Binaural cross-correlation and auditory localization in the barn owl: a theoretical study

doi:10.1016/S0893-6080(98)00107-5

Neural Networks

Volume 12, Issue 1, January 1999, Pages 31-42

https://doi.org/10.1016/S0893-6080(98)00107-5 Get rights and content

Abstract

The barn owl is a nocturnal predator that is able to capture mice in complete darkness using only sound to localize prey. Two binaural cues are used by the barn owl to determine the spatial position of a sound source: differences in the time of arrival of sounds at the two ears for the azimuth (interaural time differences (ITDs)) and differences in their amplitude for the elevation (interaural level differences (ILDs)). Neurophysiological investigations have revealed that two different neural pathways starting from the cochlea seem to be specialized for processing ITDs and ILDs. Much evidence suggests that in the barn owl the localization of the azimuth is based on a cross-correlation-like treatment of the auditory inputs at the two ears. In particular, in the external nucleus of the inferior colliculus (ICx), where cells are activated by specific values of ITD, neural activation has been recently observed to be dependent on some measure of the level of cross-correlation between the input auditory signals. However, it has also been observed that these neurons are less sensitive to noise than predicted by direct binaural cross-correlation. The mechanisms underlying such signal-to-noise improvement are not known. In this paper, by focusing on a model of the barn owl's neural pathway to the optic tectum dedicated to the localization of the azimuth, we study the mechanisms by which the ITD tuning of ICx units is achieved. By means of analytical examinations and computer simulations, we show that strong analogies exist between the process by which the barn owl evaluates the azimuth of a sound source and the generalized cross-correlation algorithm, one of the most robust methods for the estimate of time delays.

Introduction

The capability of localizing the position of a sound source through audition is common to many animal species including humans. Accurate and fast spatial localization of a sound source is crucial for the successful capture of prey or escape from predators. In the case of audition, spatial information is not explicitly represented at the input stages, but needs to be extracted by the brain. A signature of the position of a sound source is usually present in several features of auditory signals: some of these cues are monaural, such as the spectral modifications induced by the configuration of the pinnae, others are binaural. The latter cues are based on comparisons of the two input signals, such as differences in the amplitude and in the time of arrival of sounds at the two ears (interaural level differences (ILDs) and interaural time differences (ITDs)).

In most species studied, the ongoing ITD appears to be an important cue for localizing the azimuth of the sound source (37., 24., 20.). Simple geometrical considerations reveal that, in the frontal field, ITD is linked to the azimuth of a sound source by a monotonic relationship (36), so that the source position along the azimuthal axis is determined by the measured ITD. It has long been hypothesized that ITDs could be derived in the brain by cross-correlating the signals at the two ears (22., 31., 32.). Several psychophysical experiments in humans and theoretical models are in agreement with this hypothesis (30., 23., 29.). According to this method, ITDs are estimated by measuring the time at which binaural cross-correlation reaches a maximum (see, for example, ⁷). Since the cross-correlation function is based on a global comparison of the two waveforms, and does not depend on ad hoc parameters, this is a robust method for estimating the ITD between two auditory signals originating from the same sound source.

In the last two decades, auditory localization has been carefully studied in the barn owl, a nocturnal predator that relies heavily on audition for hunting (for a review, see ²⁰). Behavioral studies have shown that the barn owl uses ongoing ITDs and ILDs almost independently for estimating the azimuth and the elevation of the sound source (25).

Neurophysiological and anatomical studies have shown that two separate neural pathways are specialized for analyzing ITDs and ILDs (21). These pathways start from the cochlea and converge at the level of the central nucleus of the inferior colliculus (ICc). Neurons in the external nucleus of the inferior colliculus (ICx) are activated by auditory stimuli originating from sources located in specific positions of the surrounding space. It has been observed that the spatial selectivity of ICx neurons is the result of a selectivity for specific values of ITD and ILD which unequivocally determine the location of their receptive fields (25., 14.). The locations of receptive fields vary systematically with the position of cells within the ICx, so as to give rise to an auditory map of space (16., 19.). Several experimental observations suggest that the ITD tuning of ICx cells emerges as a result of a cross-correlation-like treatment of the signals at the two ears. In particular, it has recently been shown that space-specific neurons within the ICx are sensitive to some measure of the level of correlation of the input signals (1) and that binaural cross-correlation predicts the response of ICx neurons under conditions simulating summing localization (10). However, it has also been observed that space-specific ICx neurons are less affected by noise than would be predicted by direct binaural cross-correlation (10). Although intrinsic connections in the inferior colliculus are believed to play an important role in this final result, the exact mechanisms by which such a noise rejection is achieved are not known.

In this paper, we analyze from a theoretical perspective the origins of the azimuthal map of space in the ICx. By means of analytic examinations, as well as computer simulations, we show that the processes by which the barn owl determines the azimuth of a sound source bear a strong resemblance to the so-called generalized cross-correlation method (12), a more robust algorithm than direct binaural cross-correlation. Generalized cross-correlation is based on a prefiltering of the input signals before the cross-correlation, so as to compensate for the errors introduced by a finite time window of observation, and by the presence of multiple sound sources or echos. By focusing on a model of the neural pathway dedicated to ITDs, we show that several elements cooperate to improve the signal-to-noise ratio in the activation of ICx units. In particular, we illustrate that the characteristic response to different ITDs of units in the nucleus laminaris together with the intrinsic connections in the inferior colliculus operate a peculiar frequency weighting of the input power spectral densities, that improve the reliability of ITD estimates in noisy environments. This operation is similar to the one performed by several versions of the generalized cross-correlation algorithm.

In Section 2 we briefly review the generalized cross-correlation method for the estimate of time delays. In Section 3 we describe a mathematical and computational model of the tectal pathway dedicated to azimuth localization in the barn owl. The results of the analysis of this model are presented in Section 4. Finally, a brief discussion is included in Section 5.

Section snippets

The generalized cross-correlation method

Let s_L(t) and s_R(t) be the auditory signals at the left and right ears respectively. In the case of a single remote sound source, s(t), a common way to model these signals is $s_{L} (t)=s(t)+n_{L} (t)$ $s_{R} (t)=ks(t+ Δ)+n_{R} (t)$ where Δ is the time delay which we want to evaluate, k is a scalar, and n_L(t) and n_R(t) are noise contributions.

A well-known method for estimating Δ is to search the maximum of the cross-correlation function c(t) between the two input signals: $max T c(t)= max T (s_{L} (t)×s_{R} (t))$ where the

Modeling the pathway for auditory azimuth localization

The overall organization of the tectal pathways for azimuth localization in the barn owl is schematically illustrated in Fig. 2. The auditory pathway starts with the magnocellular cochlear nuclei, and meets in the optic tectum the visual pathway composed of direct retino-tectal projections. In the tectum, visual and auditory maps of space are aligned with a motor map, which participates in the production of orienting behavior (15). As shown by electrophysiological recordings, an auditory map of

Model analysis

In order to investigate the contributions of different components of the system to the final activity of ICx units, we have analyzed the model in two distinct steps. First, to study the effect of the feed-forward connectivity we simplified the network by removing the intrinsic connections in the inferior colliculus. When this simplification is made an analytic examination of the activation of ICx units is possible. The results of this analysis are shown in the first part of this section. In the

Conclusions

A large amount of experimental evidence suggests that the barn owl localizes the azimuth of a sound source by means of a cross-correlation-like treatment of the signals at the two ears. Space-specific neurons in the inferior colliculus have been reported to be sensitive to the level of binaural cross-correlation, so that the cross-correlation at different time-lags appears to be represented by the activation of units tuned for different ITDs in the ICx map. Two main components seem to be

Acknowledgements

This work was carried out as part of the theoretical neurobiology program at The Neurosciences Institute, which is supported by Neurosciences Research Foundation. Support for this program is received in part from the van Ameringen Foundation and Novartis Pharmaceutical Corporation.

References (37)

M. Konishi
Centrally synthesized maps of sensory space
Trends Neurosci.
(1986)
Y. Albeck et al.
Responses of neurons in the auditory pathway of the barn owl to partially correlated binaural signals
J. Neurophysiol.
(1995)
Bendat, J.S., & Piersol, A.G. (1993). Engineering applications of correlation and spectral analysis. New York:...
Blauert, J. (1983). Spatial hearing. Cambridge, MA: MIT...
Bracewell, R.N. (1986). The Fourier Transform and its applications. New York:...
C.E. Carr et al.
Organization of the nucleus magnocellularis and the nucleus laminaris in the barn owl: encoding and measuring interaural time differences
J. Comparative Neurol.
(1993)
C.E. Carr et al.
A circuit for detection of interaural time differences in the brain stem of the barn owl
J. Neurosci.
(1990)
C.G. Carter
Special issue on time delay estimation
IEEE Trans. Acoustic Speech Signal Processing
(1981)
I. Fujita et al.
The role of GABAergic inhibition in processing of interaural time difference in the owl's auditory system
J. Neurosci.
(1991)
L.A. Jeffres
A place theory of sound localization
J. Comparative Physiol. Psychol.
(1948)

C.H. Keller et al.

Binaural cross-correlation predicts the responses of neurons in the owl's auditory space map under conditions simulating summing localization

J. Neurosci.

(1996)

C.H. Keller et al.

Responses to simulated echos by neurons in the barn owl's auditory space map

J. Comparative Physiol.

(1996)

C.H. Knapp et al.

The generalized correlation method for estimation of time delay

IEEE Trans. Acoustics Speech Signal Processing

(1976)

E.I. Knudsen

Subdivision of the inferior colliculus in the barn owl (Tyto alba)

J. Comparative Neurol.

(1983)

Knudsen, E.I. (1984). Synthesis of a neural map of auditory space in the owl. In: Edelman, G.M., Cowan, W.M., & Gall,...

E.I. Knudsen et al.

Creating a unified representation of visual and auditory space in the brain

Annu. Rev. Neurosci.

(1995)

E.I. Knudsen et al.

A neural map of auditory space in the owl

Science

(1978)

E.I. Knudsen et al.

Space and frequency are represented separately in auditory midbrain of the owl

J. Neurophysiol.

(1978)

Cited by (14)

Sound source localization using window function filtering and weighted cumulative histogram method
2012, IFAC Proceedings Volumes (IFAC-PapersOnline)
The importance of security cameras has been emphasised because of the effect of reducing crime rate. This paper describes a new approach to a sound source localization system and a new method for security cameras with an array of microphones in the outdoor environments. This study details the robust method to reduce the influence of environmental noises in the sound source localization using an array of microphones, the new approaches to enhance the time delay of arrival by filtering the general cross correlation's coefficient with accuracy and to calculate cumulative angles weighted by a general window function.
Development of auditory-evoked reflexes: Visuo-acoustic cues integration in a binocular head
2002, Robotics and Autonomous Systems
Citation Excerpt :
Moreover, ICx neurons showed to be sensitive to the delay between the sound signals: the final response of the neurons in this region seems to reflect a cross-correlation-like processing, albeit with a high immunity to noise and echoes [17]. It has been argued that the generalized cross-correlation [18] by compensating for noise and echoes can provide a response similar to that of the neural pathway of the barn owl [33]. Given the ITD and the ILD there are still two separate problems to be solved: (i) how to integrate visual and acoustic spatial information; (ii) how to convert the spatial percept in an appropriate motor command.
The goal of this paper is to propose a biologically plausible, functional model of the acquisition of visual, acoustic and multi-modal motor responses. Within this context visual and acoustic spatial cues are considered, fused in a coherent percept and eventually employed to control the orienting behavior of a humanoid robot. The rationale of the approach lies in the possibility to test and empirically prove the correctness of the model through the embodiment and the real interaction of the system with the environment.
The model takes into account the fact that (i) acoustic and visual cues are represented with respect to different coordinate frames (head-centric versus retino-centric) and consequently they need to be “aligned” in order to be properly fused, (ii) a teaching signal has to be generated in order to inform the system that the motor performance is not adequate to perform the task (i.e. orient toward the stimulus) and thus adaptation is required, and (iii) vision plays a major role in driving the acquisition of the appropriate map of space but other sources of feedback might be employed as well.
Robust localization of auditory and visual targets in a robotic barn owl
2000, Robotics and Autonomous Systems
Citation Excerpt :
As shown by the data in Table 1, after having been exposed to sufficient sensorimotor experience, the system was able to localize auditory targets with high accuracy. The origin of such precision in the determination of the position of a sound source, or equivalently the development of narrow tuning by ICx units to specific values of ITD, was the subject of a second study [24]. It has long been hypothesized that ITDs could be derived in the brain of the barn owl by cross-correlating the signals at the two ears.
In the last two decades, the barn owl, a nocturnal predator with accurate visual and auditory capabilities, has become a common experimental system for neuroscientists investigating the biological substrate of spatial localization and orienting behavior. As a result, much data are now available regarding the anatomy and physiology of many neural structures involved in such processes. On the basis of this growing body of knowledge, we have recently built a computer model that incorporates detailed replicas of several important neural structures participating in the production of orienting behavior. In order to expose this model to sensorimotor and environmental conditions similar to those experienced by a barn owl, the computer simulations of the neural structures were coupled to a robot emulating the head of a barn owl, which was presented with auditory and visual stimulation. By using this system we have performed a number of studies on the mechanisms underlying the barn owl’s calibration of orienting behavior and accurate localization of auditory targets in noisy environments. In this paper we review the main results that have emerged from this line of research. This work provides a concrete example of how, by coupling computer simulations of brain structures with robotic systems, it is possible to gain a better understanding of the basic principles of biological systems while producing robust and flexible control of robots operating in the real world.
An application of passive human-robot interaction: Human tracking based on attention distraction
2002, IEEE Transactions on Systems, Man, and Cybernetics Part A:Systems and Humans.
Improved ITD estimation in reverberant environments
2011, Proceedings of the AES International Conference
Multiplicative auditory spatial receptive fields created by a hierarchy of population codes
2009, PLoS ONE

View all citing articles on Scopus

View full text

Contributed ArticleBinaural cross-correlation and auditory localization in the barn owl: a theoretical study

Abstract

Introduction

Section snippets

The generalized cross-correlation method

Modeling the pathway for auditory azimuth localization

Model analysis

Conclusions

Acknowledgements

Trends Neurosci.

Responses of neurons in the auditory pathway of the barn owl to partially correlated binaural signals

J. Neurophysiol.

Organization of the nucleus magnocellularis and the nucleus laminaris in the barn owl: encoding and measuring interaural time differences

J. Comparative Neurol.

A circuit for detection of interaural time differences in the brain stem of the barn owl

J. Neurosci.

Special issue on time delay estimation

IEEE Trans. Acoustic Speech Signal Processing

The role of GABAergic inhibition in processing of interaural time difference in the owl's auditory system

J. Neurosci.

A place theory of sound localization

J. Comparative Physiol. Psychol.

Binaural cross-correlation predicts the responses of neurons in the owl's auditory space map under conditions simulating summing localization

J. Neurosci.

Responses to simulated echos by neurons in the barn owl's auditory space map

J. Comparative Physiol.

The generalized correlation method for estimation of time delay

IEEE Trans. Acoustics Speech Signal Processing

Subdivision of the inferior colliculus in the barn owl (Tyto alba)

J. Comparative Neurol.

Creating a unified representation of visual and auditory space in the brain

Annu. Rev. Neurosci.

A neural map of auditory space in the owl

Science

Space and frequency are represented separately in auditory midbrain of the owl

J. Neurophysiol.

Contributed Article
Binaural cross-correlation and auditory localization in the barn owl: a theoretical study