brain
or
descend
to
lower
ganglia.
Local
brain
cells
were
classified
according
to
anatomical
criteria.
Four
major
classes
of
auditory
cells
could
be
discerned.
BNCI
(Brain
Neurons
Class
1)
neurons
have
arborizations
in
the
main
projec-
tion
field
of
the
ascending
neurons.
In
this
area
BNCI
cells
never
show
bouton-like
end
specializations
but
exhibit
very
fine
branches.
So
BNCI
cells
could
be
postsyn-
aptic
and
direct
followers
of
the
ascending
neurons.
Some
BNCI
neurons
arborize
only
within
this
field
(local
BNCI),
whereas
others
have
extensive
ipsilateral
branches
in
a
more
central
to
posterior
region
between
deuto-
and
protocerebrum.
Here,
BNCI
neurons
exhibit
thickened
terminal
end
structures,
indicating
possible
presynaptic
sites.
BNC2
neurons
are
located
in
this
central
to
posterior
region
of
the
diffuse
neuropil
of
the
brain.
These
cells
do
not
overlap
with
ascending
neurons
but
with
arborizations
of
the
BNCI
neurons.
BNC2
neurons
rarely
have
two
distinct
projection
fields
and
the
appearence
and
position
of
the
terminal
branches
provide
no
clear
evidence
of
spatial
separation
of
the
main
input
and
output
regions.
GN
neurons
(extrinsic
glomerular
neurons)
comprise
a
third
class
of
cells
that
project
into
the
glomerular
neuropil
of
the
mushroom
bodies
or
the
central
body.
Due
to
their
widespread
arborizations,
GN
neurons
overlap
either
with
ascending,
BNCI,
BNC2
or
with
neurons
of
all
classes.
Detailed
anatomy
and
physiology
of
these
cells
is
described
elsewhere
(SCHILDBERGER,
1984a).
In
a
fourth
class
there
exist
neurons
with
a
cell
body
and
arborizations
in
the
brain
and
an
axon
that
descends
to
the
ventral
cord.
The
descending
neu-
rons
identified
so
far
project
mostly
in
a
ventral
and
posterior
region
of
the
brain.
They
can
overlap
with
projections
of
BNCI
and
BNC2
cells,
but
also
with
some
ventral
projections
of
the
ascending
cells.
These
anatomical
results
as
well
as
latency
measurements
indicate
that
there
are
several
pathways
for
processing
auditory
information.
One
pathway
leads
from
the
ascending
neurons
via
the
BNCI
to
the
BNC2
cells
and
then
possibly
to
descending
neurons.
This
pathway
could
include
the
mushroom
bodies
and
the
central
body
as
additional
stations
between
BNCl/BNC2
and
descending
cells;
another
possibility
is
a
more
direct
coupling
of
ascending
and
descending
neurons.
PHYSIOLOGY
Threshold
curves
and
intensity-response
functions
were
measured
for
all
of
the
above
cells
and
are
given
in
detail
elsewhere
(SCHILDBERGER,
1984b).
In
general,
brain
neurons
of
all
anatomical
classes
can
be
classified
into
two
physiological
types
according
to
their
frequency
specificity
(Fig.
I):
neurons
with
a
maximum
sensitivity
at
about
5
kHz
(low-frequency
tuned)
which
is
the
carrier
frequency
of
the
conspecific
calling
song,
and
neurons
with
a
maximum
sensitivity
in
the
higher
frequency
range
(high-frequency
tuned).
To
test
for
temporal
pattern
specificity
in
brain
neurons,
snythetic
acoustic
stimuli
(chirps)
were
presented
which
differed
in
their
intrachirp
temporal
pattern
(syllable
structure).
These
chirps
were
of
constant
length
and
energy
and
the
syllable
repetition
interval
(SRI),
was
varied
from
10-98
ms
(Fig.
6).
This
range
was
used
because
the
phonotactically
attractive
range
extends
from
about
35
to
55
ms.
Examples
of
neuronal
responses
to
these
temporal
patterns
are
shown
in
Fig.
2.
In
addition
to
these
constant
energy
patterns
(50%
duty
cycle),
which
necessarily
have
varying
SRI,
syllable
duration
and
syllable
number,
other
chirp
patterns
were
used
with
varying
SRI
but
constant
syllable
duration
and
syllable
number
(varying
duty
cycle).
Recognition
of
the
conspecific
calling
song
may
be
based
on
two
kinds
of
temporal
pattern
specificity.
A
neuron
may
respond
to
a
variety
of
temporal
patterns
but
synchronize
its
firing
to
the
temporal
patterns
only
at
a
characteris-
tic
syllable
repetition
interval.
At
rates
lower
or
higher
then
the
most
preferred
1
44
SCHILDBERGER
1
lilt
BNC
10
_IL
"_11,
--N—hv
_i*A"""r•NviA,As&
BNC2b
BNC
2
a
DN
I
DN
2
L
40000000000000—
-0-0-0-0-0-0-0-0--
Fig.
2.
Responses
of
different
brain
neurons
to
chirps
with
varying
syllable
repetition
interval.
Only
AN1
copies
all
patterns;
local
and
descending
brain
neurons
show
specific
filter
properties.
SRI's
were
18
ms
(left),
34
ms
(middle)
and
98
ms
(right)
at
5
kHz
and
80
dB
SPL;
calibration:
abscissa,
80
ms;
ordinate,
50
my
(traces
1-3)
and
25
my
(traces
4-8).
one,
the
pattern
of
spikes
would
become
more
random.
Alternatively
or
additional-
ly,
a
neuron
could
act
as
a
temporal
filter
by
spiking
maximally
at
a
particular
syllable
repetition
interval,
lower
or
higher
intervals
would
elicit
a
less
vigorous
response.
Synchronization
coefficients
(see:
SCHILDBERGER,
1984b)
were
calculated
for
different
neurons
and
plotted
as
a
function
of
syllable
repetition
interval
(Fig.
3).
The
coefficient
measures
mainly
the
degree
to
which
the
neuron
copies
the
temporal
pattern.
In
the
case
of
low-frequency
tuned
ascending
neurons,
the
coefficient
improves
as
SRI
increases.
The
coefficient
for
BNC1
neurons
increased
far
more
slowly
with
SRI;
some
coefficients
reached
the
significance
level
either
not
at
all
or
only
at
SRI's
of
50
ms
or
more.
BNC2
neurons
exhibited
no
correlation
between
the
coefficient
and
SRI
and
rarely
reached
the
significance
level.
Descending
neurons
did
not
show
any
synchronization
between
response
and
chirp
pattern
(Fig.
2).
As
a
measure
of
the
magnitude
of
the
response,
the
number
of
action
poten-
tials
discharged
per
chirp
was
counted
for
each
of
the
various
constant-energy
patterns
for
the
different
neurons
(Fig.
3).
For
a
given
pattern,
the
response
was
always
greatest
for
ANI,
followed
by
those
of
BNC1
and
BNC2.
With
stimuli
well
above
threshold
intensity,
the
AN1
responses
did
not
depend
on
SRI
but
on
sound
intensity.
A
similar
result
was
obtained
with
many
but
not
all
BNC1
neurons;
that
is,
there
was
no
preferential
response
to
certain
SRI
but
the
re-
sponse
was
intensity
dependent.
By
contrast,
some
BNC2
neurons
(BNC2a
and
BNC2c)
responded
more
strongly
to
chirps
with
SRI's
between
30-50
ms
(the
range
eliciting
phonotaxis)
than
to
those
with
longer
or
shorter
SRI's.
With
sound
intensities
more
than
10
dB
above
threshold,
this
pattern
dependence
was
unaf-
fected
by
intensity.
AN
1
1111111WL
AN
2
----
Auditory
brain
neurons
45
Synchronisation
1
BNC
2
5
kHz
80
dB
1(
Spikes/Chirp
4a
dB
4,--.
80
—•
70
60
2
1
BNC
1
5
kHz
80
dB
15-
5-
1
I I
TTTTTTT
AN
1
5
kHz
80
dB
40-
20-
0
1
26
50
74
98
26
50
74
98
III
II
Syllable
Repetition
Interval
Ims1
Fig.
3.
Response
properties
of
various
auditory
brain
neurons
as
a
function
of
syllable
repetition
interval.
Left:
synchronization
coefficient
as
a
function
of
SRI;
each
continuous
line
connects
the
data
from
a
single
neuron,
several
neurons
of
each
class
are
represented;
the
dashed
line
marks
the
5%
significance
level
(chi-square);
sound
frequency:
5
kHz,
intensity:
80
dB
SPL.
Right:
response
magni-
tude
as
a
function
of
SRI;
data
points
represent
the
average
of
12
responses
of
the
same
neuron;
sound
frequency:
5
kHz,
intensity:
60-80
dB
SPL,
chirp
repetition
interval:
500
ms.
The
filter
characteristics
of
band-pass
neurons
could
be
realized
by
an
ANDing
of
low-pass
and
high-pass
filter
neurons.
Besides
these
bandpass
filters
there
exist
in
the
brain
neurons
that
responded
to
short
and
intermediate
SRI's
but
not
to
long
ones
(a
high-pass
filter
in
terms
of
syllable
repetition
rate)
and
others
responsive
to
long
and
intermediate
SRI's
but not
to
short
ones
(a
low-pass
filter).
Various
mechanisms
could
underlie
these
filter
properties.
The
high-
46
SCHILDBERGER
100
1
5 0
100
nnnnnnnnnn
rn
4--•
AN 1
.--
DBNC
1
d
50
100
Syllable
repetition
interval
I.rns1
rrm
nnnn
—•
AN 1
BNC
1
d
50
100
Syllable
pause
duration
(ms]
Fig.
4.
Relative
response
magnitude
of
an
ascending
and
two
BNCId
neurons
to
different
chirp
paradigms.
Left:
50%
duty
cycle
(chirp
duration
and
sound
energy
constant
but
SRI,
syllable
length
and
syllable
number
varying),
response
is
plotted
as
a
function
of
SRI.
Right:
varying
duty
cycle
(chirp
duration,
sound
energy
and
SRI
vary
but
syllable
length
and
syllable
number
were
constant);
response
is
plotted
as
a
function
of
syllable
pause
duration;
broken
lines:
syllable
length
30
ms,
solid
lines:
syllable
length
20
ms.
Sound
freqeuncy:
5
kHz,
intensity:
80
dB
SPL.
pass
characteristics
shown
in
Fig.
2
(BNC2b)
and
Fig.
6
may
be
due
to
long
lasting
inhibition
that
builds
up
during
a
long
syllable
and
persits
over
the
next
syllable
period.
This
could
prevent
the
neuron
from
firing
as
a
response
to
the
next
syllable,
whereas
with
shorter
syllable
length
and
repetition
rate
the
excita-
tion
dominates
the
inhibition.
The
records
in
Fig.
2
suggest
that
the
low-pass
characteristics
of
neuron
(BNC1c)
results
from
temporal
summation
of
the
graded
potentials;
the
relatively
long
syllables
at
intermediate
and
long
SRI's
permit
these
potentials
to
become
large
enough
to
generate
action
potentials,
whereas
with
short
syllables
they
do
not
build
up
to
the
threshold
level.
A
further
experimental
constraint
on
low-pass
behaviour
occurs
in
the
re-
sponses
of
the
neurons
in
Fig.
4.
Here
the
filter
properties
of
an
ascending
and
two
BNCId
neurons
are
compared
for
50%
(Fig.
4
left)
and
for
varying
(Fig.
4
right)
duty
cycles.
In
both
paradigms
the
neuronal
responses
of
the
ascending
neuron
are
independent
of
SRI.
By
contrast,
the
responses
of
the
two
BNC1
neurons
increase
with
SRI
and
reach
saturation
at
about
35
ms
SRI
which
is
the
lower
margin
of
the
phonotactically
effective
range
(DOHERTY,
pers.
comm.).
This
low-pass
characteristic
occurs
in
the
50%
as
well
as
in
the
varying
duty
cycle
paradigm.
So
the
low-pas
filter
characteristic
does
not
depend
on
syllable
duration
or
syllable
number but
only
on
SRI.
Another
property
of
these
low-pass
neurons
showed
up
in
the
unequal
duty
cycle
experiment.
The
response
falls
when
the
pause
between
the
syllables
is
shortened
to
about
7
ms,
increases
again
with
even
shorter
pause
durations
and
reaches
maximum
with
pause
durations
of
1
ms
which
is
nearly
a
tone
burst
of
about
120
ms.
This
neuronal
property
correlates
with
some
aspects
of
phonotactic
behaviour.
Animals
that
show
phonotactic
track-
ing
to
tone
bursts
fail
to
do
so
when
the
pause
between
the
syllables
is
7-10
ms,
but
start
tracking
again
when
the
pause
is
increased
above
that
value
(DOHERTY,
in
prep.).
Some
properties
of
band-pass
neurons
are
shown
in
more
detail
in
Fig.
5.
Here
the
neuronal
responses
were
tested
for
a
greater
number
of
stimulus
re-
presentations
with
different
temporal
patterns.
The
same
sound
energy
and
the
same
number
of
chirps
elicit
2-5
times
more
spikes
given
a
pattern
within
N•57
•09•419
Auditory
brain
neurons
47
N•130
nn.a.m..m..a.na.amaannuo
Own
no
.
60
.19713
422
20
30
10
N•
293
rL
A
ir
ir
100
ZO
ms
100
200
ms
100
200
ens
Fig.
5.
Responses
of
the
neuron
BNC2a
to
chirps
with
different
syllable
repetition
intervals.
SRI
was
18
ms
(left),
34
ms
(middle)
and
98
ms
(right);
PST-histograms
(top)
and
sequential
dot
displays
(bottom)
are
shown
with
identical
time
scale
for
73
chirps
for
each
pattern;
black
bars
indicate
the
stimulus
pattern;
the
insets
show
recording
examples;
N
gives
the
total
number
of
spikes
for
all
chirps,
x:
the
mean
spikes
per
chirp;
sound
frequency:
5
kHz,
intensity:
80
dB
SPL.
the
natural
range
of
the
calling
song
than
outside
that
range.
The
number
of
action
potentials
does
not
decrease
with
the
number
of
consecutive
chirps
but
during
a
single
chirp.
Therefore,
a
longer
chirp
and
more
syllables
should
not
elicit
more
spikes.
To
test
for
this
the
syllable
number
in
a
chirp
was
varied
between
I
and
15
at
optimal
SRI's.
Band-pass
neurons
exhibited
about
the
same
number
of
action
potentials
to
chirps
with
varying
syllable
numbers
but
constant
SRI
when
the
syllable
number
was
5
or
more.
So
it.
seems
that
in
these
band-pass
cells
the
responses
do
not
vary
-
above
a
critical
value
-
with
syllable
number
and
chirp
duration
when
the
SRI
is
optimal.
Though
precise
synchronization
of
the
response
with
the
syllable
onset
and
copying
of
the
syllable
length
do
not
occur
in
a
single
chirp,
pooling
of
a
greater
number
of
chirps
reveals
some
synchrony
at
intermediate
and
long
SRI's.
Not
every
syllable
elicits
a
spike
-
especially
the
late
syllables
in
a
chirp
often
fail
to
do
so
-,
but
a
spike
is
mostly
correlated
with
a
syllable.
This
might
be
ex-
plained
partly
by
the
temporal
behaviour
of
the
graded
potentials.
They
are
syn-
chronized
with
the
stimulus
pattern,
but
not
all
syllables
evoke
a
graded
potential
and
not
all
graded
potentials
elicit
an
action
potential.
So
the
cell
may
get
input
from
neurons
that
copy
all
patterns,
but
the
band-pass
properties
can
not
be
explained
yet
by
summation
of
the
graded
potentials
that
were
evoked
by
nonspecific
copying
neurons
and
additional
inputs
seem
to
be
necessary.
If
a
band-pass
neuron
detects
only
the
conspecific
syllable
repetition
interval
then
it
should
not
fire
before
at
least
one
syllable
period
of
time
has
been
proc-
essed
by
the
previous
network.
This
is
at
least
30-45
ms
plus
the
time
the
infor-
mation
needs
to
reach
the
brain
(about
20
ms).
So
a
latency
of
the
band-pass
neuron
of
at
least
50-60
ms
is
expected.
When
pooled
for
a
number
of
chirp
representations
the
latencies
of
band-pass
cells
are
in
most
cases
indeed
60-80
ms
(Fig.
5).
But
there
are
some
cases
where
a
first
response
occurs
at
35-50
ms,
which
is
too
early
for
recognition
of
the
correct
SRI.
So
it
might
be
that
the
band-pass
cells
also
get
input
from
the
shorter-latency
auditory
neurons
that
have
no
temporal
filter
characteristics
like
low-pass
or
high-pass
cells.
The
non-
specific
inputs
could
depolarize
the
band-pass
cells
near
to
or
possibly
over
48
SCHILDBERGER
7
•••••,,,
E
,s,
Response
Magnitude
/
Chirp
90
0
/0
_//
Phonotox
is
---
O`
A
0
0
._
\0
0
\
1,
-75
25
a
I I
i
t
0
ke
0 0
0
\
0
0
-25
26
50
74
98
Syllable
Repetition
Interval
lrnS1
Fig. 6.
Relative
response
magnitude
of
auditory
brain
neurons
with
specific
filter
characteristics
to
chirps
varying
in
syllable
repetition
interval.
Temporal
structures
of
stimuli
are
illustrated
above;
data
points
with
squares
are
from
the
neuron
BNC1d,
triangles
from
BNC2b
and
circles
from
BNC2a.
Closed
symbols
for
three
such
cells
are
connected
by
lines.
The
open
symbols
show
the
responses
of
other
examples
of
these
identified
neurons
in
different
animals
to
indicate
variability.
The
hatched
areas
show
the
relative
effectiveness
(right
ordinate)
of
the
syllable
repetition
intervals
in
eliciting
phonotaxic
tracking
in
Gryllus
bim
sculatus
(negative
slope
hatch,
replotted
from
DOHERTY,
in
prep.);
Sound
frequency:
5
kHz,
intensity:
80
dB
SPL.
threshold.
if
excitation
now
arrives
via
the
low-pass/high-pass
ANDing
the
cell
will
start
to
fire.
Therefore,
in
addition
to
low-
and
high-pass
filters,
unspecific
inputs
to
the
band-pass
cells
might
be
necessary
for
the
recognition
process.
The
responses
of
descending
neurons
in
Fig.
2
showed
either
long
latencies
and
band-pass
characteristics
or
short
latencies
and
responses
that
were
similar
for
all
patterns.
Together
with
the
overlapping
of
projection
fields
of
descending
neurons
with
BNC2,
BNCI
and
ascending
cells,
this
may
indicate
that
auditory
information
travels
via
different
pathways
to
structurally
and
functionally
differ-
ent
descending
neurons.
Slow
but
pattern
specific
signals
could
reach
descending
neurons
via
the
various
filters,
while
more
rapid
but
nonspecific
signals
could
reach
descending
cells
directly
via
ascending
or
nonspecific
BNCI
cells.
In
the
cricket
brain,
there
are
neurons
with
responses
dependent
upon
the
temporal
structure
of
the
stimulus.
As
summarized
in
Fig.
6,
neurons
with
low-
pass,
high-pass
and
band-pass
characteristics
are
found.
The
pattern
specificity
Auditory
brain
neurons
49
of
the
latter
is
closely
correlated
with
that
of
the
animals,
when
their
phonotactic
behaviour
is
measured
with
identical
paradigms.
Concerning
the
temporal
selectivi-
ty
similar
types
of
auditory
filter
neurons
are
found
in
the
cricket
as
well
as
in
the
amphibian
brain
(ROSE
&
CAPRANICA,
1984).
It
remains
unclear
what
mechanisms
underlie
this
neural
filtering
and
if
these
mechanisms
are
similar
in
the
vertebrate
and
the
invertebrate
brain.
Acknowledgements.
I
thank
F.
HUBER,
J.
DOHERTY
and
J.
THORSON
for
critical
comments
on
the
manuscript.
In
particular,
some
of
the
ideas
on
band-pass
neurons
were
developed
together
with
J.
THORSON.
I
also
like
to
thank
H.
BAM-
BERG
and
M.
OBERMAYER
for
help
with
the
photography.
REFERENCES
BOYD,
P.,
R.
KOHNE,
S.
SILVER
&
B.
LEWIS,
1984:
Two-tone
suppression
and
song
coding
by
ascending
neurons
in
the
cricket
Gryllus
cam
pestris
L.
J.
comp.
Physiol.
A
154:
423-430.
ESCH,
H.,
F.
HUBER
&
D.W.
WOHLERS,
1980:
Primary
auditory
neurons
in
crickets:
Physiology
and
central
projections.
J.
comp.
Physiol.
137:
27-38.
HUTCHINGS,
M.
&
B.
LEWIS,
1981:
Response
properties
of
primary
auditory
fibres
in
the
cricket
Teleogryllus
oceanicus
(Le
Guillou).
J.
comp.
Physiol.
143:
129-134.
NOCKE,
H.,
1972:
Physiological
aspects
of
sound
communication
in
crickets
(Gryllus
ca
m
pestris
L.).
J.
comp.
Physiol.
80:
141-162.
POPOV,
A.V.
&
V.F.
SHUVALOV,
1977:
Phonotactic
behavior
in
crickets.
J.
comp.
Physiol.
119:
111-126.
-
&
A.M.
MARKOVICH,
1982:
Auditory
interneurons
in
the
prothoracic
ganglion
of
the
cricket,
Gryllus
bimaculatus.
J.
comp.
Physiol.
146:
351-359.
ROSE,
G.
&
R.R.
CAPRANICA;
1984:
Processing
of
amplitude
modulated
sounds
by
the
auditory
midbrain
of
two
species
of
toads:
matched
temporal
filters.
J.
comp.
Physiol.
A
154:
211-219.
SCHILDBERGER;
K.,
1984a:
Multimodal
interneurons
in
the
cricket
brain:
Prop-
erties
of
identified
extrinsic
mushroom
body
cells.
J.
comp.
Physiol.
A
154:
71-79.
-
1984b:
Temporal
selectivity
of
identified
auditory
neurons
in
the
cricket
brain.
J.
comp.
Physiol.
A
(in
press).
THORSON,
J.,
T.
WEBER
&
F.
HUBER,
1982:
Auditory
behavior
of
the
cricket
II.
Simplicity
of
calling
song
recognition
in
Gryllus,
and
anomalous
phono-
taxis
at
abnormal
carrier
frequencies.
J.
comp.
Physiol.
146:
361-378.
WEBER,
T.,
J.
THORSON
&
F.
HUBER,
1981:
Auditory
behavior
of
the
cricket.
I.
Dynamics
of
compensated
walking
and
discrimination
paradigms
on
the
Kramer
treadmill.
J.
comp.
Physiol.
141:
215-232.
WOHLERS,
D.W.
&
F.
HUBER,
1978:
Intracellular
recording
and
staining
of
cricket
auditory
interneurons
(Gryllus
cam
pestris
L.,
Gryllus
bimaculatus
de
Geer).
J.
comp.
Physiol.
127:
11-28.
-
&
F.
HUBER,
1982:
Processing
of
sound
signals
by
six
types
of
neurons
in
the
prothoracic
ganglion
of
the
cricket,
Gryllus
cam
pestris
L.
J.
comp.
Physiol.
146:
161-173.
Address:
Max-Planck-Institut
far
Verhaltensphysiologie,
Abt.
Huber,
D-8131
Seewiesen,
Fed.
Rep.
Germany