J Neurophysiol 106: 3119 –3128, 2011.
First published September 14, 2011; doi:10.1152/jn.00294.2011.
Duration tuning in the inferior colliculus of the mustached bat
Silvio Macías,1 Emanuel C. Mora,1 Julio C. Hechavarría,2 and Manfred Kössl2
1
Department of Animal and Human Biology, Faculty of Biology, Havana University, Ciudad de La Habana, Cuba;
and 2Institut für Zellbiologie und Neurowissenschaft, J. W. Goethe Universität Frankfurt, Frankfurt am Main, Germany
Submitted 5 April 2011; accepted in final form 12 September 2011
audition; constant-frequency bats; duration selectivity; echolocation
DURING ECHOLOCATION, bats must track the temporal pattern of
emitted calls and returning echoes. Bats exhibit a variety of
echolocation signals ranging from steep, frequency-modulated
(FM) signals to narrowband, quasi-constant-frequency (QCF)
and constant-frequency (CF) signals with and without additional harmonics (Kalko 1995; Schnitzler and Henson 1980).
The species-specific signal structure is determined mostly by
ecological constraints on each species’ echolocation task. Aerial hawking echolocating bats, independent of the design of the
echolocation calls, systematically increase pulse repetition rate
and decrease call duration as they switch from the search to the
final phase of prey capture (Griffin 1958; Simmons et al. 1978).
The reduction in sound duration while repetition rate is increased is used to avoid an overlap of the echo with the bat’s
own emitted signal. Therefore, the analysis of call and/or echo
durations is vital for the success of this navigation system
(Neuweiler 1990). Neurons that are selective for the duration
of sound have been described in the central auditory system of
Address for reprint requests and other correspondence: S. Macías, Dept. of
Animal and Human Biology, Faculty of Biology, Havana Univ., calle 25 No.
455 entre J e I, Vedado, CP. 10400, Ciudad de La Habana, Cuba (e-mail:
[email protected]).
www.jn.org
several vertebrates including frogs (Narins and Capranica
1980; Potter 1965), chinchillas (Chen 1998), mice (Brand et al.
2000), rats (Pérez-González et al. 2006), and bats (Casseday et
al. 1994; Feng et al. 1978; Fremouw et al. 2005; Fuzessery and
Hall 1999; Galazyuk and Feng 1997; Luo et al. 2008; Mora and
Kössl 2004). In all species studied so far, duration-tuned
neurons have a temporal tuning preference for the durations of
behaviorally relevant sounds, suggesting that the filter mechanism that produces duration tuning could supplement the auditory filters for sound frequency and intensity, to guarantee a
preferred perception of the species’ important acoustic information. In the FM bats Eptesicus fuscus, Myotis lucifugus,
Antrozous pallidus, and Molossus molossus the observed neuronal duration tuning approximates the duration of their own
echolocation calls, mostly shorter than 10 ms (Casseday et al.
1994; Feng et al. 1978; Fremouw et al. 2005; Fuzessery and
Hall 1999; Galazyuk and Feng 1997; Mora and Kössl 2004). In
the horseshoe bat, Rhinolophus pusillus, neurons from the
inferior colliculus (IC) are tuned to a range of duration of
10 – 60 ms, covering the range of pulse durations emitted
during echolocation (Luo et al. 2006, 2008).
Duration tuning can be explained under a unified model
based on the coincidence of onset and offset depolarizing
events and onset tonic inhibitory postsynaptic potential (IPSP)
(Aubie et al. 2009). For example, short-pass responses arise
when the onset depolarization is suprathreshold and has a long
latency (i.e., longer than the latency of the onset IPSP). In this
situation the anticoincidence of postsynaptic potentials results
in the neurons’ duration selectivity (Fuzessery and Hall 1999;
Narins and Capranica 1980). Band pass duration tuning could
be created in two different ways: 1) by anticoincidence of
postsynaptic potentials, when the onset excitatory postsynaptic
potential (EPSP) occurs after the onset IPSP and it produces
spikes only for stimuli with high energy (i.e., stimulus is longer
than certain duration) (Aubie et al. 2009), and 2) by coincidence of subthreshold events, e.g., when onset EPSP is then
subthreshold and it occurs after the onset IPSP. In such a
situation the offset EPSP (probably created as a rebound from
tonic inhibition) is also needed. The neuron will only respond
when both onset and offset subthreshold EPSPs are temporally
summated (Casseday et al. 1994; Covey et al. 1996; Ehrlich et
al. 1997). Anticoincidence and coincidence of postsynaptic
potentials in duration-tuned neurons is supported by in vivo
data. As an example, duration-selective units in the anuran
midbrain exhibit sustained inhibition for the duration of tone
bursts and delayed excitation (Leary et al. 2008), characteristics that are present in both the coincidence model and the
anticoincidence model, and the relationship between tonic
inhibition and delayed excitation causes the cell to track the
offset of the stimulus (Casseday et al. 1994; Covey et al. 1996;
0022-3077/11 Copyright © 2011 the American Physiological Society
3119
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
Macías S, Mora EC, Hechavarría JC, Kössl M. Duration
tuning in the inferior colliculus of the mustached bat. J Neurophysiol 106: 3119 –3128, 2011. First published September 14,
2011; doi:10.1152/jn.00294.2011.—We studied duration tuning in
neurons of the inferior colliculus (IC) of the mustached bat. Durationtuned neurons in the IC of the mustached bat fall into three main
types: short (16 of 136), band (34 of 136), and long (29 of 136) pass.
The remaining 51 neurons showed no selectivity for the duration of
sounds. The distribution of best durations was double peaked with
maxima around 3 and 17 ms, which correlate with the duration of the
short frequency-modulated (FM) and the long constant-frequency
(CF) signals emitted by Pteronotus parnellii. Since there are no
individual neurons with a double-peaked duration response profile,
both types of temporal processing seem to be well segregated in the
IC. Most short- and band-pass units with best frequency in the CF2
range responded to best durations ⬎ 9 ms (66%, 18 of 27 units).
However, there is no evidence for a bias toward longer durations as
there is for neurons tuned to the frequency range of the FM component of the third harmonic, where 83% (10 of 12 neurons) showed best
durations longer than 9 ms. In most duration-tuned neurons, response
areas as a function of stimulus duration and intensity showed either V
or U shape, with duration tuning retained across the range of sound
levels tested. Duration tuning was affected by changes in sound
pressure level in only six neurons. In all duration-tuned neurons,
latencies measured at the best duration were longer than best durations, suggesting that behavioral decisions based on analysis of the
duration of the pulses would not be expected to be complete until well
after the stimulus has occurred.
3120
DURATION TUNING IN IC OF MUSTACHED BAT
Pteronotus parnellii. The mustached P. parnellii emits biosonar signals with at least four harmonics, each containing a long
CF component followed by a brief frequency modulation
(Schnitzler and Henson 1980). Duration of calls is usually
longer than 25 ms, and the CF component is ⬃20 ms (Macías
et al. 2006; Schnitzler and Henson 1980). We studied the
encoding of sound duration in IC neurons of the mustached P.
parnellii by presenting pure tones. Consistent with results from
other bat species, best durations were in the range of durations
of the species’ echolocation calls and response latencies were
always longer than best durations.
MATERIALS AND METHODS
Animals. The study was conducted on the IC of nine adult P.
parnellii (4 males and 5 females). The animals were captured at the
entrance of their diurnal refuge (a cave located 30 km southeast of
Havana, Cuba) during their evening exodus and kept in captivity in a
room with temperature, humidity, and photoperiod conditions similar
to those of the bats’ natural environments. All experiments were
performed in the Laboratory of Neurobiology and Biological Sensors
of the Institute of Cell Biology and Neuroscience at the University of
Fig. 1. Representative examples of 3 types of all-pass neurons in the inferior colliculus (IC) of the mustached bat. Left: the neuron’s discharge patterns shown
in a dot-raster display. Center: the neuron’s duration-tuning curves (circles) and response latency (triangles) for 2 levels. Right: duration response areas in
dependence of sound pressure level [sound duration is on the x-axis, sound level is on the y-axis, and spike count is indicated by the grayscale varying from dark
(lowest) to white (highest)]. The frequency and sound pressure level of the stimuli used for obtaining the dot-raster display are given at top right of the dot-raster
display. The threshold of each neuron is given at top of the duration response areas. A: neuron with an on-off phasic response. B: neuron with an onset response.
C: neuron with a sustained response.
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
Ehrlich et al. 1997). Hence, behavioral decisions based on
analysis of the duration of the pulses would not be expected to
be complete until well after the stimulus has occurred. In the
FM bats studied so far, in which echolocation calls are usually
shorter than 10 ms, the latency of every duration-tuned neuron
(measured at its best duration) is longer than the duration
independently of the neuron’s response pattern (Casseday et al.
1994; Faure et al. 2003; Feng et al. 1978; Fremouw et al. 2005;
Fuzessery and Hall 1999; Galazyuk and Feng 1997; Mora and
Kössl 2004). A similar behavior was observed in the IC
neurons from chinchillas, mice, and rats (Brand et al. 2000;
Chen 1998; Pérez-González et al. 2006), which were tuned to
longer durations (40 –200 ms). In the IC of the long CF bat
Rhinolophus pusillus, in response to complex stimuli (i.e., a
combination of FM and CF sounds), a subset of band-pass
neurons had latencies shorter than the best duration (Luo et al.
2008). However, it is not clear whether this behavior will
persist if the neuronal response is assessed with pure tone
stimuli.
Here we studied duration selectivity to pure tone stimuli in
the IC neurons of another CF bat species, the mustached bat
DURATION TUNING IN IC OF MUSTACHED BAT
gan) and band-pass filtering between 200 Hz and 5 kHz, the neuronal
signal was digitized by a Microstar DAP board (sampling rate 33
kHz), processed by the computer program, and stored for further
analysis.
Analysis of the measured electrophysiological data was performed
with MatLab 6.5 (Mathworks) scripts custom-written by Cornelius
Abel (Inst. of Cell Biology and Neuroscience, Univ. of Frankfurt) and
J. C. Hechavarría. The recorded multiunit activity was amplitude filtered so that for the subsequent analysis we selected only those
spikes whose amplitude was at least three standard deviations above
the recording baseline. The selected spikes were sorted by using the
first three principal components of all spike waveforms (Lewicki
1998). To separate the spikes originated from different neurons, the
automatic clustering algorithm “KlustaKwik” (Harris et al. 2000;
http://klustakwik.sourceforge.net/) was fed with the first three principal components of each spike. This method has been used to
discriminate single-unit from multiunit activity in other studies (Abel
and Kössl 2009).
Once a neuronal response was found, the unit’s frequency response
area (FRA) was determined with randomly presented pure-tone bursts
(10-ms duration with 0.5-ms rise/fall time presented at a repetition
period of 300 ms) with variable frequency and level combinations
(usually 20 –95 kHz and 0 – 80 dB SPL). As a rule, five averages were
sufficient to obtain an accurate estimate of the neuronal FRA. For each
neuron we determined the best frequency (BF), which is the stimulus
frequency that evoked responses ⬎25% of the maximum spike count
at the lowest stimulus intensity (minimum threshold).
To test for duration tuning the frequency was adjusted to the
neuron’s BF, and at each level between 0 and 90 dB SPL (0 dB SPL ⫽
20 ␮Pa, 10-dB steps) the duration of the stimulus was changed,
between 1 and 31 ms, at steps of 2 ms. Both level and duration of the
stimuli were presented randomly.
For analysis of responses, poststimulus time histograms (PSTHs,
1-ms bin size) were constructed. Response latency was measured at
the onset slope of the PSTHs, when it reached 50% of the PSTH peak
value. The responses of those neurons in which the number of spikes
increased in response to increasing stimulus duration were classified
as sustained; otherwise, the responses were considered as phasic. In
the onset responses, the response latency had a constant temporal
Fig. 2. Representative examples of 2 types of long-pass neurons in the IC of the mustached bat. The same conventions as in Fig. 1 are used. A: neuron with an
on-off phasic response. B: neuron with a sustained response. The same conventions as in Fig. 1 are used.
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
Frankfurt (Frankfurt, Germany). The animal use in this study was
authorized by the Centre for the Inspection and Control of the
Environment, Ministry of Science, Technology, and Environment,
Cuba.
Surgical procedures. Bats were prepared for surgery by anesthetizing them with pentobarbital sodium (0.05 mg/g body wt) via a
subcutaneous injection in the neck. A longitudinal midline incision
was made through the skin overlying the skull, and the underlying
temporal musculature was reflected from the incision along the
midline. Wound surfaces were treated with a lidocaine solution
applied topically. A custom-made metal rod was then glued to the
skull with dental cement. We let the animals rest for 24 h before
starting the electrophysiological recordings. After recovery, during
the experiment, the awake bats were placed in a body mold made of
plastic foam. The head was tightly held by the rod fixed in a metal
holder. With the use of skull and brain surface landmarks, a small hole
(less than ⬃1-mm diameter) was made over the IC with a scalpel
blade. The hole was covered with saline solution during the experiments, and care was taken to prevent desiccation.
A microelectrode (see below) was then inserted through the hole in
the skull. The experiments were conducted inside a soundproof room
(temperature: 27–32°C) for ⬍6 h. After a recording session the
exposed skull was covered with sterile bone wax, and the animal was
returned to its individual cage. Bats could be studied for several
consecutive days. All experiments were in accordance with the Declaration of Helsinki and also with German federal regulations.
Acoustic stimulation and recording. Acoustic stimuli were delivered from a ScanSpeak Revelator R2904/7000 speaker (Avisoft Bioacoustics, Berlin, Germany) placed ⬃10 cm away from the bat’s ear.
The speaker response was flat (⫾5 dB) in the frequency range from 10
to 100 kHz, and intensity of the presented pure tone stimuli was online
adjusted in accordance with the calibration frequency-response curve
of the speaker. Stimuli were controlled by custom-made software.
Pure tone stimuli were used. For all measurements, stimuli were
presented at the contralateral ear.
Extracellular neuronal recordings were made with carbon electrodes (Carbostar 0.4 – 0.8 M⍀). The depth of the recording electrode
was controlled and adjusted by a piezo-microstepper (PM 10 –1,
Maerzhaeuser). After amplification (Differential amplifier EX1, Da-
3121
3122
DURATION TUNING IN IC OF MUSTACHED BAT
RESULTS
Duration tuning and basic response properties. Duration
tuning was studied in 136 neurons from the IC of P. parnellii
that had BFs between 24 and 93 kHz depending on penetration
depth (maximally 1,200 ␮m). Thresholds of all IC neurons
ranged from ⫺5 to 70 dB SPL, and the latencies (in response
to a 10-ms pure tone presented at the unit’s BF and 10 dB
above threshold) were between 3.2 and 31.7 ms (17.01 ⫾ 7.23
ms). Sixty-three percent of IC neurons, whether or not they
were duration selective, had no spontaneous activity. In those
neurons with spontaneous activity (determined in a 25-ms time
window prior to the stimulus presentation), it was between 3
and 6 spikes/s.
Of the 136 neurons studied, 79 (58%) showed some type of
duration sensitivity. Fifty-one (37%) were not affected by
sound duration (all pass). Among the all-pass neurons 15
showed phasic onset responses (15 of 37), nine neurons had
on-off phasic responses, and in the remaining 27 the spike
counts increased as a function of stimulus duration (Fig. 1).
Of all duration-sensitive neurons, 27% (29 of 106) showed
long-pass responses (Fig. 2). Most long-pass units (24 of 29)
had sustained responses, while the other five showed on-off
discharge patterns. Short-pass responses were found in 15% of
neurons (16 of 106) that exhibited some form of duration
tuning (Fig. 3). Most short-pass neurons (9 of 16) showed a
phasic onset response pattern, three units displayed an offset
response, and the rest displayed a phasic on-off response.
Band-pass responses were found in 32% (34 of 106) of the
Fig. 3. Representative examples of 3 types of short-pass neurons in the IC of the mustached bat. The same conventions as in Fig. 1 are used. A: neuron with
an onset response. B: neuron with an on-off phasic response. C: neuron with an offset response.
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
relation to the beginning of the sound stimulus. In the offset responses, the response latency increased with the increment of the
stimuli duration.
To determine whether a neuron was sensitive to sound duration, we
plotted the spike count as a function of duration. We distinguished
three types of duration sensitivity (short pass, band pass, and long
pass) and a type of response that was not tuned to stimulus duration
(all pass).
A neuron was considered to be short pass if the spike count
dropped below 50% of maximum response at durations longer than
best duration. In the band-pass responses the spike count dropped
below 50% of maximum response at durations both longer and shorter
than best duration. Long-pass neurons did not show a best duration
and responded only when the duration of a best excitatory frequency
stimulus exceeded some minimum duration, with little or no spiking
in response to shorter-duration signals. In long-pass units, the cell’s
first-spike latency and minimum duration necessary to elicit spiking
do not continue to decrease as signal energy (amplitude) increases
(Brand et al. 2000; Faure et al. 2003; Pérez-González et al. 2006).
DURATION TUNING IN IC OF MUSTACHED BAT
response at 10 –50 dB SPL and again to a band-pass response
with a shorter best duration at 60 – 80 dB SPL. Also, this
neuron changed its response pattern from onset to offset in the
level range of 50 – 80 dB SPL. The neuron in Fig. 6B acted as
a band-pass filter at 40 –50 dB SPL and changed to a short-pass
response at 60 –70 dB SPL and finally to an all-pass response
at 80 dB SPL. This neuron showed no change in its temporal
response pattern with increasing SPL. In the example displayed
in Fig. 6C, the all-pass response at 40 – 60 dB SPL, with an
onset pattern, changed to an on-off short-pass filter at 70 – 80
dB SPL. In the three neurons shown, the latency decreases with
increase in the stimulus level.
The latency of short- and band-pass neurons remained relatively stable with increasing stimulus amplitude (Fig. 7), with
exception of the six neurons that were not tolerant to changes
in sound level. There was no significant difference between the
mean latency at 10 dB above threshold and at higher sound
levels (paired t-test, all P ⬍ 0.01). In none of the neurons
studied did we observe paradoxical latency shifts in form of an
increase in latency with increasing stimulus amplitude as has
been described in cortical (Sullivan 1982a, 1982b) and IC
(Galazyuk and Feng 2001; Galazyuk et al. 2005; Ma and Suga
2008) neurons of FM bats. In all short-pass and band-pass
Fig. 4. Representative examples of 3 types of band-pass neurons in the IC of the mustached bat. The same conventions as in Fig. 1 are used. A: neuron with an
offset response. B and C: neurons with onset responses.
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
units (Fig. 4). Neurons with band-pass responses showed offset
(9) and onset (25) discharge patterns.
The distribution and range of best durations in the population of short- and band-pass neurons studied are shown in
Fig. 5A. The distribution of best durations is bimodal, with one
peak at 3 ms and a second peak at 17 ms. In the IC of P.
parnellii, eight units had best durations longer than 20 ms.
Figure 5B plots best duration as a function of BF. Neurons with
duration-tuned responses process mainly the frequency range
of the second and third harmonics. Most duration-tuned units
that had their BF in the CF2 range (59 – 62 kHz) responded to
best durations ⬎ 9 ms (66%, 18 of 27 units). However, there is
no evidence for a bias toward longer durations as there is for
neurons tuned to the frequency range of the FM component of
the third harmonic, where 83% (10 of 12 neurons) showed best
durations longer than 9 ms.
Effect of sound level on duration tuning. Most durationtuned neurons kept the same type of duration tuning across all
SPLs tested. In six neurons, changes in SPL produced pronounced changes in the type of filter characteristics (Fig. 6),
with a tendency for a decrease of best duration with increasing
SPL, and these changes were accompanied by changes in
discharge pattern. For example, in the neuron shown in Fig. 6A,
the long-pass response at 0 dB SPL changed to a band-pass
3123
3124
DURATION TUNING IN IC OF MUSTACHED BAT
duration-tuned neurons the response latency was longer than
the best duration (Fig. 8).
DISCUSSION
In the midbrain of the FM bats E. fuscus, A. pallidus, and M.
molossus between 36% and 53% of all auditory neurons exhibit
duration tuning, and most of them are tuned to short sound
durations ⬍10 ms (Casseday et al. 1994; Ehrlich et al. 1997;
Fuzessery and Hall 1999; Mora and Kössl 2004). In the long
CF bat R. pusillus, approximately two-thirds of all durationtuned neurons were either short pass or band pass (Luo et al.
2008). In the IC of the mustached bat, 47% (50 of 106) were
classified as short pass or band pass; these ratios are similar to
those described for other bats. In contrast to other studies (Luo
et al. 2008; Mora and Kössl 2004), we did not find neurons that
showed multipeaked or band-reject duration selectivity.
Mechanisms underlying duration tuning. The coincidence
mechanism requires offset excitation or rebound from inhibition in duration-tuned neurons. The high number of durationtuned neurons with offset responses in the IC of E. fuscus
(⬎50%), A. pallidus (42%), and M. molossus (39%) supports
this fact. In the mustached bat IC, only 12 neurons were offset
responders (3 short-pass and 9 band-pass neurons). In those
neurons with offset responses, the coincidence model could
easily explain duration tuning (Casseday et al. 1994; Covey et
al. 1996; Ehrlich et al. 1997).
The temporal selectivity of short-pass neurons with an onset
response pattern can be explained by the anticoincidence of
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
Fig. 5. A: distribution of best durations of short-pass (open bars) and band-pass
(filled bars) neurons from the IC of the mustached bat Pteronotus parnellii.
The histogram was constructed using a bin size of 2 ms. B: best frequency of
each duration tuned neuron (n ⫽ 50 units). The graphic does not include the
best durations obtained with each of the intensities tested for those neurons that
changed duration tuning when the sound level was changed. Open circles,
short pass; filled circles, band pass.
postsynaptic potentials (Aubie et al. 2009; Fuzessery and Hall
1999). In the anticoincidence model, the onset excitatory input
will generate spikes until duration is such that a shorter latency
sustained inhibition input lasts long enough to coincide with
the transient excitatory response and cancels it.
Similar to those described in the IC of the long CF bat R.
pusillus (Luo et al. 2008), in which all band-pass neurons
showed onset responses, most types of duration-tuned neurons
exhibited on-responses for all durations tested. A possible
mechanism responsible for band-pass tuning in onset neurons
could be that at short stimulus duration inhibition and excitatory rebound are over before the arrival of the delayed
excitatory input, which can still cause the neuron to fire. With
increasing durations, the excitatory rebound coincides with the
excitatory input and the neuron responds maximally. As the
duration increases further, the inhibition is still ongoing while
the delayed excitatory input arrives, which reduces or even
eliminates the response (Luo et al. 2008). In the band-pass
duration-tuned neurons of the IC of P. parnellii the response
does not disappear when stimulus duration is increased, suggesting that the onset-evoked excitation is suprathreshold and
cannot be suppressed by onset inhibition. This result is in
agreement with a computational model for duration-tuned
neurons (Aubie et al. 2009; Sayegh et al. 2011).
Results of intracellular recordings from a single band-pass
duration-selective neuron in the IC of the anuran (Leary et al.
2008) suggested that band-pass selectivity could be attributable, in part, to the enhancement of excitation that occurred at
certain stimulus durations. Short-duration stimuli did not elicit
any detectable depolarization of the cell, whereas longer stimuli effectively produced excitation after a certain value of
stimulus duration. At the stimulus duration that best excited the
cell, spiking occurred slightly after the inhibition had reached
its peak. This suggests that, although no postinhibitory rebound
was observed when only inhibition was present, the peak
activity occurs during the release from inhibition and the
occurrence of the excitation, producing an onset band-pass
response (Leary et al. 2008). This result is different from that
found in intracellular recordings in the IC of E. fuscus (Covey
et al. 1996). Whole cell recording of a typical duration-tuned
cell showed that leading inhibition dominated the early part of
the response and that there was an inward current associated
with sound offset. These observations demonstrate the importance of the interaction of excitatory and inhibitory inputs that
are temporally shifted against each other in creating tuning to
simple temporal features of a stimulus (Casseday and Covey
1996).
In both models, coincidence and anticoincidence, duration
tuning to longer or shorter sounds can be created by having
longer or shorter latency in the onset excitatory input. According to what is found in other species with shorter durations of
echolocation pulses (Casseday et al. 2000; Faure et al. 2003),
in the IC of the mustached bat response latencies were always
longer than corresponding best durations of duration-tuned
neurons. The same has been found in nonecholocating mammals. For example, in the rat band-pass neurons responding
maximally to sound durations between 25 and 160 ms showed
longer latencies (Pérez-González et al. 2006). In the IC of
guinea pigs the subset of onset neurons showing band-pass
responses showed significantly longer latencies than onset
DURATION TUNING IN IC OF MUSTACHED BAT
3125
all-pass neurons (Wang et al. 2006). Also, in mice most of the
band-pass neurons had offset responses (Brand et al. 2000).
Only two studies (Luo et al. 2008; Yin et al. 2008) report
duration-tuned neurons whose response latencies were shorter
than best durations (R. pusillus: Luo et al. 2008; guinea pigs:
Yin et al. 2008). In these two studies either complex acoustic
stimuli (i.e., a combination of FM and CF sounds in the study
by Luo et al. 2008) or narrow time analysis windows were used
to assess the neuronal response. Therefore it is hard to know
whether the duration selectivity of these neurons could be
explained by coincidence or anticoincidence models.
Behavioral role of duration-tuned neurons in the CF bat
Pteronotus parnellii. The behavioral role of duration tuning is
still poorly understood. One hypothesis is that in bats neuronal
Fig. 7. Relation between response latency for stimulation
with tones at best duration and 10 dB above threshold
compared with response latency for a best duration tone at
20 (A) and 30 (B) dB above threshold in short- and bandpass neurons (n ⫽ 50 responses). Solid line, 1:1 line.
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
Fig. 6. Examples of units with variations in duration tuning properties due to intensity changes. A: the response changes from a long-pass response at 0 dB SPL
to a band-pass response between 10 and 50 dB SPL and to a short-pass response with levels above 60 dB SPL. B: a band-pass response changes to short pass
and to all pass with successive increases in SPL. C: an all-pass response changes to short pass when SPL is increased. Included are the dot-raster plots obtained
for 2 intensities in each neuron. Insets: 1 ms of the action potential waveforms (y-axis in arbitrary units) obtained after the spike sorting. Duration response areas
and dot-raster displays were calculated, in each neuron, by using the cluster with the higher number of spikes.
3126
DURATION TUNING IN IC OF MUSTACHED BAT
Fig. 8. Relation between best duration and latency for the
best duration in duration-tuned neurons at 10 (A) and 30 (B)
dB above threshold (n ⫽ 50 responses). Solid line indicates
stimulus offset.
J Neurophysiol • VOL
assess whether duration-tuned neurons will support target ranging (at least not in the form in which delay-tuned neurons do).
However, it is possible that they do help bats to keep tracking
their position relative to a specific target. For example, different populations of duration-tuned neurons could be activated
during different echolocation phases. Drastic changes in call
duration are known to occur in all insectivorous bats. In P.
parnellii the duration of the CF component is reduced from
⬃20 ms (search phase) to ⬍5 ms (final buzz). We found an
overrepresentation of long and short best durations in the IC of
P. parnellii. This could indicate that selective activation of
different subsets of duration-tuned neurons can trigger different motor behavioral programs about two possible stages: far
from target (search phase) and close to target (final buzz).
These two phases of echolocation demand very different behaviors, which could be triggered by the different neuronal
subsets. Of course, the duration-tuned neurons constitute only
one of several components of the sonar imaging process. For
creating an accurate image of the surroundings the durationtuned neurons have to operate together with other populations
of neurons (i.e., the delay-tuned neurons).
The fact that the duration-tuned neurons in P. parnellii
largely are level insensitive would contribute to this purpose.
Such behavior emphasizes the functional role assigned to
duration tuning because it indicates that the activation of a
specific neuron might mark the duration of an incoming sound,
without being affected by its intensity (Fremouw et al. 2005;
Zhou and Jen 2001). In a very small subpopulation of collicular
neurons (6 of 136), however, the intensity of the stimuli affects
duration tuning. In these neurons, a change in duration tuning
to shorter best duration at higher sound levels often was
accompanied by a decrease in the response latency, which is
different from what is found in duration-tuned neurons in
low-duty cycle bats (Fremouw et al. 2005). Of course, extracellular recordings in which the spiking pattern changes during
larger changes in sound pressure level could have been produced by activation of different neurons within a multiunit
cluster. An onset cell sensitive at low to moderate SPLs and an
offset cell responding at higher SPLs could produce the observed spike pattern changes in an extracellular multiunit
recording. However, only spikes of large amplitude (at least 3
times the amplitude of the base noise) were fed into the spike
sorting algorithm. Spikes were sorted by using a principal
component analysis (PCA). In the PCA, to represent any
particular data point (i.e., a spike) the principal components are
scaled and added together. Because the components are or-
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
tuning for stimulus duration is a mechanism that operates
during the processing of echolocation calls and/or echoes. This
hypothesis is supported by the correlation between echolocation pulse duration and neuronal best durations, although this is
not exclusive to bats, since neuronal tuning for sound duration
has been found in a number of nonecholocating species (Brand
et al. 2000; He et al. 1997; Leary et al. 2008; Pérez-González
et al. 2006; Wang et al. 2006). Other evidence suggesting that
duration tuning might be important for the processing of
echolocation calls and/or echoes is the finding that duration
tuning improves as the acoustic stimulus mimics the natural
scenarios of the species. In E. fuscus duration selectivity
improves during presentation of tone pairs at a certain delay
mimicking call and echo (Wu and Jen 2006), and frequency
tuning and amplitude coding of the duration-tuned neurons can
improve at the best duration (Wu and Jen 2007, 2008). On the
other hand, the same argument could be made for communication calls. The presentation of two stimuli could lead to
phenomena such as backward and forward masking, which
would shift thresholds upward, thereby appearing to sharpen
any sort of tuning. Faure et al. (2003), in a two-pulse experiment in the IC of E. fuscus, showed that the relative time
between best duration and nonexcitatory pulses affects the
activity of the neurons, suggesting that inhibition could explain
temporal masking phenomena.
In the species of FM bats studied, the range of best durations
reported extends beyond the duration of the echolocation
pulses and is usually shorter than 10 ms (Ehrlich et al. 1997;
Fuzessery and Hall 1999; Mora and Kössl 2004). In the long
CF bat R. pusillus, duration of echolocation calls ranged from
16 to 58 ms (Luo et al. 2006). In this species, Luo et al. (2008)
found that most of the best durations in short- and band-pass
neurons covered the range of pulse duration emitted during
echolocation.
The acoustic behavior of the mustached bat during prey
pursuits and capture has not been studied in detail. However,
during hunting sequences a progressive shortening of the
emitted pulse can be expected. In enclosed rooms, during
exploration flights, the duration of echolocation signals of P.
parnellii was around 22 ms (Macías et al. 2006). Overall, the
range of best duration encompassed the range of duration of the
echolocation calls.
Theoretically, duration-tuned neurons can also be involved
in the perception of target distance (Sayegh et al. 2011).
However, at present there is no evidence of duration-tuned
neurons being also tuned to echo delays. It is thus difficult to
DURATION TUNING IN IC OF MUSTACHED BAT
dered in terms of how much variability they capture, adding
together the first k components will describe the most variation
in the data. Lewicki (1998) and Wheeler and Heetderks (1982)
showed that in most cases the first three components (which is
the number of components used in our spike sorting procedure)
are enough for an accurate characterization of the spike waveforms. Wheeler and Heetderks (1982) also showed that PCA is
more robust than other spike sorting methods such as, for
example, the cluster analyses. Neurons in which the duration
properties are dependent on the sound level were also found to
be well represented in the IC of the FM bat M. molossus (Mora
and Kössl 2004), which exhibits a high plasticity in its echolocation calls (Mora et al. 2004). It is still an open question
whether and by which mechanism the ambiguous durationtuned neurons could contribute to the perception of sounds.
This work was supported by fellowships from the Deutscher Akademischer
Austauschdienst (DAAD) to S. Macías and by the Deutsche Forschungsgemeinschaft (DFG).
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
AUTHOR CONTRIBUTIONS
Author contributions: S.M., E.C.M., and M.K. conception and design of
research; S.M. performed experiments; S.M. and J.C.H. analyzed data; S.M.,
E.C.M., J.C.H., and M.K. interpreted results of experiments; S.M. prepared
figures; S.M. and E.C.M. drafted manuscript; S.M., E.C.M., J.C.H., and M.K.
edited and revised manuscript; S.M., E.C.M., and M.K. approved final version
of manuscript.
REFERENCES
Abel C, Kössl M. Sensitive response to low-frequency cochlear distortion
product in the auditory midbrain. J Neurophysiol 101: 1560 –1574, 2009.
Aubie B, Becker S, Faure PA. Computational models of millisecond level
duration tuning in neural circuits. J Neurosci 29: 9255–9270, 2009.
Brand A, Urban R, Grothe B. Duration tuning in the mouse auditory
midbrain. J Neurophysiol 84: 1790 –1799, 2000.
Casseday JH, Covey E. A neuroethological theory of the operation of the
inferior colliculus. Brain Behav Ecol 47: 311–336, 1996.
Casseday JH, Ehrlich D, Covey E. Neural tuning for sound duration: role of
inhibitory mechanism in the inferior colliculus. Science 264: 847– 850,
1994.
Casseday JH, Ehrlich D, Covey E. Neural measurement of sound duration:
control by excitatory-inhibitory interactions in the inferior colliculus. J
Neurophysiol 84: 1475–1487, 2000.
Chen GD. Effects of stimulus duration on responses of neurons in the
chinchilla inferior colliculus. Hear Res 122: 142–150, 1998.
Covey E, Casseday JH. Timing in the auditory system of the bat. Annu Rev
Physiol 61: 457– 476, 1999.
Covey E, Kauer JA, Casseday JH. Whole-cell patch-clamp recording reveals
subthreshold sound-evoked postsynaptic currents in the inferior colliculus of
awake bats. J Neurosci 16: 3009 –3018, 1996.
Ehrlich D, Casseday JH, Covey E. Neural tuning to sound duration in the
inferior colliculus of the big brown bat, Eptesicus fuscus. J Neurophysiol 77:
2360 –2372, 1997.
Faure PA, Fremouw T, Casseday JH, Covey E. Temporal masking reveals
properties of sound-evoked inhibition in duration-tuned neurons of the
inferior colliculus. J Neurosci 23: 3052–3065, 2003.
Feng AS, Simmons JA, Kick SA. Echo detection and target ranging neurons
in the auditory system of the bat Eptesicus fuscus. Science 202: 645– 648,
1978.
Fremouw T, Faure PA, Casseday JH, Covey E. Duration selectivity of
neurons in the inferior colliculus of the big brown bat: tolerance to changes
in sound level. J Neurophysiol 94: 1869 –1878, 2005.
J Neurophysiol • VOL
Fuzessery ZM, Hall JC. Sound duration selectivity in the pallid bat inferior
colliculus. Hear Res 137: 137–154, 1999.
Galazyuk AV, Feng AS. Encoding of sound duration by neurons in the
auditory cortex of the little brown bat, Myotis lucifugus. J Comp Physiol A
180: 301–311, 1997.
Galazyuk AV, Feng AS. Oscillation may play a role in time domain central
auditory processing. J Neurosci 21: RC147, 2001.
Galazyuk AV, Lin W, Llano D, Feng AS. Leading inhibition to neural
oscillation is important for time domain processing in the auditory midbrain.
J Neurophysiol 94: 314 –326, 2005.
Griffin DR. Listening in the Dark. New Haven, CT: Yale Univ. Press, 1958.
Harris R, Henze DA, Csicsvari J, Hirase H, Buzáki G. Accuracy of tetrode
spike separation as determined by simultaneous intracellular and extracellular measurements. J Neurophysiol 84: 401– 414, 2000.
He JF, Hashikawa T, Ojima H, Kinouchi Y. Temporal integration and
duration tuning in the dorsal zone of cat auditory cortex. J Neurosci 17:
2615–2625, 1997.
Kalko EKV. Echolocation signal design, foraging habitats and guild structure
in six Neotropical sheath-tailed bats (Emballonuridae). Symp Zool Soc Lond
67: 259 –273, 1995.
Leary CJ, Edwards CJ, Rose GJ. Midbrain auditory neurons integrate
excitation and inhibition to generate duration selectivity: an in vivo wholecell patch study in anurans. J Neurosci 28: 5481–5493, 2008.
Lewicki MS. A review of methods for spike sorting: the detection and
classification of neural action potentials. Network 9: R53–R78, 1998.
Luo F, Li AA, Ma J, Wu FJ, Liang B, Chen QC, Zhang SY. Spectrotemporal characteristics of echolocation pulses and frequency tuning by inferior
collicular neuron of Rhinolophus pusillus. Acta Zool Sin 52: 528 –535, 2006.
Luo F, Metzner W, Wu FJ, Zhang SY, Chen QC. Duration-sensitive
neurons in the inferior colliculus of horseshoe bats: adaptations for using
CF-FM echolocation pulses. J Neurophysiol 99: 284 –296, 2008.
Ma X, Suga N. Corticofugal modulation of the paradoxical latency shifts of
inferior collicular neurons. J Neurophysiol 100: 1127–1134, 2008.
Macías S, Mora EC, García A. Acoustic identification of mormoopid bats: a
survey during the evening exodus. J Mammal 87: 324 –330, 2006.
Mora EC, Macías S, Vater M, Coro F, Kössl M. Specializations for
aerial-hawking in the echolocation system of Molossus molossus (Molossidae, Chiroptera). J Comp Physiol A 190: 561–574, 2004.
Mora EC, Kössl M. Ambiguities in sound-duration selectivity by neurons in
the inferior colliculus of the bat Molossus molossus from Cuba. J Neurophysiol 91: 2215–2226, 2004.
Narins PM, Capranica RR. Neural adaptation for processing the two-note
call of the Puerto Rican treefrog, Eleuthereodactylus coqui. Brain Behav
Evol 18: 48 – 66, 1980.
Neuweiler G. Auditory adaptations for prey capture in echolocating bats.
Physiol Rev 70: 615– 641, 1990.
Pérez-González D, Malmierca MS, Moore JM, O, Covey E. Hernandez
Duration-selective neurons in the inferior colliculus of the rat: topographic
distribution and relation of duration sensitivity to other response properties.
J Neurophysiol 95: 823– 836, 2006.
Potter HD. Patterns of acoustically evoked discharges of neurons in the
mesencephalon of the bullfrog. J Neurophysiol 28: 1155–1184, 1965.
Sayegh R, Aubie B, Faure PA. Duration tuning in the auditory midbrain of
echolocating and non-echolocating vertebrates. J Comp Physiol A (February
9, 2011). doi:10.1007/s00359-011-0627-8.
Schnitzler HU, Henson OW. Performance of airborne animal sonar system.
I. Microchiroptera. In: Animal Sonar Systems, edited by Busnel RG, Fish JF.
New York: Plenum, 1980, p. 109 –181.
Simmons JA, Lavender WA, Lavender BA, Childs JE, Hulebak K, Rigden
MR, Sherman J, Woolman B. Echolocation by free-tailed bats (Tadarida).
J Comp Physiol A 125: 291–299. 1978.
Sullivan WE. Neural representation of target distance in auditory cortex of the
echolocating bat Myotis lucifugus. J Neurophysiol 48: 1011–1032, 1982a.
Sullivan WE. Possible neural mechanisms of target distance coding in auditory system of the echolocating bat, Myotis lucifugus. J Neurophysiol 48:
1033–1047, 1982b.
Wang J, van Wijhe R, Chen Z, Yin S. Is duration tuning a transient process
in the inferior colliculus of the guinea pig? Brain Res 1114: 63–74, 2006.
Wheeler BC, Heetderks WJ. A comparison of techniques for classification of
multiple neural signals. IEEE Trans Biomed Eng 29: 752–759, 1982.
Wu CH, Jen PH. GABA-mediated echo duration selectivity of inferior
collicular neurons of Eptesicus fuscus, determined with single pulses and
pulse-echo pairs. J Comp Physiol A 192: 985–1002, 2006.
106 • DECEMBER 2011 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
GRANTS
3127
3128
DURATION TUNING IN IC OF MUSTACHED BAT
Wu CH, Jen PH. Neurons in the inferior colliculus of the big brown bat show
maximal amplitude sensitivity at the best duration. Chin J Physiol 50:
258 –268, 2007.
Wu CH, Jen PH. Bat inferior collicular neurons have the greatest frequency
selectivity when determined with best-duration pulses. Neurosci Lett 438:
362–367, 2008.
Yin S, Chen Z, Yu D, Feng Y, Wang J. Local inhibition shapes duration
tuning in the inferior colliculus of guinea pigs. Hear Res 237: 32– 48,
2008.
Zhou X, Jen PH. The effect of sound intensity on duration-tuning characteristics of bat inferior collicular neurons. J Comp Physiol A 187:
63–73, 2001.
Downloaded from http://jn.physiology.org/ by 10.220.32.246 on June 17, 2017
J Neurophysiol • VOL
106 • DECEMBER 2011 •
www.jn.org