Identified auditory neurons in the cricket Gryllus rubens: temporal processing in calling song sensitive units


Farris, H.E.; Mason, A.C.; Hoy, R.R.

Hearing Research 193(1-2): 121-133

2004


This study characterizes aspects of the anatomy and physiology of auditory receptors and certain interneurons in the cricket Gryllus rubens. We identified all U-shaped ascending interneuron tuned to frequencies > 15 kHz (57 dB SPL threshold at 20 kHz). Also identified were two intrasegmental 'omega'-shaped interneurons that were broadly tuned to 3-65 kHz, with best sensitivity to frequencies of the male calling song (5 kHz, 52 dB SPL). The temporal sensitivity of units excited by calling song frequencies were measured using sinusoidally amplitude modulated stimuli that varied in both modulation rate and depth, parameters that vary with song propagation distance and the number of singing males. Omega cells responded like low-pass filters with a time constant of 42 ms. In contrast, receptors significantly coded modulation rates up to the maximum rate presented (85 Hz). Whereas omegas required [approximately]65% modulation depth at 45 Hz (calling song AM) to elicit significant synchrony coding, receptors tolerated a [approximately]50% reduction in modulation depth up to 85 Hz. These results suggest that omega cells in G. rubens might not play a role in detecting song modulation per se at increased distances from a singing male.

Available
online
at
www.sciencedirect.com
.
1%
SCIENCE
CLDIRECT*
HEARING
RESEARC
H
ELSEVIER
Hearing
Research
193
(2004)
121-133
www.elsevier.com/locate/heares
Identified
auditory
neurons
in
the
cricket
Gryllus
rubens:
temporal
processing
in
calling
song
sensitive
units
Hamilton
E.
Farris
*,
Andrew
C.
Mason
1
,
Ronald
R.
Hoy
Section
of
Neurobiology
and
Behavior,
Mudd
Hall,
Cornell
University,
Ithaca,
NY
14850,
USA
Accepted
16
February
2004
Available
online
21
April
2004
Abstract
This
study
characterizes
aspects
of
the
anatomy
and
physiology
of
auditory
receptors
and
certain
interneurons
in
the
cricket
Gryllus
rubens.
We
identified
an
`L'-shaped
ascending
interneuron
tuned
to
frequencies
>15
kHz
(57
dB
SPL
threshold
at
20
kHz).
Also
identified
were
two
intrasegmental
'omega'-shaped
interneurons
that
were
broadly
tuned
to
3-65
kHz,
with
best
sensitivity
to
frequencies
of
the
male
calling
song
(5
kHz,
52
dB
SPL).
The
temporal
sensitivity
of
units
excited
by
calling
song
frequencies
were
measured
using
sinusoidally
amplitude
modulated
stimuli
that
varied
in
both
modulation
rate
and
depth,
parameters
that
vary
with
song
propagation
distance
and
the
number
of
singing
males.
Omega
cells
responded
like
low-pass
filters
with
a
time
constant
of
42
ms.
In
contrast,
receptors
significantly
coded
modulation
rates
up
to
the
maximum
rate
presented
(85
Hz).
Whereas
omegas
required
—65%
modulation
depth
at
45
Hz
(calling
song
AM)
to
elicit
significant
synchrony
coding,
receptors
tolerated
a
—50%
reduction
in
modulation
depth
up
to
85
Hz.
These
results
suggest
that
omega
cells
in
G.
rubens
might
not
play
a
role
in
detecting
song
modulation
per
se
at
increased
distances
from
a
singing
male.
©
2004
Elsevier
B.V.
All
rights
reserved.
Keywords.•
Sinusoidal
amplitude
modulation;
Temporal
modulation
transfer
function;
Omega
cells;
Afferents;
INT-1;
AN2;
Ultrasound;
Gryllus
bimaculatus;
Teleogryllus
oceanicus;
Calling
song
1.
Introduction
In
most
species
of
crickets
(Gryllidae),
males
pro-
duce
calling
songs
that
function
as
sexual
advertise-
ment
signals
allowing
conspecifics
to
discriminate
among
singing
males
(Walker,
1957;
Alexander,
1960).
In
addition
to
song
carrier
frequency
(Popov
and
Corresponding
author.
Present
address:
Neuroscience
Center,
Louisiana
State
University
Health
Science
Center,
2020
Gravier
Street,
New
Orleans,
LA
70112,
USA.
Tel.:
+1-504-599-0865;
fax:
+1-504-599-0891.
E-mail
addresses.•
hfarri@lsuhsc.edu
(H.E.
Farris),
amason@
utsc.utoronto.ca
(A.C.
Mason).
1
Present
address:
Division
of
Life
Sciences,
University
of
Toronto-
Scarborough,
1265
Military
Trail,
Scarborough,
Ont.,
Canada
M1C
1A4.
Abbreviations.•
hfAN,
high-frequency
ascending
neuron;
ON1,
ON2,
omega
neuron
1,
2;
SAM,
sinusoidal
amplitude
modulation;
TMTF,
temporal
modulation
transfer
function;
INT-1,
interneuron
1;
AN2,
ascending
neuron
2
0378-5955/$
-
see
front
matter
©
2004
Elsevier
B.V.
All
rights
reserved.
doi:10.1016/j.heares.2004.02.008
Shuvalov,
1977)
and
relative
song
intensity
(Farris
et
al.,
1997),
one
of
the
most
important
parameters
for
discriminating
songs
is
the
temporal
structure
of
a
song's
amplitude
modulation
(AM).
Indeed,
phono-
tactic
decisions
which
are
known
to
be
ultimately
correlated
to
species
recognition
and
character
dis-
placement
(Walker,
1974;
Shaw
and
Herlihy,
2000)
have
been
shown
to
be
based
on
the
AM
rate
of
calling
songs
(Walker,
1957;
Doherty,
1985;
Popov
and
Shuvalov,
1977;
Ulagaraj
and
Walker,
1975;
Doolan
and
Pollack,
1985).
In
the
field,
however,
variance
in
song
AM
is
not
limited
to
AM
rate.
As
a
song
propagates
through
a
complex
environment,
the
AM
structure
can
be
distorted
due
to
reverberations
(i.e.,
multipaths),
frequency-dependent
absorption
and
the
addition
of
sounds
produced
by
other
animals
(see
Forrest,
1994;
Romer,
1998
for
review).
In
the
time
domain,
such
environmental
filtering
smears
the
AM
structure
of
a
calling
song
and
effectively
removes
the
intervals
between
song
pulses
(Simmons,
1988).
122
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
Compared
to
frogs
(i.e.,
species
that
produce
repetitive
or
cricket-like
song;
Rose
and
Capranica,
1985),
however,
there
are
relatively
few
data
for
the
effects
of
variance
in
song
modulation
depth
on
the
physiolog-
ical
coding
or
behavioral
attractiveness
of
calling
songs
in
insects
in
general
(Machens
et
al.,
2001,
Prinz
and
Ronacher,
2002)
and
crickets
in
particular;
most
studies
of
temporal
sensitivity
in
acoustic
insects
have
been
concerned
only
with
sensitivity
to
AM
rate
(i.e.,
pulse
duration,
inter-pulse
interval,
etc.),
while
fore-
going
tests
on
AM
depth.
Although
several
studies
have
indirectly
assayed
the
effects
of
temporal
distor-
tion
by
measuring
responses
to
either
the
simultaneous
presentation
of
multiple
songs
(i.e.,
two
stimuli
with
ideal
pulsed
or
"DC-on"
and
"DC-off"
structures;
Schildberger,
1984;
Doherty,
1985;
Pollack,
1986;
Romer
and
Krusch,
2000)
or
the
presentation
of
songs
at
various
distances
(Romer
and
Bailey,
1986),
these
experimental
paradigms
do
not
systematically
measure
the
limits
of
detection
or
discrimination
of
changes
in
modulation
depth.
From
a
psychophysical
point
of
view,
a
more
general
measure
of
temporal
sensitivity
can
be
derived
from
measuring
a
system's
temporal
modulation
transfer
function
(TMTF).
Considered
a
systems
analysis
ap-
proach,
the
TMTF
explains
the
amount
of
modulation
necessary
for
the
detection
of
modulation
at
each
modulation
frequency
(Viemeister,
1977,
1979).
Al-
though
the
structure
of
the
AM
stimuli
used
to
measure
TMTFs
may
vary
across
experiments
[e.g.,
sinusoidal
AM
(SAM),
pulsed
AM
and
square
AM]
the
indepen-
dent
variables
are
always
the
modulation
frequency
and
depth.
In
this
paper,
we
used
SAM
stimuli
that
varied
in
modulation
frequency
and
depth,
to
measure
the
TMTF
of
identified
auditory
neurons
in
the
cricket,
Gryllus
rubens
(Gryllidae,
Gryllinae).
The
calling
song
of
G.
rubens
consists
of
a
trill
(i.e.,
sound
pulses
repeated
at
a
constant
interval,
Romer,
1998)
of
11
ms
pulses
of
4.5
kHz
at
45
pulses/s
at
22
°C
(Walker,
1962;
Bentley
and
Hoy,
1972;
Doherty
and
Callos,
1991).
Although
no
prior
studies
have
characterized
the
anatomy
and
physiology
of
individual
auditory
units
in
G.
rubens,
the
rationale
for
using
this
cricket
to
measure
temporal
sensitivity
is
related
to
its
trilled
calling
song.
Without
the
added
complexity
of
chirps
(i.e.,
more
than
one
in-
terpulse
interval),
the
trilled
song
of
G.
rubens
is
ideal
for
examining
sensitivity
to
amplitude
modulation
at
one
time
scale
only,
the
temporal
structure
of
the
pulses
in
the
trill.
Although
the
stimuli
used
in
previous
studies
of
temporal
sensitivity
in
crickets
almost
certainly
contain
more
modulation
than
that
found
for
songs
in
the
field,
the
sinusoidally
modulated
stimuli
used
here
may
rep-
resent
the
other
extreme
in
that
there
are
no
true
'offs'
to
the
stimulus.
The
amplitude
of
the
SAM
stimulus
is
always
changing
and
only
instantaneously
goes
to
0
when
100%
modulated.
Thus,
whereas
the
artificial
stimuli
used
in
most
previous
studies
represent
a
best
case
scenario
for
signal
transmission,
the
SAM
stimuli
used
here
hypothetically
represent
signal
transmission
that
is
less
than
ideal.
2.
Methods
2.1.
Subject
animals
The
breeding
colony
of
G.
rubens
was
started
in
1997
from
individuals
sound
trapped
in
Alachua
County,
Florida,
reared
under
a
14
L/10
D
h.
light
schedule
and
fed
Cat
Chow'
ad
libitem.
This
project
adhered
to
the
Cornell
University
guidelines
for
animal
care
and
use.
2.2.
Acoustic
stimuli
All
stimuli
were
generated
using
Tucker
Davis
Technologies
(TDT)
16
bit,
digital-to-analog
convert-
ers
and
custom
written
software
(6
and
10
µs
sampling
period
for
tuning
curve
and
SAM
stimuli,
respec-
tively).
Stimuli
were
amplified
using
a
Harman/Kar-
don
HK6150
integrated
amplifier
and
broadcast
from
Radio
Shack
Super
tweeters
(cat.
no.
40-1310b)
located
30
cm
from
the
preparation
and
positioned
normal
to
the
longitudinal
axis
of
the
cricket.
Stimulus
amplitude
was
adjusted
using
TDT
PA4
programmable
attenuators.
The
stimuli
were
calibrated
at
the
position
of
the
test
animal
using
online
com-
parisons
of
the
r.m.s.
voltage
of
the
stimuli
to
that
generated
by
a
B&K
4220
pistonphone
calibrator
(125
ms
duration
samples).
The
calibration
system
included
a
B&K
4135
1/4
inch
microphone
(0°
angle
of
inci-
dence),
B&K
2639
preamp
and
a
B&K
5935
micro-
phone
power
supply.
All
sound
pressure
levels
(dB
SPL)
are
referenced
to
20
µPa.
The
onset
and
offset
ramps
of
pulses
used
to
measure
tuning
curves
are
raised
cosine.
Total
harmonic
distortion
of
the
system
was
determined
to
be
<1%
(-40
dB)
using
a
Hewlett—
Packard
3562A
signal
analyzer.
2.3.
Neurophysiological
recordings
The
experimental
procedures
used
here
are
similar
to
those
in
Mason
et
al.
(1998).
Briefly,
cold-anethe-
sized
female
crickets
were
mounted
on
a
platform
ventral
side
up
in
a
foam
lined
Faraday
cage
that
reduced
acoustic
and
electrical
noise.
The
prothoracic
legs
were
extended
laterally
and
waxed
at
the
tarsi
to
stainless
steel
bars.
After
exposing
the
prothoracic
ganglion
by
removing
the
ventral
cuticle,
the
ganglion
was
supported
on
a
stainless
steel
"spoon"
that
also
served
as
the
ground
electrode.
Electrical
activity
in
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
123
auditory
units
in
the
prothoracic
ganglion
was
re-
corded
using
thin-walled
(1.0
mm
o.d.)
borosilicate
glass,
micropipette
electrodes
filled
at
the
tip
with
2.5%
Lucifer
Yellow
(Sigma)
and
backfilled
with
0.1
mol
1
-1
LiCl.
Electrode
resistance
varied
from
30
to
120
MO
.
The
search
stimulus
used
to
locate
auditory
units
consisted
of
a
pair
of
20
ms
pulses
of
5
and
40
kHz
presented
asynchronously
(25
ms
interpulse
in-
terval).
After
amplification
(AM
systems
model
1600
DC
followed
in
some
experiments
by
a
custom
built
AC
amplifier),
the
neural
responses
were
digitized
(100
µs
sampling
period)
using
a
TDT
AD3
and
System
II
Array
Processor.
By
injecting
hyperpolarizing
current
(0.2-1.2
nA)
after
recording,
cells
were
stained
with
Lucifer
Yellow
for
anatomical
identification.
After
staining,
the
ganglia
were
dissected
and
fixed
(16-24
h)
in
4%
paraformaldehyde
(not
buffered),
dehydrated
in
an
ascending
ethanol
series
and
cleared
using
me-
thyl
salicylate.
Stained
cells
were
photographed
and
digitized
as
whole
mounts
using
a
Leitz
Dialux
20
and
BioRad
MRC-600
confocal
microscopes,
respectively.
No
cells
were
included
in
the
study
unless
staining
provided
adequate
identification.
2.4.
Experimental
protocols
After
penetrating
a
cell,
we
measured the
temporal
sensitivity
of
auditory
units
excited
by
5
kHz
by
presenting
a
series
of
sinusoidal
amplitude
modulated
(SAM)
tones
(5
kHz,
90
dB
SPL)
that
varied
in
modulation
rate
and
depth.
The
SAM-series
consisted
of
five
replicates
of
four
modulation
frequencies
(25,
45,
65,
85
Hz)
at
five
modulation
depths
(100%,
75%,
50%,
25%,
0%)
to
yield
a
total
of
100
stimuli.
Each
stimulus
consisted
of
ten
cycles
of
modulation
(i.e.,
stimulus
duration
varied
from
118
to
400
ms).
Al-
though
neural
adaptation
is
likely
to
occur
throughout
the
durations
of
these
gated
stimuli
(i.e.,
stimulus
duration
is
potentially
shorter
than
the
neural
adap-
tation
time
and
thus
spike-train
analysis
includes
onset
responses;
Givois
and
Pollack,
2000;
Gleich
and
Klump,
1995),
we
chose
the
10-cycle
SAMs
for
the
following
reasons:
(1)
stimuli
are
at
least
twice
the
number
of
cycles
required
to
elicit
phonotaxis
in
gryllids
and
are
thus
behaviorally
relevant
(Nolen
and
Hoy,
1986a);
(2)
best
measures
of
temporal
processing
are
not
necessarily
made
while
auditory
units
are
in
a
more
adapted
state
(Epping,
1990;
Prinz
and
Ron-
acher,
2002)
and
longer
duration
stimuli
can
reduce
sensitivity
to
amplitude
modulation
(Smith
and
Brachman,
1980;
Smith
et
al.,
1985;
see
Frisina,
2001
for
review);
(3)
psychophysical
studies
in
humans
comparing
continuous
and
gated
stimuli
show
gating
only
reduces
thresholds
at
the
slowest
modulation
rates
without
systematically
affecting
the
low-pass
cutoff
frequency
(Yost
and
Sheft,
1997);
(4)
use
of
gated
stimuli
facilitates
comparisons
with
data
from
numerous
other
studies
of
temporal
processing
in
crickets
and
other
insects
(e.g.,
Mason
et
al.,
1998;
Horseman
and
Huber,
1994;
Schildberger,
1984;
Tougaard,
1998);
(5)
it
is
not
known
whether
the
conditions
for
adaptation
are
met
in
the
field.
To
re-
duce
the
spectral
splatter
associated
with
decreased
modulation
depth,
0.25
ms
ramps
(i.e.,
longer
than
one
wavelength)
were
applied
to
the
onsets
and
offsets
of
the
SAM
stimuli.
Cells
which
exhibited
spontane-
ous
activity
well
after
penetration
(>30
s)
were
not
included.
Behavioral
and
neural
detection
thresholds
are
de-
termined
by
temporal
integration
of
pressure
and/or
power
(Heil
and
Neubauer,
2001;
Gollisch
et
al.,
2002;
Farris
and
Hoy,
2000).
The
average
power
of
a
SAM
stimulus
varies
with
modulation
depth
as
(1
+
mod
depth
2
/
2
)
x
J,
where
h
is
the
average
power
when
mod
depth
=
0
(Viemeister,
1979).
To
ensure
that
the
variance
in
neural
discharges
was
due
to
the
cells'
ability
to
follow
changes
in
amplitude
modulation
and
not
overall
stimulus
power
(Frisina
et
al.,
1990;
Pollack
and
El-Feghaly,
1993),
SAM
stimulus
amplitude
was
corrected
online
so
that
each
SAM
stimulus
had
the
same
energy (i.e.,
the
integral
of
power)
per
modulation
cycle
(90
dB
re.
20
µPa
tone).
Furthermore,
this
correction
facilitated
com-
parison
of
the
cells'
ability
to
detect
sinusoidal
am-
plitude
modulation
with
the
ability
to
detect
changes
in
the
amplitude
of
single
pulses
(see
below).
Spike
times
were
measured
from
the
digitized
voltage
traces
using
a
custom
written
window
discriminator
of
voltage
level
or
voltage
change
(derivative).
The
SAM
series
was
presented
once
per
preparation.
Following
the
presentation
of
the
SAM
series,
the
frequency
tuning
of
a
cell
was
determined
by
mea-
suring
the
minimum
sound
pressure
level
(±3
dB)
re-
quired
to
elicit
at
least
one
action
potential
in
3/5
stimulus
presentations.
Stimuli
were
pure
tones
that
varied
in
frequency
from
3
to
65
kHz
(10
ms
pulses,
1
ms
ramps,
400
ms
ISI)
and
were
presented
in
as-
cending
order.
Using
the
responses
recorded
during
measurement
of
the
tuning
curve,
input/output
(I/O)
functions
at
5
kHz
were
measured
by
calculating
the
mean
number
of
spikes
elicited
for
the
five
repetitions
presented
at
each
intensity
step.
2.5.
Analysis
Methods
used
to
assess
the
amount
of
modulation
in
the
voltage
response
of
cells
sensitive
to
5
kHz
are
modeled
after
those
in
Rees
and
Palmer
(1989).
Analysis
of
response
modulation
was
accomplished
by
measuring
the
vector
strength
(r)
(i.e.,
length
of
the
mean
vector)
of
the
spike
times,
which
is
calculated
by
the
following
equation:
124
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
2
7
,
2
r=
(
cos(a
i
))
E
sin(a
i
))
(
n n
where
a
is
the
angular
conversion
of
the
spike
times
relative
to
the
modulation
period
and
n
is
the
number
of
spikes
(Rees
and
Palmer,
1989;
but
see
Zar,
1999).
The
vector
strength
coefficient,
r
is
inversely
correlated
to
the
variance
in
the
spike
times
and
can
vary
from
0
to
1;
values
near
0
represent
the
case
in
which
there
is
no
modulation
in
the
spike
train,
whereas
values
close
to
1
represent
the
opposite
case
in
which
spike
times
are
perfectly
concentrated
at
some
phase
in
the
modulation
frequency.
Statistical
significance
of
r
(assessing
whether
the
spike
times
are
uniformly
distributed
in
the
modu-
lation
cycle)
was
calculated
using
a
Rayleigh
test
for
circular
uniformity
(Zar,
1999).
Vector
strength
values
were
also
used
to
calculate
the
modulation
gain
(in
dB)
in
the
spike
train
using
the
following
equation:
Gain
=
20
x
log
Percentage
of
modulation
depth
spike
train
Percentage
of
modulation
depth
stimulus
where
the
modulation
depth
of
the
spike
train
is
defined
as
200r;
because
a
spike
train
with
100%
sinusoidal
modulation
produces
a
vector
strength
of
0.5,
conver-
sion
of
r
to
percent
modulation
depth
requires
a
factor
of
200
(Rees
and
Palmer,
1989;
Gleich
and
Klump,
1995).
This
conversion
of
vector
strength
to
gain
(in
dB)
facilitates
the
analysis
of
the
filter
shape
(Gleich
and
Klump,
1995;
Frisina
et
al.,
1990;
Kim
et
al.,
1990)
and
thus
calculation
of
time
constants.
Furthermore,
this
conversion
quantifies
the
signal-to-noise
ratio
received
by
cells
post-synaptic
to
the
recording
site
independent
of
spike
number
(Rees
and
Palmer,
1989;
Frisina
et
al.,
1990;
but
see
Palmer,
1995
for
review).
Comparisons
of
modulation
gain
between
cell
classes
were
done
using
a
t
test,
which
conservatively
assumes
unequal
variances
(Welch's
approximate
t;
Zar,
1999).
We
used
the
slope
of
the
I/O
curves
at
5
kHz
(spikes/
single
10
ms
pulse)
to
assess
the
relationship
between
AM
detection
and
intensity
discrimination
(Wojtczak
and
Viemeister,
1999).
I/O
responses
were
normalized
to
those
for
the
90
dB
SPL
pulses,
the
power
of
the
SAM
stimulus,
to
control
for
any
variance
in
intensity
dis-
crimination
threshold
that
may
be
correlated
to
the
power
of
the
carrier
(Farris
et
al.,
1997;
McGill
and
Goldberg,
1968;
Long
and
Cullen,
1985;
Kohlrausch,
1993;
Wojtczak
and
Viemeister,
1999).
Subsequently,
we
solved
for
each
cell's
relative
I/O
slope
by
finding
the
least-squares
solution
of
the
following
logistic
equation:
Proportion
of
the
response
to
90
dB
exp(slope
x
dB
stimulus
+
intercept)
1
+
exp(slope
x
dB
stimulus
+
intercept)'
which
is
appropriate
for
analyzing
dependent
variables
bounded
by
0
and
1.
Differences
in
the
mean
I/O
slopes
(a)
U
(c)
U
200
gm
(b)
U
(d±
7.0.
1
C
Fig.
1.
Examples
of
the
anatomy
of
auditory
neurons
recorded
in
the
prothoracic
ganglion
(schematics;
ventral
view;
top
is
anterior)
of
G.
rubens.
Sketches
were
produced
by
digitally
tracing
confocal
micrographs.
(a)
hfAN
(high-frequency
ascending
neuron).
(b)
ON1
(omega
neuron
1).
(c)
ON2
(omega
neuron
2).
(d)
Receptor
neuron.
I
-
(a)
.1
I I
1 1 1 1
1
I I I
1
-
(b)
_
1
I I
1
1
1
1
1
1 1 1 1 1
-
(c)
1
111•11
90
85
80
75
70
65
60
55
90
85
80
75
70
65
60
55
50
T
hres
ho
ld
dB
(
SPL)
90
85
80
75
70
65
60
55
50
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
125
of
different
cell
classes
were
tested
using
a
two
sample
t
test
(two-tailed).
Because
one
cell
class
had
a
small
sample
size
(i.e.,
receptors,
see
Section
3)
potentially
affecting
I/O
slope
variance,
a
variance
ratio
test
was
used
to
determine
whether
equal
or
unequal
variance
t
tests
were
applied
(Zar,
1999)
and
included
only
those
slopes
that
were
statistically
significant
(P
<
0.05;
but
worst
fit,
P
=
0.006).
3.
Results
3.1.
Characterized
cells
and
frequency
sensitivity
Three
types
of
interneurons
were
recorded
and
stained
in
the
prothoracic
ganglion
of
G.
rubens
(Fig.
1).
Nomenclature
for
these
cells
is
based
on
their
similarity
to
auditory
units
well
known
in
other
gryllines
(Casaday
and
Hoy,
1977;
Wohlers
and
Huber,
1982;
Stiedl
et
al.,
1997).
The
high-frequency
ascending
neuron
(hfAN;
Fig.
1(a)),
characterized
by
its
laterally
extending
den-
dritic
processes,
ascending
axon
and
contralateral
cell
body
is
sensitive
to
frequencies
>15
kHz
with
best
sen-
sitivity
at
20
kHz
(57
dB
SPL,
N
=
3,
Fig.
2(a)).
In
addition
to
this
neuron's
excitatory
response
to
ultra-
sound,
frequencies
from
3
to
6
kHz
elicited
inhibition
in
some
preparations
(Figs.
2(a)
and
3).
The
anatomy
and
physiology
of
this
cell
are
indistinguishable
from
that
of
AN2
and
Int-1
in
Gryllus
bimaculatus
and
Teleogryllus
oceanicus,
respectively
(Moiseff
and
Hoy,
1983;
Nolen
and
Hoy,
1986b;
but
see
Schildberger
et
al.,
1989;
Hennig,
1988
for
review).
We
also
recorded
from
several
omega-shaped
intra-
segmental
neurons
(N
=
10,
Fig.
1(b)
and
(c)).
Cells
of
this
shape
have
been
previously
described
in
other
ensiferan
Orthoptera
(Casaday
and
Hoy,
1977;
Popov
et
al.,
1978;
Wohlers
and
Huber,
1982;
Romer
et
al.,
1988;
Mason
et
al.,
1998)
and
are
broadly
tuned
to
sounds
with
frequen-
cies
from
3
to
65
kHz,
with
a
sensitivity
peak
at
5
kHz
(
,
-
,
52
dB
SPL;
Fig.
2(b)
and
(c)).
As
in
previous
studies,
we
have
sorted
omega
neurons
into
two
subclasses,
omega
neu-
rons
1
and
2
(ON1
and
ON2).
Both
cells
display
the
classic
omega
shape
with
bilateral
arborizations
connected
by
a
large
arched
process
that
crosses
the
midline
of
the
gan-
glion
(Wohlers
and
Huber,
1982).
The
cell
body
extends
laterally
from
the
end
of
the
arched
midline
process.
Al-
though
based
on
only
one
preparation
which
did
not
in-
clude
staining
of
the
cell
body,
we
have
distinguished
a
second
omega
cell
class
(ON2)
due
to
the
presence
of
a
large
process
that
crosses
the
midline
of
the
ganglion
be-
tween
the
area
of
greatest
arborization
(Fig.
1(c)).
This
second
large
contralateral
process
is
also
the
distin-
guishing
character
for
ON2
in
G.
bimaculatus
(Wohlers
and
Huber,
1982).
Auditory
receptors
were
characterized
by
the
pres-
ence
of
an
axon
extending
laterally
from
the
prothoracic
3
10
70
Frequency
(kHz)
Fig.
2.
Frequency
tuning
of
interneurons.
Squares
are
the
mean
thresholds
(±SE)
necessary
to
elicit
an
action
potential
in
3/5
presen-
tations
(10
ms
duration,
1
ms
ramps).
Sample
sizes
were
3,
9
and
1
for
hfAN
(a),
ON1
(b)
and
ON2
(c),
respectively.
Open
circles
in
(a)
show
the
mean
thresholds
for
eliciting
inhibitory
post-synaptic
potentials
in
certain
preparations
of
hfAN
(N
=
2)
(Fig.
3).
ganglion
down
the
tympanal
nerve
(Fig.
1(d)).
Receptor
tuning
(N
=
4)
varied
from
frequencies
<10
kHz
to
broad
tuning
across
the
entire
stimulus
range
(Fig.
4).
Although
this
pattern
of
variable
tuning
appears
com-
mon
for
receptors
(Nocke,
1972,
Esch
et
al.,
1980;
Hutchings
and
Lewis,
1981;
Imaizumi
and
Pollack,
1999),
the
presence
of
units
tuned
below
4
kHz
raises
the
possibility
that
they
originate
in
the
subgenual
rather
than
the
tympanal
organ
[i.e.,
similar
to
the
low
126
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
(a)
(c)
I
(d)
(b)
Fig.
3.
Voltage
traces
of
responses
to
single
pulse
stimuli
of
10
ms
duration
(stimulus
example
in
bottom
trace).
(a)
Response
of
hfAN
to
20
kHz
(87
dB
SPL).
(b)
Inhibition
of
spontaneous
activity
in
hfAN
by
4
kHz
(74.7
dB
SPL).
The
spontaneous
activity,
which
was
only
initially
encountered
upon
penetrating
this
cell,
is
displayed
here
to
illustrate
the
scale
of
the
IPSP.
(c)
Response
of
ON1
to
5
kHz
(81.2
dB
SPL).
(d)
Response
of
a
receptor
to
5
kHz
(81.2
dB
SPL).
Calibration
scale:
10
mV
and
25
ms.
90
84
77
71
64
58
51
45
3
10
Frequency
(kHz)
Fig.
4.
Frequency
tuning
of
four
receptors.
Symbols
are
the
thresholds
necessary
to
elicit
an
action
potential
in
3/5
presentations
(10
ms
duration,
1
ms
ramps).
frequency
receptor
recorded
by
Nocke
(1972)].
Never-
theless,
their
anatomical
projections
into
the
auditory
neuropile
as
well
as
their
sensitivity
to
short
duration
(10
ms)
pulses
of
5
kHz
(i.e.,
—72.5
dB
SPL)
suggest
that
they
could
detect
and
process
the
calling
songs
of
gryl-
lines
(Nocke,
1971,
Kavanagh,
1987,
Nolen
and
Hoy,
1986a).
Thus,
we
have
included
this
small
sample
in
the
analysis
of
temporal
processing
below.
3.2.
Temporal
sensitivity
A
complete SAM
series
was
recorded
for
10
ONs
(including
one
ON2)
and
four
receptors
(data
include
one
ON
recording
in
which
a
frequency
tuning
curve
was
not
measured).
Example
voltage
traces
and
peri-
stimulus
time
histograms
(PSTHs)
of
the
responses
to
the
SAM
series
in
the
two
cell
classes
are
shown
in
Figs.
5-7,
respectively.
Whereas
periodicity
at
the
modulation
Depth
ON
I
Receptor
100%
t
75%
50%
V
25%
111
0%
222
ms
(
0
cyc
es)
45
Hz
SAM
Fig.
5. 5.
Voltage
trace
responses
of
an
ON1
and
a
receptor
to
a
single
SAM
series
at
45
Hz
AM
rate.
Peri-stimulus
time
histograms
of
the
responses
of
these
same
cells
to
the
entire
SAM
series
are
shown
in
Figs.
6
and
7.
frequency
in
the
receptor
histograms
(Fig.
7)
is
notice-
able
across
all
of
the
modulation
frequencies,
modula-
tion
in
the
PSTH
of
the
exemplar
ON1
response
(Fig.
6)
is
displayed
only
up
to
45
Hz.
Drawing
from
convention
in
the
field
of
psychoacoustics,
TMTFs
for
ONs
and
receptors
are
plotted
in
Fig.
8
as
the
modulation
depth
necessary
to
elicit
a
spike
train
with
statistically
signifi-
cant
(a
=
0.05)
modulation
(at
the
stimulus
modulation
frequency)
as
a
function
of
the
modulation
frequency
of
the
stimulus.
ONs
demonstrate
a
low-pass
filter
response
T
hres
ho
ld
dB
(
SPL)
A
,A
Tei
7
/
8
/
/
X
R
b
1
1
1 1
1
1
1
t
I
70
8
6
4
2
0
25
8
6
4
2
0
45
500
250
8
8
6
6
4
4
2
2
0
500
0
250
8
6
6
4
4
2
2
0
0
0
500
250
8
8
6
6
4
4
2
2
0
0
500
250
8
8
6
6
4
4
2
2
0
0
I
T*1 1
14
11
7
500
250
Nu
m
be
r
o
f
Sp
ike
s
65
85
8
6
4
2
0
8
6
4
2
0
100
100
200
100
8
6
4
2
0
411107
8
6
4
75
2
0
100
200
100
8
6
4
2
0
6
4
50
2
0
100
200
100
8
6
4
2
0
8
6
4
25
2
0
100
200
100
6
4
2
0-
1
'
h
1
100
200
8
6
4
0
2
0
100
(%
)
tp
daa
uog
emp
ow
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
127
Modulation
Rate
(Hz)
Time
(ms)
Fig.
6.
Peri-stimulus
time
histograms
of
the
response
of
an
ON1
to
the
20
different
modulation
stimuli.
Each
frame
shows
the
summed
response
to
five
replicates
of
the
10-cycle
stimulus.
Note
that
for
these
histograms,
binwidth
equals
0.1
times
the
modulation
period
(e.g.,
binwidth
at
25
Hz
is
0.1
times
40
ms)
which
gives
the
appearance
of
variation
in
spike
number
between
the
tones.
whereas
receptors
show
an
all-pass
filter
response
(Fig.
8).
Whereas
vector
strength
values
(r)
that
were
not
significantly
clumped
(i.e.,
when
P
>
0.05)
were
thrown
out
for
the
TMTF
in
Fig.
8,
the
values
still
represent
the
relative
distribution
of
spikes
in
the
modulation
cycle
period.
In
order
to
make
use
of
all
of
the
r
values
to
assess
the
response
characteristics
of
the
two
classes
of
neurons,
we
calculated
the
amount
of
modulation
gain
(see
Section
2)
in
the
spike
train
(Fig.
9).
These
calcu-
lations
reveal
that
in
contrast
to
receptors,
in
which
modulation
is
enhanced
in
the
neural
code
at
most
modulation
frequencies,
modulation
in
the
spike
trains
of
omegas
is
reduced
at
all
but
the
slowest
rates.
Fur-
thermore,
across
all
four
modulation
frequencies,
gain
in
receptor
spike
trains
was
significantly
greater
than
that
for
omegas
at
all
depths
except
25%
(Table
1).
The
low-pass
like
responses
of
omegas
were
modeled
as
one-pole
low-pass
filters
using
a
least-squares
fit
of
the
following
equation
for
each
modulation
depth
ex-
cept
0%
(tones):
Modulation
gain
(f)
=
20
log
1
+
f2T2
+
c,
where
C
is
a
constant
shifting
the
filter
above
0
(as
the
name
implies,
measures
of
gain
may
reveal
increases
in
128
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
Modulation
Rate
(Hz)
10
8
6
25
10
8
6
45
10
8
6
65
10
8
6
85
4
4
4
4
100
2
2
2
2
500
250
0
100
200
100
10
10
10
10
8
6
6
6
6
4
4
4
4
75
2
2
2
2
0
0
500
250
11
100
200
100
10
10
10
10
8
8
1
8
6
6
6
6
4
4
4
4
50
2
2
2
2
0
0
0
0
500
250
0
100
200
100
10
10
10
10
8
8
8
3
6
6
6
6
4
4
4
4
25
2
2
2
2
0
0
0
0
500
0
250
0
100
200
100
10
10
10
10
8
8
8
6
6
6
6
4
4
4
0
2
0
2
0
4
—**
2
2
500
0
250
100
200
0
100
Time
(ms)
Fig.
7.
Peri-stimulus
time
histograms
of
the
response
of
a
receptor
to
the
20
different
modulation
stimuli.
Each
frame
shows
the
summed
response
to
five
replicates
of
the
10-cycle
stimulus.
Note
that
for
these
histograms,
binwidth
equals
0.1
times
the
modulation
period
(e.g.,
binwidth
at
25
Hz
is
0.1
times
40
ms)
which
gives
the
appearance
of
variation
in
spike
number
between
the
tones.
Nu
m
ber
o
f
Sp
ikes
(%
)
11
1-
dau
uop
vi
np
ow
modulation),
and
f
and
i
are
the
modulation
frequency
and
time
constant
of
the
filter,
respectively.
Although
a
simple
linear
regression
of
modulation
gain
verses
log
modulation
frequency
showed
slopes
slightly
below
that
expected
for
a
simple
low-pass
filter
(i.e.,
—6
dB/octave;
Table
2),
the
low-pass
filter
model
explained
a
signifi-
cant
proportion
of
the
variance
in
the
responses
of
omega
neurons
at
all
depths
except
25%.
The
mean
cutoff
frequencies
calculated
from
our
sample
of
omega
cells
using
gain
functions
is
,
-
,
24
Hz
(varying
with
modulation
depth;
Table
2),
slower
than
that
for
re-
ceptors
(T
=
15
ms,
1/T
=
67
Hz,
R
2
=
0.365,
P
=
0.013).
3.3.
Input/output
at
5
kHz
Whereas
no
difference
was
found
between
the
omega
and
receptor
I/O
slopes
measured
on
an
absolute
scale
(Fig.
10(a),
mean
ON
s
i
ope
=
0.132,
N
=
8,
mean
recep-
tor
s
i
ope
=
0.296,
N
=
4,
P
=
0.324),
the
relative
I/O
slopes
differed
significantly
(mean
ON
s
i
ope
=
0.115,
N
=
8,
mean
receptor
s
i
ope
=
0.307,
N
=
4,
P
=
0.024;
due
to
either
an
incomplete
intensity
series
or
a
statis-
tically
insignificant
I/O
slope,
two
ONs
were
not
in-
cluded
in
these
comparisons).
This
means
that
in
comparison
with
omegas,
this
sample
of
receptors
re-
quire
smaller
changes
in
intensity
(from
the
90
dB
car-
I
I
I
I
T
I
I
I
I
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
129
25
45
65
85
Modulation
Frequency
(Hz)
Fig.
8.
Average
temporal
modulation
transfer
functions
showing
modulation
threshold
as
a
function
of
stimulus
modulation
frequency.
Circles
and
squares
are
the
mean
(±SE)
modulation
index
[i.e.,
20
x
log(mod
depth)]
necessary
for
significant
(P
<
0.05)
modulation
in
the
spike
trains
at
the
four
modulation
frequencies
for
receptors
(N
=
4)
and
omega
neurons
(N
=
10),
respectively.
There
were
no
omega
preps
which
displayed
statistically
significant
modulation
at
65
Hz
(open
square)
and
only
one
prep
at
85
Hz.
Right
axis
shows
the
modulation
index
in
percent
depth.
Shaded
area
marks
the
calling
song
modulation
frequency
at
22°.
25
45
65
85
Modulation
Frequency
(Hz)
Fig.
9.
Average
modulation
gain
functions
when
calculated
across
all
modulation
depths
for
receptors
(circles)
and
omegas
(squares).
Whereas
points
above
the
dashed
line
(positive
gain)
represent
cases
in
which
stimulus
modulation
is
enhanced
in
the
spike
train,
negative
values
represent
distortion
of
the
stimulus
AM.
Overall
modulation
gain
is
significantly
greater
in
receptors
than
omegas
(Table
1).
Unlike
Fig.
8,
when
expressed
as
modulation
gain,
the
TMTF
of
receptors
could
be
fit
with
a
low-pass
filter
(T
=
15
ms,
1/T
=
67
Hz,
R
2
=
0.365,
P
=
0.013).
Shaded
area
marks
the
calling
song
modulation
frequency
at
22°.
rier
level)
to
produce
greater
relative
changes
in
spike
number,
a
result
consistent
with
the
greater
AM
sensi-
tivity
demonstrated
above
(Fig.
10(b)).
4.
Discussion
4.1.
Ultrasound
sensitivity
in
hfAN
In
a
previous
study
with
G.
rubens,
extracellular
recordings
in
the
neck
connectives
revealed
at
least
one
auditory
unit
sensitive
to
ultrasound
(Farris
et
al.,
1998).
The
results
from
the
present
study
suggest
that
this
high-frequency
sensitivity
could
be
mediated
by
hfAN,
as
the
frequency
response
in
the
neck
connec-
tives
recorded
by
Farris
et
al.
(1998)
is
similar
to
those
measured
here
in
intracellular
recordings
of
hfAN
(Fig.
2).
Candidate
homologues
to
hfAN,
units
of
the
AN2
type
occur
in
at
least
three
other
species
of
gryllids
and
are
indistinguishable
from
hfAN
in
both
anatomy
and
physiology
(Casaday
and
Hoy,
1977;
Wohlers
and
Huber,
1982).
For
example,
like
hfAN,
Int-1
in
T.
oceanicus
is
most
sensitive
to
sounds
with
frequencies
from
15
to
30
kHz
(
,
-
,
58
dB
SPL,
Moiseff
and
Hoy,
1983)
and
inhibited
by
frequencies
from
3
to
8
kHz
(65
dB
SPL
at
6
kHz,
Nolen
and
Hoy,
1986a,b).
Int-1
has
been
shown
to
mediate
negative
steering
responses
to
bat-like
ultrasound
(Moiseff
and
Hoy,
1983;
Nolen
and
Hoy,
1984).
In
G.
rubens,
pulsed
ultrasound
broadcast
simultaneously
with
a
calling
song
has
been
shown
to
reduce
that
song's
relative
attractiveness
(Farris
et
al.,
1998).
Based
solely
on
the
similarity
between
the
anatomy
and
physiology
of
Int-1
in
T
oceanicus
and
hfAN
in
G.
rubens,
our
results
suggest
that
hfAN
could
play
a
similar
role
to
that
of
Int-1
by
mediating
the
repulsive
effect
of
ultrasound
in
the
context
of
anti-predator
behavior.
4.2.
Temporal
sensitivity
in
omega
neurons
Cells
of
the
omega
class
have
been
characterized
across
several
families
of
the
ensiferan
Orthoptera
(see
Mason
et
al.,
1998).
With
respect
to
the
true
crickets
(Gryllidae),
ONs
have
been
shown
to
be
broadly
tuned
with
best
sensitivity
at
frequencies
near
that
of
the
calling
song
carrier
(Popov
et
al.,
1978;
Schildberger
et
al.,
1989).
Such
a
pronounced
sensitivity
peak
was
evident
in
ON1
and
ON2
and
suggests
a
specialization
for
the
processing
of
calling
songs.
Unlike
the
sharp
spectral
tuning
at
5
kHz,
however,
previous
studies
have
found
no
evidence
to
suggest
such
a
bias
in
the
temporal
tuning
of
omegas
for
stimuli
(i.e.,
100%
modulated)
with
pulse
rates
like
those
found
in
the
songs
of
conspecific
males
(Wohlers
and
Huber,
1982).
Rather,
omega
cells
have
been
shown
to
respond
to
a
broad
range
of
stimuli
including
heterospecific
songs
with
temporal
structures
different
from
those
in
conspecific
songs
(Pollack,
1986).
In
T.
oceanicus
for
example,
omega
neurons
are
able
to
code
pulse
rates
of
,
-
,
15-32
Hz,
encompassing
the
range
25
32
40
tz
50
-;
63
79
100
2
0
co
-
,
a
,
_
2
3
-4
-6
-8
130
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
Table
1
Modulation
depth
(%)
25
50
75
100
All
depths
P
value
receptor
gain
vs.
omega
gain
0.0920
0.0255
0.0014
0.0002
<0.0001
Modulation
gain
in
receptors
is
greater
than
that
in
omegas
at
depths
>25%
(Welch's
approximate
t;
Zar,
1999).
Table
2
Mod
depth
(%)
T
(s)
1/T
(Hz)
Filter
constant
R
2
Plow-pass
Gain
slope
"octave
(dB)
(dB)/octave
100
0.049
20.40
2.024
0.274
<0.001
—4.287
<0.001
75
0.057
17.54
4.165
0.255
<0.001
—4.397
<0.001
50
0.042
23.81
3.568
0.192
<0.005
—4.046
<0.005
25
0.02
50.0
2.165
0.092
0.057
—2.836
0.055
Columns
1-6
are
parameters
for
the
least
squares
fit
of
a
one-pole
low-pass
filter
to
the
ON
modulation
gain
functions
at
each
modulation
depth.
Columns
are
the:
modulation
depth
(%),
filter
time
constant
(T),
filter
cutoff
frequency
(Hz),
filter
constant
(C),
R
2
and
P
values
for
the
model.
Columns
7
and
8
are
the
slopes
(dB)
and
P
values
for
a
linear
regression
of
modulation
gain
vs.
octave
modulation
frequency.
found
across
the
T.
oceanicus
and
T.
commodus
songs
(Pollack,
1986;
Bentley
and
Hoy,
1972).
Pollack
(1986)
did
not
test,
however,
whether
the
amount
of
temporal
coding
was
correlated
to
the
variance
in
the
modulation
rates
found
in
the
complex
songs
of
Teleogryllus
spp.
In
voltage
traces
shown
by
Wohlers
and
Huber
(1982),
temporal
coding
of
stimulus
modulation
in
the
response
of
ON1
in
G.
bimaculatus
appears
to
be
lost
at
modu-
lation
frequencies
between
29
and
66
Hz,
frequencies
just
above
those
typical
for
the
male
calling
song
(20-29
Hz,
Doherty,
1985).
In
addition
to
changes
in
modulation
rate,
changes
in
modulation
depth
due
to
reverberations
and
overlap
with
multiple
singers
can
temporally
distort
a
calling
song
(Forrest,
1994).
Investigations
of
the
response
of
omega
cells
to
the
presentation
of
multiple
songs
have
been
conducted
in
the
context
of
examining
the
ability
of
these
cells
to
code
song
directionality
as
well
as
the
mechanisms
underlying
omega
cell
adaptation
(Sobel
and
Tank,
1994;
Pollack,
1998;
Romer
and
Krusch,
2000).
For
example,
when
presented
with
two
songs,
the
biphasic
response
of
ONs
results
in
the
'selective
atten-
tion'
of
ONs
for
the
song
with
the
greatest
relative
in-
tensity,
which
may
be
produced
either
by
the
auditory
system's
directionality
(e.g.,
,
-
,
20
dB
difference
at
90°)
or
the
power
output
of
the
songs
themselves
(Pollack,
1986;
but
see
Pollack,
1998
for
review).
Excitatory
responses
to
the
louder
song
subsequently
elicit
an
outward
hy-
perpolarizing
current
which
adapts
the
cell
and
prevents
responses
to
quieter
songs.
For
the
adapted
cell,
the
quieter
song
is
no
longer
above
threshold
as
the
I/O
curve
is
shifted
to
higher
intensities
(see
below).
While
the
directionality
of
the
auditory
system
and
omega
cell
adaptation
work
in
concert
to
selectively
code
certain
sounds
coming
from
different
directions,
the
effect
of
these
mechanisms
is
reduced
as
the
sources
move
to-
gether
(Pollack,
1986).
Thus,
probes
of
selective
atten-
tion
in
omega
neurons
using
two
songs
broadcast
from
different
directions
are
not
tests
of
temporal
sensitivity
per
se
(variance
in
coding
is
not
correlated
to
a
stimulus'
temporal
structure,
but
to
its
direction
and
thus
relative
intensity).
The
two-song
experiments
more
closely
re-
semble
tests
of
directional
biases
in
loudness
and/or
the
position
of
the
modulating
stimulus
on
the
cells
I/O
curve
(which
we
address
below
for
our
data).
Building
on
these
experiments
(e.g.,
Pollack,
1986),
we
examined
the
coding
of
temporally
distorted
stimuli
that
would
hypothetically
model
songs
emanating
from
the
same
direction.
Under
this
experimental
condition,
the
direc-
tional
mechanisms
mentioned
above
for
coding
only
one
song
in
many
should
be
eliminated
and
thus
allow
us
to
measure
the
limits
of
temporal
distortion
tolerated
by
omega
cells.
In
G.
rubens,
the
response
of
an
omega
neuron
to
SAM
stimuli
resembles
that
of
a
low-pass
filter
with
a
time
constant
(r)
of
,
-
,
42
ms
(Table
2).
The
TMTF
(Fig.
8)
shows
that
whereas
only
,
-
,
50%
AM
depth
is
required
to
elicit
a
significantly
modulated
response
to
25
Hz
AM,
a
,
-
,
65%
modulation
depth
is
necessary
at
the
faster
AM
rates
typical
of
male
calling
songs
(45
Hz;
Walker,
1962;
Bentley
and
Hoy,
1972;
Doherty
and
Callos,
1991).
These
results
suggest
that
ON1
and
ON2
might
not
play
a
role
in
the
detection
of
song
modula-
tion
structure
at
longer
distances
from
a
singing
male
where
temporal
distortion
is
greatest
(Forrest,
1994;
Simmons,
1988).
Omega
neurons
are
likely
important
to
ascending
temporal
processing,
however.
From
a
comparative
point
of
view,
the
aspects
of
the
central
auditory
cir-
cuitry
involved
in
the
processing
of
song
AM
have
been
characterized
in
a
congener
of
G.
rubens
(Schildberger,
1984).
At
least
one
role
for
ON's
is
to
lateralize
and
till mil
-
(a)
I I I I I
I
I I
I
-
(b)
.-
m'
o
U-
0
45
50
55
60
65
70
75
80
85
90
dB
(SPL)
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
131
8
0
1.0
710.75
rn
0.5
0.25
rx
Fig.
10.
Input/output
(I/O)
functions
at
5
kHz
(10
ms
duration,
1
ms
ramps).
(a)
Circles
and
squares
are
the
mean
number
of
spikes
per
stimulus
as
a
function
of
stimulus
amplitude
(dB
SPL)
for
receptor
and
omega
neurons,
respectively.
(b)
Curves
represent
the
least-squares-fit
of
the
exponential
model
to
the
normalized
I/O
responses
(re.
90
dB
SPL)
of
single
omega
and
receptor
cells
(symbols
same
as
in
(a)).
The
exponential
slopes
for
these
two
particular
cells
(omega
=
0.106,
re-
ceptor
=
0.336)
are
similar
to
the
sample
means,
which
are
significantly
different
(mean
ONo„
pe
=
0.115,
mean
receptoro
ope
=
0.307,
N
=
12,
t
=
2.654,
P
=
0.024).
Note,
the
mean
relative
response
is
not
shown
in
(b)
because
the
mean
of
sigmoidal
responses
with
different
intercepts
does
not
illustrate
the
mean
slope.
modulate
input
to
prothoracic
ascending
neurons
(AN1)
(Horseman
and
Huber,
1994),
which
are
likely
one
synapse
from
two
classes
of
brain
neurons
with
tempo-
ral
sensitivity
matched
to
behavior.
Thus,
unlike
verte-
brates
in
which
identification
of
post-synaptic
cells
is
difficult,
our
measures
of
modulation
gain
in
cricket
ON
spike
trains
allow
us
to
consider
the
signal-to-noise
ratio
impinging
on
identified
ascending
units
and
assess
whether
subsequent
stages
of
processing
are
improving,
degrading
or
following
ascending
information.
For
ex-
ample,
Schildberger
(1984)
measured
spike
synchroni-
zation
to
100%
modulated
pulsed
stimuli
in
AN1
and
two
classes
of
brain
neurons
(BNC1&2)
in
G.
bimacul-
atus.
AN1
responded
like
a
low-pass
filter
with
a
cutoff
frequency
of
,
-
,
50
Hz,
an
octave
higher
than
the
ON's
measured
here.
Assuming
a
common
response
for
ON's
in
G.
rubens
and
G.
bimaculatus,
such
modulation
rates
would
generate
signals
pre-synaptic
to
AN1
with
de-
graded
modulation
(Fig.
9).
Although
several
reasons
could
invalidate
this
comparison
between
studies
(e.g.,
the
criteria
used
to
determine
significant
sychronization
are
less
strict
for
the
pulsed
stimuli
in
Schildberger's
study),
our
use
of
modulation
gain
to
describe
ON's
response
is
important
because
it
suggests
that
AN1
must
either
amplify
the
incoming
AM
and/or
receive
input
from
faster
cells,
such
as
receptors.
4.3.
Temporal
sensitivity
and
intensity
coding
in
receptors
In
contrast
to
ON1
and
ON2,
there
was
much
greater
sensitivity
to
amplitude
modulation
in
the
small
sample
of
receptors.
Over
the
range
of
modulation
rates
pre-
sented,
on
average
<50%
modulation
depth
was
re-
quired
to
elicit
a
significantly
modulated
response
in
receptors,
which
is
comparable
to
that
found
in
the
analogous
receptors
of
locusts
(Prinz
and
Ronacher,
2002).
Note,
however,
that
because
Prinz
and
Ronacher
(2002)
use
a
linear
extrapolation
between
responses
at
different
modulation
rates
to
estimate
statistically
sig-
nificant
temporal
coding
(even
though
the
change
in
response
vs.
percent
depth
does
not
appear
linear),
their
study
may
be
methodologically
biased
to
faster
re-
sponses
than
those
shown
in
Fig.
8.
Although
we
found
no
evidence
in
receptor
responses
for
any
specialized
bandpass
tuning
to
the
AM
rates
in
calling
songs,
re-
ceptors
must
play
a
role
in
the
temporal
processing
of
song
AM,
as
all
temporal
information
available
to
the
system
must
go
through
them.
In
order
to
explore
the
difference
in
temporal
pro-
cessing
in
receptors
and
omegas,
we
examined
their
ca-
pabilities
for
coding
intensity.
Because
they
both
require
the
detection
of
changes
in
stimulus
amplitude,
intensity
discrimination
and
the
detection
of
AM
are
thought
to
be
related
(Wojtczak
and
Viemeister,
1999).
In
humans
for
example,
there
is
a
positive
correlation
between
the
just-noticeable-difference
for
intensity
discrimination
and
the
amount
of
modulation
required
for
the
detec-
tion
of
AM
(Wojtczak
and
Viemeister,
1999).
From
a
physiological
perspective,
assuming
that
stimulus
in-
tensity
corresponds
to
a
certain
point
on
a
cell's
I/O
curve
(e.g.,
the
intensity
that
elicits
a
50%
response),
steeper
I/O
functions
should
produce
greater
intensity
resolution.
By
measuring
I/O
slopes
relative
to
90
dB
(the
power
of
the
SAM
stimulus)
at
5
kHz
for
auditory
receptor
(high-cutoff
TMTF)
and
omega
(low-cutoff
TMTF)
cells,
we
compared
the
intensity
resolution
of
these
two
'fast'
and
'slow'
cell
classes,
respectively.
On
an
absolute
response
scale,
the
change
in
spike
number
with
stimulus
intensity
was
similar
between
receptors
and
omegas
(Fig.
10(a)).
On
a
relative
response
scale,
however,
the
I/O
slopes
of
receptors
are
significantly
steeper
than
those
of
omegas.
In
the
context
of
detec-
tion
of
AM,
this
means
that
smaller
AM
depths
are
132
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
necessary
to
elicit
a
change
in
receptor
spike
number
than
those
for
omegas.
Consistent
with
the
psychoa-
coustic
relationship
between
increment
and
modulation
detection
(Wojtczak
and
Viemeister,
1999),
we
consider
this
result
to
be
an
independent
confirmation
of
the
re-
sults
(i.e.,
faster
and
more
sensitive
receptors
and
slower
omegas)
collected
in
the
SAM
series.
In
summary,
the
present
study
reveals
several
im-
portant
points.
First,
from
a
comparative
point
of
view,
we
found
that
the
anatomy
and
physiology
of
identifi-
able
auditory
units
in
G.
rubens
are
similar
to
those
in
other
gryllines.
Second,
using
a
systems
analysis
ap-
proach,
we
have
shown
that
temporal
distortion
affects
the
physiological
detection
of
AM
in
the
omega
cells
and
receptors
of
G.
rubens.
This
effect
varied
with
cell
type,
however,
as
receptors
could
detect
smaller
modulation
depths
than
omegas.
It
is
important
to
note
that
our
results
suggest
nothing
about
which
AM
rates
are
at-
tractive,
only
which
rates
are
effectively
coded
by
a
sample
of
auditory
units.
Acknowledgements
Three
anonymous
reviewers
offered
many
helpful
comments
on
this
paper.
Special
thanks
to
D.
Bodnar
for
advice
on
temporal
processing
and
data
analysis.
We
thank
T.
Walker
for
collecting
animals
and
helping
to
start
our
colony.
M.
Oshinski
and
T.
Forrest
provided
help
with
physiological
techniques
and
computer
pro-
gramming.
In
addition,
T.
Forrest
provided
useful
in-
sight
into
TMTFs
and
signal
detection
theory.
Thanks
to
H.E.
Bass
at
the
National
Center
of
Physical
Acoustics
for
providing
equipment.
This
study
was
presented
in
partial
fulfillment
of
a
doctoral
degree
at
Cornell
University.
H.E.
Farris
was
supported
by
NIH
Grant
No.
RO1
DC00103
to
R.R.
Hoy.
References
Alexander,
R.D.,
1960.
Sound
communication
in
Orthoptera
and
Cicadidae.
In:
Lanyon,
W.E.,
Tavolga,
W.N.
(Eds.),
Animal
Sounds
and
Communication.
AIBS
Publication
No.
7,
Washing-
ton,
DC,
pp.
38-92.
Bentley,
D.R.,
Hoy,
R.R.,
1972.
Genetic
control
of
the
neuronal
network
generating
cricket
(Teleogryllus
oceanicus)
song
patterns.
Anim.
Behay.
20,
478-492.
Casaday,
G.B.,
Hoy,
R.R.,
1977.
Auditory
interneurons
in
the
cricket
Teleogryllus
oceanicus:
physiological
and
anatomical
properties.
J.
Comp.
Physiol.
121,
1-13.
Doherty,
J.A.,
1985.
Trade-off
phenomena
in
calling
song
recognition
and
phonotaxis
in
the
cricket,
Gryllus
bimaculatus
(Orthoptera,
Gryllidae).
J.
Comp.
Physiol.
A
156,
787-801.
Doherty,
J.A.,
Callos,
J.D.,
1991.
Acoustic
communication
in
the
trilling
field
cricket,
Gryllus
rubens
(Orthoptera:
Gryllidae).
J.
Ins.
Behay.
4,
67-82.
Doolan,
J.M.,
Pollack,
G.S.,
1985.
Phonotactic
specificity
of
the
cricket
Teleogryllus
oceanicus:
intensity-dependent
selectivity
for
temporal
parameters
of
the
stimulus.
J.
Comp.
Physiol.
A
157,
223-
233.
Epping,
W.J.,
1990.
Influence
of
adaptation
on
neural
sensitivity
to
temporal
characteristics
of
sound
in
the
dorsal
medullary
nucleus
and
torus
semicircularis
of
the
grassfrog.
Hear.
Res.
45,
1-13.
Esch,
H.,
Huber,
F.,
Wohlers,
D.W.,
1980.
Primary
auditory
neurons
in
crickets:
physiology
and
central
projections.
J.
Comp.
Physiol.
137,
27-38.
Farris,
H.E.,
Forrest,
T.G.,
Hoy,
R.R.,
1997.
The
effects
of
calling
song
spacing
and
intensity
on
the
attraction
of
flying
crickets
(Orthoptera:
Gryllidae:
Nemobiinae).
J.
Insect.
Behay.
10,
639-653.
Farris,
H.E.,
Forrest,
T.G.,
Hoy,
R.R.,
1998.
The
effect
of
ultrasound
on
the
attractiveness
of
acoustic
mating
signals.
Physiol.
Entomol.
23,
322-328.
Farris,
H.E.,
Hoy,
R.R.,
2000.
Ultrasound
sensitivity
in
the
cricket,
Eunemobius
carolinus
(Gryllidae,
Nemobiinae).
J.
Acoust.
Soc.
Am.
107,
1727-1736.
Forrest,
T.G.,
1994.
From
sender
to
receiver:
propagation
and
environmental
effects
on
acoustic
signals.
Am.
Zool.
34,
644-654.
Frisina,
R.D.,
2001.
Subcoritcal
neural
coding
mechanisms
for
auditory
temporal
processing.
Hear.
Res.
158,
1-27.
Frisina,
R.D.,
Smith,
R.L.,
Chamberlain,
S.C.,
1990.
Encoding
of
amplitude
modulation
in
the
Gerbil
cochlear
nucleus:
1.
A
hierarchy
of
enhancement.
Hear.
Res.
44,
99-122.
Givois,
V.,
Pollack,
G.S.,
2000.
Sensory
habituation
of
auditory
receptor
neurons:
implication
for
sound
localization.
J.
Exp.
Biol.
203,
2529-2537.
Gleich,
0.,
Klump,
G.M.,
1995.
Temporal
modulation
transfer
functions
in
the
European
Starling
(Sturnus
vulgaris):
II.
Responses
of
auditory-nerve
fibers.
Hear.
Res.
82,
81-92.
Gollisch,
T.,
Schutze,
H.,
Benda,
J.,
Herz,
A.V.M.,
2002.
Energy
integration
describes
sound-intensity
coding
in
an
insect
auditory
system.
J.
Neurosci.
22,
10434-10448.
Heil,
P.,
Neubauer,
H.,
2001.
Temporal
integration
of
sound
pressure
determines
thresholds
of
auditory-nerve
fibers.
J.
Neurosci.
21,
7404-7415.
Hennig,
R.M.,
1988.
Ascending
auditory
interneurons
in
the
cricket
Teleogryllus
commodus
(Walker):
comparative
physiology
and
direct
connections
with
afferents.
J.
Comp.
Physiol.
163,
135-143.
Horseman,
G.,
Huber,
F.,
1994.
Sound
localisation
in
crickets.
1.
Contralateral
inhibition
of
an
ascending
auditory
interneuron
(AN1)
in
the
cricket
Gryllus
bimaculatus.
J.
Comp.
Physiol.
A
175,
389-398.
Hutchings,
M.,
Lewis,
B.,
1981.
Response
properties
of
primary
auditory
fibers
in
the
cricket
Teleogryllus
oceanicus
(Le
Guillou).
J.
Comp.
Physiol.
143,
129-134.
Imaizumi,
K.,
Pollack,
G.S.,
1999.
Neural
coding
of
sound
frequency
by
cricket
auditory
receptors.
J.
Neurosci.
19,
1508-1516.
Kavanagh,
M.W.,
1987.
The
efficiency
of
sound
production
in
2
cricket
species,
Gryllotalpa
australis
and
Teleogryllus
commodus
(Orthop-
tera,
Grylloidea).
J.
Exp.
Biol. 130,
107-119.
Kim,
D.O.,
Sirianni,
J.G.,
Chang,
S.O.,
1990.
Responses
of
DCN-
PVCN
neurons
and
auditory
nerve
fibers
in
unanethetized
decer-
ebrate
cats
to
AM
and
pure
tones:
analysis
with
autocorrelation/
power
spectrum.
Hear.
Res.
44,
95-113.
Kohlrausch,
A.,
1993.
Comment
on
Temporal
modulation
transfer
functions
in
patients
with
cochlear
implants
[J.
Acoust.
Soc.
Am.
91,
2156-2164].
J.
Acoust.
Soc.
Am.
93,
1649-1650.
Long,
G.R.,
Cullen,
K.,
1985.
Intensity
difference
limens
at
high
frequencies.
J.
Acoust.
Soc.
Am.
78,
507-513.
Machens,
C.K.,
Stemmler,
M.B.,
Prinz,
P.,
Krahe,
R.,
Ronacher,
B.,
Herz,
A.V.M.,
2001.
Representation
of
acoustic
communication
signals
by
insect
auditory
receptors
neurons.
J.
Neurosci.
21,
3215-
3227.
Mason,
A.C.,
Forrest,
T.G.,
Hoy,
R.R.,
1998.
Hearing
in
mole
crickets
(Orthoptera:
Gryllotalpidae)
at
sonic
and
ultrasonic
frequencies.
J.
Exp.
Biol.
201,
1967-1979.
H.E.
Farris
et
al.
/
Hearing
Research
193
(2004)
121-133
133
McGill,
W.J.,
Goldberg,
J.P.,
1968.
Pure
tone
intensity
discrimination
and
energy
detection.
J.
Acoust.
Soc.
Am.
44,
576-581.
Moiseff,
A.,
Hoy,
R.R.,
1983.
Sensitivity
to
ultrasound
in
an
identified
auditory
interneuron
in
the
cricket:
a
possible
neural
link
to
phonotactic
behavior.
J.
Comp.
Physiol.
152,
155-167.
Nocke,
H.,
1971.
Biophysik
der
schallerzeugung
durch
die
vorderfiugel
der
grillen.
J.
Comp.
Physiol.
74,
272-314.
Nocke,
H.,
1972.
Physiological
aspects
of
sound
communication
in
crickets
(Gryllus
campestris
L.).
J.
Comp.
Physiol.
80,
141-162.
Nolen,
T.G.,
Hoy,
R.R.,
1984.
Initiation
of
behavior
by
single
neurons:
the
role
of
behavioral
context.
Science
226,
992-994.
Nolen,
T.G.,
Hoy,
R.R.,
1986a.
Phonotaxis
in
flying
crickets.
I.
Attraction
to
the
calling
song
and
avoidance
of
bat-like
ultrasound
are
discrete
behaviors.
J.
Comp.
Physiol.
A
159,
423-439.
Nolen,
T.G.,
Hoy,
R.R.,
1986b.
Phonotaxis
in
flying
crickets.
II.
Physiological
mechanisms
of
two-tone
suppression
of
the
high
frequency
avoidance
steering
behavior
by
the
calling
song.
J.
Comp.
Physiol.
A
159,
441-456.
Palmer,
A.R.,
1995.
Neural
signal
processing.
In:
Moore,
B.C.J.
(Ed.),
Hearing.
Academic
Press,
New
York,
pp.
75-121.
Pollack,
G.S.,
1986.
Discrimination
of
calling
song
models
by
the
cricket,
Teleogryllus
oceanicus:
the
influence
of
sound
direction
on
neural
encoding
of
the
stimulus
temporal
pattern
and
on
phono-
tactic
behavior.
J.
Comp.
Physiol.
158,
549-561.
Pollack,
G.S.,
1998.
Neural
processing
of
acoustic
signals.
In:
Hoy,
R.R.,
Popper,
A.N.,
Fay,
R.R.
(Eds.),
Comparative
Hearing:
Insects.
Springer,
New
York,
pp.
139-196.
Pollack,
G.S.,
El-Feghaly,
E.,
1993.
Calling
song
recognition
in
the
cricket
Teleogryllus
oceanicus:
comparison
of
the
effects
of
stimulus
intensity
and
sound
spectrum
on
selectivity
for
temporal
pattern.
J.
Comp.
Physiol.
A
171,
759-765.
Popov,
A.V.,
Markovich,
A.M.,
Andjan,
A.S.,
1978.
Auditory
interneurons
in
the
prothoracic
ganglion
of
the
cricket,
Gryllus
bimaculatus
deGeer.
J.
Comp.
Physiol.
126,
183-192.
Popov,
A.V.,
Shuvalov,
V.F.,
1977.
Phonotactic
behavior
of
crickets.
J.
Comp.
Physiol.
119,
111-126.
Prinz,
P.,
Ronacher,
B.,
2002.
Temporal
modulation
transfer
functions
in
auditor
receptor
fibres
of
the
locust
(Locusta
migratoria
L.).
J.
Comp.
Physiol.
A
188,
577-587.
Rees,
A.,
Palmer,
A.R.,
1989.
Neuronal
responses
to
amplitude-
modulated
and
pure
tone
stimuli
in
the
guinea
pig
inferior
colliculus,
and
their
modification
by
broadband
noise.
J.
Acoust.
Soc.
Am.
85,
1978-1994.
Romer,
H.,
1998.
The
sensory
ecology
of
acoustic
communication
in
insects.
In:
Hoy,
R.R.,
Popper,
A.N.,
Fay,
R.R.
(Eds.),
Compar-
ative
Hearing:
Insects.
Springer,
New
York,
pp.
63-96.
Romer,
H.,
Bailey,
W.J.,
1986.
Insect
hearing
in
the
field
II.
Male
spacing
behaviour
and
correlated
acoustic
cues
in
the
bushcricket
Mygalopsis
marki.
J.
Comp.
Physiol.
A
159,
627-638.
Romer,
H.,
Krusch,
M.,
2000.
A
gain-control
mechanism
for
processing
of
chorus
sounds
in
the
afferent
auditory
pathway
of
the
bushcricket
Tettigonia
virklissima
(Orthoptera;
Tettigoniidae).
J.
Comp.
Physiol.
A
186,
181-191.
Romer,
H.,
Marquart,
V.,
Hardt,
M.,
1988.
Organization
of
a
sensory
neuropile
in
the
auditory
pathway
of
two
groups
of
Orthoptera.
J.
Comp.
Neurol.
275,
201-215.
Rose,
G.J.,
Capranica,
R.R.,
1985.
Sensitivity
to
amplitude
modulated
sounds
in
the
anuran
auditory
nervous
system.
J.
Neurophysiol.
53,
446-465.
Schildberger,
K.,
1984.
Temporal
selectivity
of
identified
audi-
tory
neurons
in
the
cricket
brain.
J.
Comp.
Physiol.
A
155,
171-
185.
Schildberger,
K.,
Huber,
F.,
Wohlers,
D.W.,
1989.
Central
auditory
pathway:
neuronal
correlates
of
phonotactic
behavior.
In:
Huber,
F.,
Moore,
T.E.,
Loher,
W.
(Eds.),
Cricket
Behavior
and
Neuro-
biology.
Cornell
University
Press,
Ithaca,
NY,
pp.
423-458.
Shaw,
K.,
Herlihy,
D.P.,
2000.
Acoustic
preference
functions
and
song
variability
in
the
Hawaiian
cricket
Laupala
cerasina.
Proc.
R.
Soc.
Lond.
B
267,
577-584.
Simmons,
L.W.,
1988.
The
calling
song
of
the
field
cricket,
Gryllus
bimaculatus
(De
Geer):
constraints
on
transmission
and
its
role
in
intermale
competition
and
female
choice.
Anim.
Behay.
36,
380-
394.
Smith,
R.L.,
Brachman,
M.L.,
1980.
Dynamic
response
of
single
auditory-nerve
fibers:
some
effects
of
intensity
and
time.
In:
van
den
Brink,
G.,
Bilsen,
F.A.
(Eds.),
Psychophysical,
Physiological
and
Behavioural
Studies
in
Hearing.
Delft
University
Press,
The
Netherlands,
pp.
312-319.
Smith,
R.L.,
Brachman,
M.L.,
Frisina,
R.D.,
1985.
Sensitivity
of
auditory-nerve
fibers
to
changes
in
intensity:
a
dichotomy
between
decrements
and
increments.
J.
Acoust.
Soc.
Am.
78,
1310-1316.
Sobel,
E.C.,
Tank,
D.W.,
1994.
In
vivo
Ca
2
+
dynamics
in
a
cricket
auditory
neuron:
an
example
of
chemical
computation.
Science
263,
823-826.
Stiedl,
0.,
Stumpner,
A.,
Mbungu,
D.N.,
Atkins,
G.,
Stout,
J.F.,
1997.
Morphology
and
physiology
of
local
auditory
interneurons
in
the
prothoracic
ganglion
of
the
cricket
Acheta
domesticus.
J.
Exp.
Zool.
279,
43-53.
Tougaard,
J.,
1998.
Detection
of
short
pure-tone
stimuli
in
the
noctuid
ear:
what
are
temporal
integration
and
integration
time
all
about?
J.
Comp.
Physiol.
A
183,
563-572.
Ulagaraj,
S.M.,
Walker,
T.J.,
1975.
Response
of
flying
mole
crickets
to
three
parameters
of
synthetic
songs
broadcast
outdoors.
Nature
253,
1278-1279.
Viemeister,
N.F.,
1977.
Temporal
factors
in
audition:
a
systems
analysis
approach.
In:
Evans,
E.F.,
Wilson,
J.P.
(Eds.),
Psycho-
physics
and
Physiology
of
Hearing.
Academic
Press,
London,
pp.
419-428.
Viemeister,
N.F.,
1979.
Temporal
modulation
transfer
functions
based
upon
modulation
thresholds.
J.
Acoust.
Soc.
Am.
66,
1364-1380.
Walker,
T.J.,
1957.
Specificity
in
the
response
of
female
tree
crickets
(Orthoptera:
Gryllidae:
Oecanthinae)
to
calling
songs
of
the
males.
Ann.
Entomol.
Soc.
Am.
50,
626-636.
Walker,
T.J.,
1962.
Factors
responsible
for
intraspecific
variation
in
the
calling
songs
of
crickets.
Evolution
16,
407-428.
Walker,
T.J.,
1974.
Character
displacement
and
acoustic
insects.
Am.
Zool.
14,
1137-1150.
Wohlers,
D.W.,
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.
A
146,
161-173.
Wojtczak,
M.,
Viemeister,
N.F.,
1999.
Intensity
discrimination
and
detection
of
amplitude
modulation.
J.
Acoust.
Soc.
Am.
106,
1917-
1924.
Yost,
W.A.,
Sheft,
S.,
1997.
Temporal
modulation
transfer
functions
of
tonal
stimuli:
gated
versus
continuous
conditions.
Audit.
Neurosci.
3,
401-414.
Zar,
J.H.,
1999.
Biostatistical
Analysis.
Prentice-Hall,
Englewood
Cliffs,
NJ.