A Dominance Hierarchy of Auditory Spatial Cues in Barn
Ilana B. Witten, Phyllis F. Knudsen, Eric I. Knudsen*
Neurobiology Department, Stanford University Medical Center, Stanford, California, United States of America
Background: Barn owls integrate spatial information across frequency channels to localize sounds in space.
Methodology/Principal Findings: We presented barn owls with synchronous sounds that contained different bands of
frequencies (3–5 kHz and 7–9 kHz) from different locations in space. When the owls were confronted with the conflicting
localization cues from two synchronous sounds of equal level, their orienting responses were dominated by one of the
sounds: they oriented toward the location of the low frequency sound when the sources were separated in azimuth; in
contrast, they oriented toward the location of the high frequency sound when the sources were separated in elevation. We
identified neural correlates of this behavioral effect in the optic tectum (OT, superior colliculus in mammals), which contains
a map of auditory space and is involved in generating orienting movements to sounds. We found that low frequency cues
dominate the representation of sound azimuth in the OT space map, whereas high frequency cues dominate the
representation of sound elevation.
Conclusions/Significance: We argue that the dominance hierarchy of localization cues reflects several factors: 1) the relative
amplitude of the sound providing the cue, 2) the resolution with which the auditory system measures the value of a cue,
and 3) the spatial ambiguity in interpreting the cue. These same factors may contribute to the relative weighting of sound
localization cues in other species, including humans.
Citation: Witten IB, Knudsen PF, Knudsen EI (2010) A Dominance Hierarchy of Auditory Spatial Cues in Barn Owls. PLoS ONE 5(4): e10396. doi:10.1371/
Editor: Andrew Iwaniuk, University of Lethbridge, Canada
Received October 25, 2009; Accepted March 22, 2010; Published April 28, 2010
Copyright: ? 2010 Witten et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the National Institutes of Health (NIH) and National Science Foundation (NSF) graduate research fellowship. The funders
had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
The central auditory system infers the location of a sound
source in space by evaluating and combining a variety of cues. The
dominant localization cues are binaural cues, based on interaural
level differences (ILD) and interaural timing differences (ITD), the
latter based on measurements of interaural phase differences (IPD)
. Because the correspondence between values of ITD and ILD
and locations in space varies with the frequency of the sound, the
auditory system measures these cues in frequency-specific channels
and evaluates them in a frequency-specific manner. The
information provided by these cues is combined to create a
representation of the most likely location of the acoustic stimulus.
Human psychophysical studies, in which localization cues from
different frequencies are put into conflict, demonstrate that these
frequency-specific sound localization cues are weighted differen-
tially in determining the location of a sound source. For example,
when humans are presented with simultaneous low frequency
(500 Hz, 1 kHz or 2 kHz) and high frequency (4 kHz) sounds
from different locations, the high frequency sound is grouped
perceptually with the low frequency sound (because the sounds are
synchronized ), and the combined stimulus is lateralized near
the position of the low frequency source . In other experiments,
low frequency sounds have been shown to alter the lateralization
of synchronous high frequency sounds, but not vice versa [4,5].
These results indicate that the human auditory system follows
the rule that low frequency localization cues dominate over high
frequency cues when localizing a sound source. The basis for the
dominance of low over high frequency cues is thought to be
related to the relative spatial resolution provided by each cue. The
discriminability index (d’), measured psychophysically as the ability
to judge whether a sound originates from the left or right of the
midline, predicts the relative dominance of localization cues when
different frequency components containing conflicting cues are
presented simultaneously [5,6].
We looked for evidence of an analogous dominance hierarchy
among sound localization cues in barn owls, and we explored
underlying factors that could account for their relative dominance.
Owls exploit the same binaural cues for localizing sounds as do
humans. However, the human auditory system is only able to
measure IPDs for frequencies up to about 1.3 kHz [7,8] whereas,
the barn owl auditory system measures IPD cues up to about
8 kHz . In addition, the barn owl’s external ears are
asymmetrical which causes the left ear to be more sensitive to
high frequency sounds (.3 kHz) from below and the right ear to
be more sensitive to high frequency sounds from above . The
ear asymmetry causes the ILDs of frequencies above 3 kHz to vary
with the elevation of a sound source. Thus, each frequency above
3 kHz provides two binaural cues: an IPD cue that varies with
azimuth and an ILD cue that varies primarily with elevation .
PLoS ONE | www.plosone.org1 April 2010 | Volume 5 | Issue 4 | e10396
Hence, for barn owls it is not obvious how sounds will be
integrated when low and high frequency cues conflict.
Using an approach similar to that of previous human
psychophysical studies [3,5], we tested the relative dominance of
sound localization cues by presenting owls with simultaneous
sounds from different locations. The human auditory system uses
temporal coincidence as a strong cue to signify that sound
components arise from a single object . By presenting owls
with synchronous sounds of different frequencies from different
locations, we were able to observe the dynamic resolution of
contradictory spatial cues as the central auditory system created a
neural representation of the inferred location of the stimuli.
Materials and Methods
Adult barn owls were housed in flight aviaries. Birds were cared
for in accordance with the US National Institutes of Health Guide
for the Care and Use of Laboratory Animals. All procedures were
approved by the Stanford University Administrative Panel on
Laboratory Animal Care (APLAC).
Owls were anesthetized with 1% halothane mixed with nitrous
oxide andoxygen(45:55).A smallmetalfastenerwasattached tothe
rear of the skull and recording chambers (1 cm diameter) were
implanted over the optic tectum on both sides, based on stereotaxic
coordinates, with dental acrylic. A local analgesic (bupivicaine HCl)
was administered to all wounds following surgery.
Three owls were used for behavioral testing. During training and
testing sessions, an owl was placed on a perch in the center of a
darkened sound attenuating chamber. The chamber was equipped
with remotely-controlled movable speakers mounted on a narrow,
horizontal semicircular track of radius 92 cm. The track held two
speakers (Audax TM025F1) mounted on a bar that were separated
in space by 30u, either in azimuth (horizontal bar) or elevation
(vertical bar). Sound bursts consisted of either low (3–5 kHz) or high
(7–9 kHz) frequency narrowband noise, 250 ms in duration, with
5 ms rise and fall times. Bandpass filtering was performed digitally
using the ‘‘ellip’’ function in Matlab; stopband attenuation was
50 dB. Sound pressure levels (dBA scale), measured at the center of
the chamberwiththeowlremoved,wereequal(within61 dB)across
frequencies. Head positions were tracked using a head-mounted
monitoring device (miniBIRD 500, Ascension Technologies).
During an initial training period, an owl learned to first fixate a
zeroing light and then orient its head towards the source of a
subsequent sound, which was either the low or high frequency
sound. Sound levels were randomly interleaved across a range of
10–60 dB above behavioral threshold measured previously for
each owl. The location of the sound source was varied randomly
across the frontal 640u in azimuth and elevation. The owl was
rewarded with a piece of meat for orienting toward the sound with
a short latency head movement (,500 ms).
During subsequent test sessions, the owl was presented either
with one sound alone (as before), or else with two simultaneous
sounds (one low and one high frequency narrowband sound) from
different locations. When only one sound was presented, the owl
was rewarded only when it turned toward the sound source. When
simultaneous sounds were presented, the owl was rewarded for any
short latency head movement following the onset of the sounds
(,500 ms). When both sounds were presented, they were
separated either in azimuth or in elevation by 30u. When the
sounds were separated in azimuth, the elevation was positioned
randomly at either +20u, 0u, or 220u; when the sounds were
separated in elevation, the azimuth was positioned randomly at
either L20u, 0u, or R20u. Each sound was presented at each
relative location with equal probability, and sound levels roved
randomly from 10–60 dB above behavioral threshold. Single and
paired stimuli were randomly interleaved. The data reported in
this paper were collected during these test sessions. We collected
20–30 orientation movements for each stimulus configuration
from each owl.
To replicate dichotically the frequency-dependent timing and
level content of sounds coming from different spatial positions,
HRTFs were recorded from 7 owls, using a method similar to that
described by Keller et al. . Briefly, each owl was secured in the
center of the sound attenuating chamber using the head fastener,
and ketamine (0.1 ml/hr) and vallium (0.025 ml/hr) were admin-
istered throughout the session. Probetubes (1.5 cm long) attached to
microphones (Knowles FG-23652-P16) were inserted into the ears.
The tip of each probetube was placed 1–2 mm from the eardrum
and the probetube was attached to the edge of the ear canal with
superglue. Broadband sounds (2–11 kHz) from a free-field speaker
were presented from positions that spanned the frontal hemisphere
in 5u increments. For each speaker position, the signal from each
microphone was digitally recorded. The HRTF was calculated for
eachear and foreachlocation of the speaker by dividing the Fourier
transformoftherecorded sound waveformbytheFourier transform
of the presented sound waveform. The HRTFs were converted into
finite impulse response (FIR) filters, or head-related impulse
responses (HRIRs) with a linear-phase FIR filter design using
least-squares error minimization . We corrected the HRIRs to
account for the filtering properties of the speaker, chamber,
probetube, microphone, and earphones (Etymotic ER-1, used for
dichotic stimulus presentation) by measuring the appropriate
transfer functions (see below), and creating inverse FIRs to cancel
out their effects. Corrected HRIRs were then used to filter sound
waveforms to simulate free field conditions. The phase angle and
amplitude from these HRTFs corresponded to the IPD and ILD as
a function of frequency.
Eleven adult barn owls were used for electrophysiological
experiments. During a recording session, an owl was suspended in
a prone position with its head stabilized using the mounted
fastener. Nitrous oxide and oxygen (45:55) were administered
continuously so that owls remained in a passive state. Sounds
bursts, consisting of low (3–5 kHz) and/or high (7–9 kHz)
frequency narrowband noise, 50 ms in duration, with 5 ms rise
and fall times, were filtered with head-related transfer functions
(HRTFs) from a typical barn owl and were presented dichotically
through earphones (Etymotic ER-1). HRTFs from different owls
are highly consistent across the frontal region of space that was
tested [11,14]. The HRTFs from this owl were chosen because the
owl was of average size and its HRTFs closely followed the
relationship between ITD and auditory azimuth of the population
average. Differences in ILD across the measured HRTFs were on
the order of a few decibels. Multiunit and single-unit responses were
isolated from the deep layers (layers 11–13) ofthe OT with insulated
tungsten microelectrodes (6–13 MV). The identification of the
tectal layers was based on distinct unit properties that have been
linked to these layers based on electrode track reconstructions .
Site selection was based on two properties: 1) robust responses to
broadband (2–10 kHz) search stimuli, and 2) neural thresholds to
Hierarchy of Auditory Cues
PLoS ONE | www.plosone.org2 April 2010 | Volume 5 | Issue 4 | e10396
13. Proakis JG, Manolakis DG (2007) Digital signal processing. Upper Saddle River,
14. Knudsen EI, Esterly SD, du Lac S (1991) Stretched and upside-down maps of
auditory space in the optic tectum of blind-reared owls: acoustic basis and
behavioral correlates. J Neurosci 11: 1727–1747.
15. DeBello WM, Knudsen EI (2004) Multiple sites of adaptive plasticity in the owl’s
auditory localization pathway. J Neurosci 24: 6853–6861.
16. Gold JI, Knudsen EI (1999) Hearing impairment induces frequency-specific
adjustments in auditory spatial tuning in the optic tectum of young owls.
J Neurophysiol 82: 2197–2209.
17. Gold JI, Knudsen EI (2000) Abnormal auditory experience induces frequency-
specific adjustments in unit tuning for binaural localization cues in the optic
tectum of juvenile owls. J Neurosci 20: 862–877.
18. Knudsen EI, Konishi M (1978) Space and frequency are represented separately
in auditory midbrain of the owl. J Neurophysiol 41: 870–884.
19. Pena JL, Viete S, Funabiki K, Saberi K, Konishi M (2001) Cochlear and neural
delays for coincidence detection in owls. J Neurosci 21: 9455–9459.
20. Fischer BJ, Pena JL, Konishi M (2007) Emergence of multiplicative auditory
responses in the midbrain of the barn owl. J Neurophysiol 98: 1181–1193.
21. Manley GA, Koeppl C, Konishi M (1988) A neural map of interaural intensity
difference in the brainstem of the barn owl Tyto alba. J Neurosci 8: 2665–2676.
22. Koppl C (1997) Phase locking to high frequencies in the auditory nerve and
cochlear nucleus magnocellularis of the barn owl, Tyto alba. J Neurosci 17:
23. Wallach H, Newman EB, Rosenzweig MR (1949) The precedence effect in
sound localization. Am J Psychol 62: 315–336.
24. Spitzer MW, Takahashi TT (2006) Sound localization by barn owls in a
simulated echoic environment. J Neurophysiol 95: 3571–3584.
25. Spitzer MW, Bala AD, Takahashi TT (2004) A neuronal correlate of the
precedence effect is associated with spatial selectivity in the barn owl’s auditory
midbrain. J Neurophysiol 92: 2051–2070.
26. Litovsky RY, Yin TC (1998) Physiological studies of the precedence effect in the
inferior colliculus of the cat. I. Correlates of psychophysics. J Neurophysiol 80:
27. Takahashi TT, Keller CH (1994) Representation of multiple sound sources in
the owl’s auditory space map. J Neurosci 14: 4780–4793.
28. Keller CH, Takahashi TT (2005) Localization and identification of concurrent
sounds in the owl’s auditory space map. J Neurosci 25: 10446–10461.
29. Middlebrooks JC, Makous JC, Green DM (1989) Directional sensitivity of
sound-pressure levels in the human ear canal. J Acoust Soc Am 86: 89–108.
Hierarchy of Auditory Cues
PLoS ONE | www.plosone.org11 April 2010 | Volume 5 | Issue 4 | e10396