| Abstract | Introduction | Methods | Results | Discussion | References |

Lateralization of Narrow Band Noise by Blind and Sighted Listeners

Helen J. Simon, Ph.D.*
Pierre L. Divenyi, Ph.D.#
Al Lotze*

*Smith-Kettlewell Eye Research Institute
2318 Fillmore St.
San Francisco, CA 94115

#Speech and Hearing Research
VA Medical Center
Martinez, CA 94553

ABSTRACT. The effects of varying Interaural Time Delay (ITD) and Interaural Intensity Difference (IID) were measured in normal-hearing Sighted and congenitally Blind subjects as a function of eleven frequencies and at sound pressure levels (SPLs) of 70 and 90 dB SPL and 25 dB sensation level (SL)1. Using an "acoustic" pointing paradigm, the subject varied the IID of a 500-Hz narrow-band (100 Hz) noise (the "pointer") to coincide with the apparent lateral position of a "target" ITD stimulus. ITDs of 0, ± 200 and ± 400 µs were obtained through total waveform delays of narrow-band noise, including envelope and fine strsucture. For both groups, the results of this experiment confirm the traditional view of binaural hearing for like stimuli: Non-zero ITDs produce little perceived lateral displacement away from 0 IID at frequencies of above 1250 Hz. To the extent that greater magnitude of lateralization for a given ITD, presentation level and center frequency can be equated with superior localization abilities, Blind listeners appear at least comparable and even somewhat better than Sighted subjects, especially when attending to signals in the periphery. The present findings suggest that Blind listeners are fully able to utilize the cues for spatial hearing and that vision is not a mandatory prerequisite for the calibration of human spatial hearing.

1. INTRODUCTION
The current work is a part of the an ongoing project designed to increase the knowledge base regarding the interaction between impaired and intact sensory modalities in localization and binaural hearing. The experiment reported here was designed to examine binaural processing, specifically, the perceived lateral position of narrow-band noise stimuli as a function of interaural time delay, center frequency and stimulus presentation level in normal-hearing sighted and congenital totally blind subjects.

The localization of sound in space, a basic function shared by all acoustically sensitive animals, is a complex perceptual process that requires the integration of multiple acoustic cues (Wightman and Kistler 1992b). The two basic steps involved are the quantification of auditory cues and the association of cue values with appropriate locations in space (localization) or, when using earphones, along the horizontal axis inside the head (lateralization.)

Localization (or lateralization) on the azimuthal or horizontal plane (HP) depends primarily on binaural difference cues: interaural intensity differences (IIDs) and interaural time differences (ITDs) (Blauert 1997). The dual mechanism or “duplex” theory of binaural localization (Rayleigh 1907) states that low frequencies (below approximately 1.5 kHz) are localized or lateralized on the basis of ITDs and high frequencies on the basis of IIDs. That lateralization on the basis of interaural timing information is possible for complex high frequency stimuli such as transients, noise, and amplitude-modulated signals has challenged the "duplex" theory (Hafter and Buell 1985; Henning 1974). It has also been shown that when wideband stimuli are produced with conflicting IID and ITD cues, listeners follow the direction of the ITD cue, as long as the stimuli include low frequencies (Wightman and Kistler 1992a). In a study from this laboratory, (Simon et al 1994) it was found that lateralization was somewhat dependent on presentation level, especially in the high frequencies. Increasing presentation level above 60 dB sound pressure level (SPL) had an effect on lateralization, increasing the slope of the IID-ITD function even at the higher frequencies.

The auditory system evaluates time and intensity differences and associates a particular cue or set of cues with the appropriate location in space. The question of how the lack of vision affects this learning process has been raised over the years and has been the subject of much research with human subjects as well as animal models.

For many years there has been discussion regarding auditory compensatory mechanisms in the blind predicated on the theory that the loss of the visual information channel results in greater emphasis on the other sensory modalities. This implies increased requirements for auditory processing (Starlinger and Niemeyer 1981). Two models for defining the role of visual experience in the development of spatial hearing in blind listeners have been advanced. (Axelrod 1959; Jones 1975). The deficit model holds that auditory space per se, does not exist, but has to be calibrated by vision (Axelrod 1959). This model assumes that other kinds of experience cannot be substituted for the visual experience in the development of spatial hearing.

The alternative model, the compensation model (Ashmead et al 1998a; Jones 1975) assumes that, although visual experience may normally play a role in the development of spatial hearing, other kinds of experiences are also important and that compensation occurs through multimodal use. Some proponents of this model predict that spatial hearing may actually be better in persons with visual disabilities since nonvisual areas of perception may become more highly developed than in sighted individuals (Ashmead et al 1998a). Both models are supported by evidence from animals (barn owl, cat, ferret) and humans.

Early studies regarding the ability of the blind to localize, a binaural task of vital significance to the blind population (eg (Muchnik et al 1991; Rice 1969; Tonning 1975) reveal inconsistencies in the results (see Ashmead et al 1998a for a comprehensive review of spatial hearing in blind listeners.) For example, Starlinger and Niemeyer (1981) and Muchnik et al (1991) found that blind subjects performed better than sighted listeners. Other studies of localization, however, found either no differences between the two groups or differences favoring the sighted groups (eg Jones 1975; Tonning 1975; Wanet and Veraart 1985).

However, methodological problems and differences between the studies such as the use of pure tone stimuli (Muchnik et al 1991), controlling for head movement or not, lack of control of the etiology of the blindness (Starlinger and Niemeyer 1981) and comparing blind children (Tonning 1975) with an unexplained normal control make comparisons between these studies difficult (cf Ashmead et al 1998a).

More recently, consistent with the compensation model, comparable and sometimes superior spatial-hearing abilities have been found in blind children (Ashmead et al 1998b) and adults, the latter group especially with monaural signals (Lessard et al 1998) or when attending to signals in the periphery (Lessard et al 1998; Röder et al 1999).

Lessard et al (1998) studied the ability of two groups of early blind subjects (totally blind subjects and blind subjects with residual peripheral vision) and two groups of sighted subjects (blindfolded and not) to localize 30 ms broad-band noise bursts in the HP. Subjects were tested under monaural and binaural conditions. In the binaural condition, the totally blind subjects were as accurate in localizing as the blindfolded and sighted control groups and better than the subjects with residual vision. In the monaural condition (one ear was blocked with a soft foam earplug and a hearing protector muff), the totally blind subjects also localized more accurately than the other groups. While the control and residual vision groups demonstrated positional biases by localizing in favor of the unobstructed ear, half of the totally blind subjects localized the sound on the appropriate side, even when it was presented on the side of the obstructed ear. Lessard et al (1998) attributed the greater accuracy in the totally blind subjects to some form of auditory compensation not available to the group with residual vision.

Röder et al (1999) recorded electrophysiologically and found that the event related potentials (ERPs) were consistent with the behavioral result discussed above: The N1 gradient was significantly steeper for the blind than the sighted subjects for the peripheral but not the center stimulus. The steeper N1 gradient was cited as evidence of sharper tuning at early spatial attention (first 100 ms) mechanisms in blind subjects, indicating compensatory enhancement of early auditory spatial selection mechanisms that may occur following visual deprivation from birth. The blind subjects also showed a posterior shift in the scalp topography of the N1 component, providing evidence for developmental reorganization. Röder et al (1999) credited the enhanced capability for sound localization in the periphery to an attentional tuning mechanism operating with a latency of 100 msec. These results were noted to be analogous to studies of the congenitally deaf by Neville and colleagues (eg. Neville and Lawson 1987) where evidence of enhanced early processing of peripheral visual events was found.

Thus, the recent results of Lessard et al (1998) and Röder et al (1999) suggest that when visual input is absent, early auditory spatial representations continue to develop. This is consistent with the compensation model and cross-modal plasticity found in cat by Rauschecker and colleagues (1995; 1983). After binocular deprivation, an increase in responsiveness to auditory and somatosensory stimulus in multimodal areas was found in the superior colliculus and the anterior ectosylvian sulcus. In addition, visually responsive units did not significantly decrease, resulting in an increased proportion of multisensory neurons (Rauschecker and Harris 1983).

Other auditory studies of blind and sighted listeners have emphasized both peripheral and central auditory functioning. Most of the early studies assessing more peripheral functions found no significant differences between blind and sighted subjects in varied tasks such as difference limen for frequency, loudness perception and hearing thresholds (Benedett and Loeb 1972; Bross and Borenstein 1982; Curtis and Winer 1969; Starlinger and Niemeyer 1981). However, the results of a variety of tests attempting to assess central auditory functioning more often than not show a superiority of functioning for the blind subjects. For example, the blind perform better than sighted subjects do in a variety of dichotic and speech discrimination tests (Muchnik et al 1991; Niemeyer and Starlinger 1981), have decreased N1 latencies (Niemeyer and Starlinger 1981), steeper response gradients (Röder et al 1999) and better gap detection (Muchnik et al 1991).

The purpose of the present study was to assess early-blind subjects' use of one of the binaural localization cues, ITD, relative to sighted subjects. Since both of these localization cues are used in natural settings in real life, the aim of the experiment was to quantitatively assess the ability to use of these cues by the two groups. The acoustic-pointing paradigm was used, which allows an assessment of the use of the binaural cues for localization independent of requiring the subject to explicitly locate the signals in space.

The experimental procedure is an adaptation of an acoustic pointing task used in experiments investigating lateralization of sinusoidally amplitude-modulated tones and narrow bands (Bernstein and Trahiotis 1985a; Bernstein and Trahiotis 1985b; Domnitz and Colburn 1977); Simon et al 1994; Trahiotis and Bernstein 1986). This pointing procedure allowed these and other investigators to measure indirectly the extent of sound image lateralization by finding the value of the IID needed to match a given ITD.

We used this acoustic pointing task to examine lateralization as a function of 11 frequencies and a wide range of overall levels, i.e. sound pressure levels of 70 and 90 dB and 25-dB sensation level (SL). 11 frequencies were employed in order to thoroughly examine the areas of the frequency domain where IID and ITD lateralization differentially dominate. Studying frequency effects and how they relate to lateralization is especially important in the context of blindness since the blind must have access to every available auditory cue. The three presentation levels, especially 25 dB SL, were chosen to allow us to compare, at a later time, the performance of normal-hearing and hearing-impaired listeners as suggested by Durlach et al (1981). This strategy facilitates the separation of reduced sensitivity versus abnormal response-process factors in binaural processing performance. Testing normal hearing and hearing-impaired listeners at the same sensation levels allows us to equate the performance of the two groups, and ensures that the stimuli are presented to all subjects at a level sufficient to allow optimum performance.

Further, this study carefully controlled for the confounding variables such as etiology of the visual impairment, hearing loss status of the subjects, and the experimental parameters. Since lateralization results have been demonstrated to generalize to localization, it is hoped that this study will help resolve the inconsistencies of previous studies.

2. METHOD
2.1 Subjects Two male and two female congenitally and totally blind subjects (no light perception) and one male and four female sighted subjects participated in this study. Audiometric evaluations were performed on all subjects using both pure tones for diagnostic purposes and narrow band signals to obtain thresholds for the 25 dB SL condition. Figure 1 (panels a and b) shows the pure tone hearing thresholds for each ear for the Sighted (panel a) and Blind (panel b) subjects. Pure-tone hearing threshold levels for all subjects were within normal limits (re: (ANSI 1996) at all tested frequencies (250-4000 Hz) except for one Blind subject at 3000 and 4000 Hz. Her data at these two frequencies were not included in the data analysis. Although the mean age of the Blind subjects (48 yr., range = 45-51) was slightly higher than that of the Sighted subjects (39 yr. range = 18-64.), the mean hearing thresholds for the two groups were within 5 dB of each other at frequencies between 250 and 1000 Hz and within 9 dB at frequencies between 2000 and 4000 Hz. Threshold differences between the two ears of both groups were no more than 5 dB. Since it has been found that the audiogram is not a good predictor of binaural abilities (e.g. Gabriel et al 1992; Koehnke et al 1992; Koehnke and Zurek 1990) we were not concerned with the small audiometric differences between the two groups. There was no evidence of middle ear pathology on either bone conduction or tympanometric tests, and reflexometry indicated no retrocochlear involvement for any of the subjects. The etiologies of the visual impairments of all our subjects was determined as retinopathy of prematurity and was not associated with any other neurological and/or sensory involvement. All of the Blind subjects were high functioning. Because they were independent travelers who use steel-tipped canes, rather than dog guides, it was reasonable to assume that they were experienced users of everyday sound localization cues. None of the subjects had prior experience in psychoacoustic experiments. All subjects underwent extensive practice prior to the experiment, in order to familiarize themselves with the experimental procedures and to start the experiment proper after their performance had reached a relatively flat portion of the learning curve. For this performance to be reached, we found that it was only necessary for the subjects to complete a series of IID adjustments at all ITDs at a single SPL and at all 11 frequencies.

Figure 1. Audiograms. The vertical axis indicates frequency; the horizontal axis indicates dB HTL (re: ANSI, 1996). Panels (a) and (b) contain data from the Sighted and Blind Groups, respectively. O indicates the response for the right ear and X for the left. Error bars indicate the standard error of the mean (SE).

2.2 Stimuli
Narrow band stimuli were chosen primarily for four reasons. First, it has been shown that the auditory system is relatively insensitive to the ITD fine structure of sinusoids at frequencies above approximately 1200 Hz (Schiano et al 1986). Second, Bernstein and Trahiotis (1985a, 1985b) and Trahiotis and Bernstein (1986) showed that SAM waveforms did not produce the same extent of laterality as narrow band signals, even at the greatest bandwidth. Third, since the three spectral components of SAM tones are dependent upon modulation frequency, some sinusoidally modulated signals may not fall within a common critical bandwidth, giving subjects at least two ambiguous cues. Fourth, a pilot study showed that subjects had much greater variability in their laterality judgments with sinusoidally modulated signals.

Using the 'quadrature' noise method (Amenta et al, 1987), 100-Hz bandwidth narrow band Gaussian pseudo-random noise stimuli at various frequencies were digitally generated off-line. The 100-Hz stimulus bandwidth was chosen to fit the low frequencies within one critical bandwidth, thereby allowing for finer resolution of stimulus frequency dependent effects and decreasing the effect of increased bandwidth on the extent of lateralization of high frequency stimuli (Trahiotis and Bernstein 1986). One bandwidth value was chosen in order to keep the time envelope fluctuations constant at different frequencies, even though the resulting bandwidths differ in critical band units.

Ten stimulus tokens were used for each center frequency. Of these, one token was randomly chosen, with replacement, by the computer each time a "target" or "pointer" stimulus was presented. The 100 ms stimuli, including 20 ms on/off raised cosine function ramps, were generated at 20 kHz with a two-channel 12 bit digital-to-analog converter (National Instruments model NB-MIO-16). Intensity was attenuated as needed in 0.3 dB steps using an in-house built digitally controlled two-channel attenuator. After conversion, the signals were filtered using a Chebyshev fourth-order anti-aliasing filter with 8 kHz as the corner frequency that yielded a 30-dB attenuation at 20 kHz.

ITDs were produced digitally through a complete waveform (envelope and fine structure) delay. Stimuli consisted of narrow band noise with center frequencies of 250, 300, 400, 500, 750, 1000, 1250, 1500, 2000, 3000 and 4000 Hz, all with 100-Hz bandwidths.

2.3 Procedure
Since our objective was to compare hearing-impaired and normal-hearing listeners at the same SPLs and SLs in subsequent experiments, stimuli were presented to all subjects at 70 dB SPL and 90 dB SPL. Stimuli were also presented at 25 dB SL to two Sighted and all four blind subjects. The 25 dB SL signals were presented at levels between 25 and 40 dB SPL. Signals were presented binaurally through ER 3-A insert receivers2 to subjects seated in an IAC sound-attenuated room.

As previously noted (eg Domnitz and Colburn 1977), subjects sometimes reported that the subjective locations of the stimuli were not completely "punctate." When this was the case, they were instructed to match the perceived centroid of the pointer to the perceived centroid of the target. To minimize learning and/or fatigue effects the target ITD was randomized between experimental runs. Each session began with a minimum of six practice trials preceding the experimental runs. Sessions lasted one and one-half hours with breaks as required.

2.4 Experimental Paradigm
The acoustic-pointing task, demonstrated to be an effective measurement tool for measuring the extent of sound image lateralization by Bernstein and Trahiotis (1985a,b), was utilized in this study. The subjects were presented with ITD "targets" (0, ± 200, ± 400 µs) at each of the test frequencies and presentation levels. The task of the subjects was to use a linear-taper potentiometer that allowed them to vary continuously the IID of a 100-Hz-wide band of noise centered at 500 Hz (the "pointer") in order to make its perceived lateral position coincide with that of the target stimulus. In other words, the IID was used to represent the perceived lateral position of the target stimulus whose position was determined by ITD alone. In order to avoid a response bias based on the absolute position of the potentiometer’s knob, a randomly chosen value of IID (less than 5 dB) was inserted into the pointer to vary the 0 IID position of the knob before each run sequence. The range of IID could be varied by as much as 40 dB (± 20 dB) if the subject needed such a large IID variation to match the larger ITDs.

Each repeating sequence of stimuli consisted of six 100-ms stimuli, three targets followed by three pointers, with 50-ms inter-stimulus-intervals. Each sequence was followed by a 400-ms pause. This stimulus sequence was repeated until the listener indicated by pressing a button on the response box that a satisfactory match had been achieved. This was considered one trial and the magnitude of the IID served as a measure of the perceived intracranial position of the target. Subjects were instructed to match the position of the pointer to the test stimulus and to ignore any differences between the target and the pointer other than position. Thus, the targets for an experimental trial consisted of one of the 11 frequencies at one ITD and one level. Each block consisted of fifteen experimental trials, three at each of the five ITDs. Center frequency and level were held constant over a block of trials. While two Sighted subjects (1 and 2) received only two blocks at each level and frequency due to their inability to complete the project, the other subjects received four or more blocks3 at each level and frequency for all 5 ITDs. The unbalanced nested ANOVA (see results) took this discrepancy in the number of blocks run per subject into account. All trials at one presentation level were consecutive. Each level took several sessions to complete, thus insuring that data for each frequency, level and ITD were gathered on more than one day.

The level of the pointer was always different from that of the target, as suggested by Bernstein and Trahiotis (1985a,b), to prevent subjects from matching the target and pointer on the basis of equal loudness. In the 70 dB and 90 dB SPL conditions, the level of the IID pointer was such that for an IID of 0 dB, the level at each ear was 3 dB lower than the level of the ITD target. In the 25-dB SL4 condition, for an IID of 0 dB, the level at each ear was 3 dB higher than the level of the ITD target.

The data obtained were values of the pointer IID set by the listener for each center frequency at each ITD5. For these stimuli it was expected that IIDs matching the position of the low-frequency narrow band signals presented at a given ITD would reflect delays within the fine structure of the stimulus. In contrast, there would be reduced lateralization of high-frequency targets due to the reduced sensitivity to ITDs above 1250 Hz (Bernstein and Trahiotis 1985a).

3. RESULTS
To analyze these results, an ANOVA using SAS (Littell et al 1991) was carried out to examine the effects of Center Frequency, Level (25 dB SL and 70 and 90 dB SPL), Group (Blind vs Sighted) and ITD (0, 200, 400 µs) with IID as the dependent variable.

The 11 center frequencies were collapsed into three bands, low (250-500Hz), mid (750-1250 Hz) and high (1500-4000 Hz). The rationale for this was based on psychophysics: ITD is the dominant lateralization cue at low frequencies (250-500 Hz), becomes less dominant from 750-1250 Hz, and is ineffective at 1500 Hz and above. We collapsed the negative and positive ITDs since there were no differences between them and we were also not interested in hemifield asymmetries. Thus, the ANOVA design consisted of 3 frequency bands X 3 levels X 3 ITDs equaling 27 conditions6. Equal SPL values were utilized bilaterally in the 25-dB SL condition since the results of a previous study (Simon et al 1994) showed that for normal-hearing subjects, a target at 0 µs ITD was matched by a pointer closer to 0 IID under the equal SPL condition than in the equal SL condition. The between subjects factor was the comparison between the Blind and Sighted subjects. The repeated-measure (within subjects) factors were the comparisons between center frequency, level and ITD7. The results of the main effects and interactions are shown in Table I.

There are significant main effects of Group, ITD, Center Frequency and Presentation Level (see Table I) and nine significant interactions. Because the main focus of this paper is the comparison of the two groups, we will focus the description and discussion of the results on Group effects and interactions involving the Group factor.

Figure 2. The interaction of ITD and Group. The vertical axis indicates the IID value inserted into the pointer; the horizontal axis indicates the ITD in µs. The parameter is Group. Error bars indicate the SE of the mean. Asterisks between two points denote the significance levels of the pairwise comparisons (* p < .05; ** p< .005; *** p <. 0005).

In figure 2, the IID (in dB) of the pointer is shown as a function of ITD (µs) for the two groups. Error bars indicate the standard error of the mean (SE). The divergence of the two groups with increasing ITD is significant (ITD* Group interaction, F (2,3447) = 8.6, p = 0.0002) and suggests clearly greater lateralization ability by the Blind subjects. The differences are significantly greater for 400 µs than for 200 µs and are confirmed by Bonferroni adjustments. The data are consistent with results of other investigators who measured IID as a function of ITD (Bernstein and Trahiotis 1985a; Bernstein and Trahiotis 1985b; Blauert 1997; Simon et al 1994; Trahiotis and Bernstein 1986). Normal-hearing subjects are expected to match 0 ITD targets by 0 IID pointers for all frequencies, and increasing ITD targets by pointers with increasing IID. The result for both groups fulfills these expectations: Mean IIDs were within +/- 1.5 dB at 0 µs. Standard errors were small for both groups, although some individual subjects in both groups showed greater variability. This is consistent with the evidence of individual differences in other lateralization studies (eg Bernstein and Trahiotis 1985a; Simon et al 1994).

Figure 3. The interaction of Center Frequency and Group. The vertical axis indicates the IID value inserted into the pointer; the horizontal axis indicates the Center Frequency. The parameter is Group. Error bars indicate the SE of the mean. Asterisks between two points denote the significance levels of the pairwise comparisons (* p < .05; ** p< .005; *** p <. 0005).

Figure 3 shows Center Frequency on the abscissa and the IID (in dB) inserted into the pointer on the ordinate, for the two groups; error bars indicate the SE. Again, it is clear that the Blind Group’s data are superior to the Sighted Group in all frequency bands, although the difference is significant only in the two lowest frequency regions (p < .0001). Consistent with the group differences, the interaction of Group*Frequency is also significant [F (2,3447) = 4.9, p=0.0077]. There were also significant differences for both groups among the three Center Frequencies: Lateralization was significantly greater at the lowest vs the mid and the mid vs the highest Center Frequency. The average IID at 750-1250 Hz for the Blind Group was not significantly different from the average IID at 250-500 Hz for the Sighted Group showing that the Blind Group lateralizes as well in the mid frequency region as the Sighted Group does at the lowest frequency region where ITD dominate cue.

Figure 4. The significant interaction of Center Frequency, Group and ITD. The axes are identical to Figure 2. Panels (a)-(c) contain data obtained at 250-500, 750-1250, and 1500-4000 Hz, respectively. The parameter is Group. Error bars indicate the SE of the mean. Asterisks between two points denote the significance levels of the pairwise comparisons (* p < .05; ** p< .005; *** p <. 0005).

In Figs. 4a-c, performance in the three frequency bands is shown in detail. Here, similar to Fig. 2 in which all frequencies were combined, the pointer IID is shown for the two groups as a function of the three ITD values tested. Although interaction of Center Frequency by Group *ITD was significant [F (4,3447) = 3.51, p = .0073], the three panels of this figure reveal that the group difference was not uniform. Figure 4a reveals no difference between the two groups except at 400 µs where the Blind Group lateralizes significantly (p < 0.0001) more than the Sighted Group. This result is even clearer in Figure 4b (mid frequencies) where the Blind Group’s lateralization is significantly greater than the Sighted Group’s at 200 and 400 µs (p < 0.01 and 0.0001, respectively). At the highest center frequency band (Figure 4c), there is no difference between the groups.

This figure also shows that the IID values used to match non-zero target ITDs are larger (or further from 0 IID) at low (500 Hz and below) and mid-frequencies (750-1250) than at high frequencies (1500 Hz and above) for the same target ITD. Bonferroni adjustments reveal significant differences between the 250-500 Hz and 750-1250 Hz bands at 400 µs. The highest frequency result is also significantly different than the other two frequency bands at non-zero target ITDs. Although not evident in the group data, the data show greater variability as the signal moves from the midline (cf Simon et al 1994).

Figure 5. The interaction of Presentation Level and Group. The vertical axis indicates the IID value inserted into the pointer; the horizontal axis indicates the Presentation Level in dB SL and SPL. The parameter is Group. Error bars indicate the SE of the mean. Asterisks between two points denote the significance levels of the pairwise comparisons (* p < .05; ** p< .005; *** p <. 0005).

Fig. 5 shows IID lateralization results for the two groups, broken down by Presentation Level (in dB SL and SPL). In this figure, the three conditions of level are shown on the abscissa and the IID (in dB) inserted into the pointer is shown on the ordinate. The Blind Group lateralized significantly more than the Sighted Group at 25 dB SL (p < .01) and 90 dB SPL (p < .0001) and both groups lateralize significantly more at 70 dB SPL than at 90 dB. The Sighted Group also shows significant differences in lateralization between 25 dB SL and 70 dB SPL and between 70 and 90 dB SPL while the Blind Group shows significant differences between 70 dB SPL and 90 dB SPL and 25 dB SL and 90 dB SPL. Thus, while the Sighted Group lateralizes significantly more at 70 dB SPL than at the other two presentation levels, the Blind Group shows no difference between 25 dB SL and 70 dB SPL.

Figure 6. The significant interaction of Group, Presentation Level and ITD. The axes are identical to Figure 2. Panels (a)-(c) contain data obtained at 25 dB SL, 70 dB SPL, and 90 dB SPL, respectively. The parameter is Group. Error bars indicate the SE of the mean. Asterisks between two points denote the significance levels of the pairwise comparisons (* p < .05; ** p< .005; *** p <. 0005).

The results shown in Fig. 5 are further analyzed in Figures 6 (a-c) that illustrate IID as a function of ITD separately for the three presentation level conditions. Figure 6a shows the result at 25 dB SL. Presentation levels for the 25 dB SL condition (in dB SPL) ranged from 25 to 55 dB and 25 to 60 dB SPL for the Sighted and Blind Groups, respectively. Again, the Blind subjects lateralize more than the Sighted subjects at the most peripheral target ITD, 400 µs. At 70 dB SPL (Figure 6b), there is no difference between the two groups at any target ITD. At 90 dB SPL (Figure 6c) the Blind Group lateralizes significantly more than the Sighted Group at every target ITD. As the difference between the group differences seen in the three panels suggests, there is a significant interaction of Group by Level by ITD [F (4,3447) = 4.38, p =0016].

Figure 7. The interaction of Center Frequency, ITD and Presentation Level. The axes are identical to Figure 2. The parameter is presentation level expressed as dB SPL or SL. Panels (a) and (b) contain data from 250-500 Hz and 1500-4000 Hz, respectively.

The 3-way interaction between Center Frequency, ITD and Level is also significant [F (8,3447) = 7.9, p = <0.0001]. The regression lines in panels (a) and (b) of Figure 7 illustrate this interaction, respectively, for the low-frequency band (250-500) and the high frequency band (1500-4000 Hz), where the performance differences between bands are most evident. The data presented reflect the average performance of the two groups. The IID (in dB) inserted into the pointer is shown on the ordinate with ITD (µs) on the abscissa. The parameter is Presentation Level (25 dB SL and 70 and 90 dB SPL). This three-way interaction reflects the fact that the two-way interaction between Center Frequency and ITD changes as a function of Level. At the lower intensities, subjects are better able to detect the ITD differences at lower frequencies than they are at the higher frequencies as reflected by the steeper slopes. However, as intensity level increases, the slopes of the high frequency functions increase indicating stronger lateralization, thus better discrimination of ITD. In fact, although the slopes at 1500-4000 Hz are always less steep than those in the low frequencies, this difference between the high and low frequency slopes decreases with increased intensity.

4. DISCUSSION
4.1 The effect of vision on IID
To the extent that greater magnitude of lateralization for a given ITD, presentation level and center frequency can be considered to represent increased localization abilities, blind listeners appear at least comparable, and often better than the Sighted subjects. This is most evident in the periphery (especially at 400 µs) where localization is poorest in sighted humans (eg Blauert 1997; Middlebrooks and Green 1991). The increased IID range for the Blind Group is especially interesting in view of the fact that ITD is processed primarily in the mid brain (eg in the medial superior olive) by an immutably wired lattice neural network (Yin and Chan 1990). Thus, what is plastic is unlikely to be the ITD but the way it is mapped onto a larger IID space that makes it possible to encode finer differences.

The fact that, across frequency, the Blind subjects in the present study evidenced greater lateralization for the most peripheral stimuli is consistent with the compensatory model discussed in the Introduction and with data of Lessard et al (1998) and Röder et al (1999) in humans and Rauschecker and Kniepert (1994) in binocularly deprived cats. Lessard et al (1998) and Röder et al (1999) found that while congenitally blind subjects and sighted subjects localized equally well for center target stimuli, the blind subjects showed a more sharply tuned or focused attention to peripheral stimuli than did the sighted subjects. Rauschecker and Kniepert (1994) found that binocularly deprived cats were more precise in sound localization (had less mean variance) in the rear-lateral positions and were more accurate than sighted controls when absolute mean errors were evaluated.

This result for the most peripheral targets may reflect the fact that blind travelers must use their hearing in order to maintain a line of travel parallel to walls (Ashmead and Wall 1999) and learn to use peripheral cues more than sighted listeners in order to avoid these obstacles. In addition, it may be that blind listeners have more experience attending to sounds occurring at peripheral azimuths, since sighted listeners have an incentive to habitually orient themselves to place objects of interest within the visual field.

The question has been previously asked whether blind subjects' superior navigation abilities are due to the relative importance (leading to better use) of auditory cues, or an ability to maintain a straight-line travel (Strelow and Brabyn 1982), a skill practiced by the blind pedestrians on a daily basis. Strelow and Brabyn (1982) suggested that since the blind listeners performed worse than blindfolded subjects as an object in the periphery became smaller (wall vs 15 and 5 cm poles) their superior abilities in navigation were due to better utilization of auditory cues rather than an underlying motor skill advantage. In addition, broadband masking noise and sound-attenuating earplugs significantly decreased the blind subjects' performance with travel along a wall.

Ashmead and colleagues (1998a; 1999) investigated the specific acoustic information available to travelers in order to detect a wall on their periphery. They found that listeners were able to detect an ambient spectral shift (not self-produced), not necessarily an overall increase in sound level. This spectral information was towards the low frequency region at an average distance of 47 cm from a wall. This finding was consistent with another result (Ashmead et al 1998b) which showed that the walking paths of blind children were straighter and more parallel in hallways that were narrower than 3 m.

Our Blind subjects were high functioning, independent travelers who use steel-tipped canes, which generate clicks as they walk. It may be that these high functioning blind travelers use the aforementioned spectral cues from the periphery and therefore have developed better peripheral localization (lateralization) cue calibration than sighted listeners (cf Loomis et al 2001).

Our findings, like those of Ashmead et al (1998b), who examined a wide variety of functional spatial localization tests such as reaching or walking to the locations of sound sources and reacting to minimum audible changes in location, show that although group differences were statistically significant, individual blind listeners are not necessarily superior in spatial hearing to sighted listeners. Our results do show us, however, that Blind listeners in this study were fully able to utilize the cues for spatial hearing. We must caution that our results cannot be generalized to all blind listeners because of the small sample size, the fact that our subjects were all high-level functioning travelers and that only those with a single congenital etiology were tested. It may be that some people who rely on guide dogs do so because they lack the superior auditory abilities in evidence in the subjects in this study. This question will be addressed in a future study.

It has been suggested that under certain conditions one can generalize from lateralization results to localization ability (Plenge 1974). The present findings suggest that the calibration of human spatial hearing is not necessarily dependent upon vision but that vision has a facilitative role (cf Ashmead et al 1998b). In sighted persons, a spatial location is registered by way of multimodal inputs such as vision and audition (smell is not reliable). The McGurk effect (1965) shows that vision will prevail in case of conflicting cues. The data only support the conclusion in that, in the absence of visual input to support or counter the auditory, the Sighted subjects did not do as well as the Blind subjects. Perhaps the difference is the edge the Blind listener obtains through learning, to compensate for the lack of visual input.

A question frequently asked in human as well as animal studies is whether the fact that, in some conditions, blind subjects perform better than sighted subjects is due to compensatory plasticity (for which there is a great deal of evidence) in the visual and auditory systems. This compensation either may be due to the reorganization of neuronal populations (eg Rauschecker 1995); Lessard et al 1998), improved learning (Lessard et al 1998), or the sharpening of the non-visual senses (Rauschecker 1995). While this question cannot be answered in the context of this paper, there has been a great deal of evidence of compensatory plasticity in the visual and auditory systems.

It is recognized that the extent of reorganization is dependent upon the time of onset (Brainard and Knudsen 1998; Hubel and Wiesel 1977) and degree of blindness (Lessard et al 1998). Recent studies have consistently demonstrated that congenially blind human subjects behave differently than blind subjects who lost their sight after puberty. For example, [18F] fluorodeoxyglucose (FDG) PET studies have demonstrated elevated metabolism in the visual cortex of early but not late blind subjects (Veraart et al 1990; Wanet-Defalque et al 1988). This finding was non task-specific (i.e. at rest or with tactile or auditory tasks). However, Buchel et al (1998) showed task specific activation of extrastriate visual and parietal association areas with Braille reading as compared to auditory processing in congenitally blind subjects. With braille-reading adults, Buchel et al (1998) and Sadato et al (1996) further found that blind subjects who lost their sight after puberty show additional activation in the primary visual cortex with the same tasks. These latter findings suggest that the presence of early visual stimulation is necessary for the primary visual cortex to respond to meaningful touch stimuli.

Since the primary visual cortex in blind-raised monkeys did not show crossmodal responses and the fact that there was differential activation of late and congenitally blind subjects, Buchel et al (1998) raise the possibility of reciprocal activation by visual imagery in subjects with early visual experience. Although there were no differences in activations between late and congenially blind subjects in auditory processing relative to Braille reading in this study, there was greater activation in bilateral temporal regions, extending more posteriorly in the blind relative to the sighted subjects. Whereas both blind groups showed a marked activation for auditory processing, the sighted group showed similar levels of activity for print reading and auditory processing (Buchel et al 1998). Although others (eg Dehay et al 1989) have found that early visual deprivation leads to changes in the structural anatomy of the visual cortex at the microscopic level in non-human primates, there was no evidence of morphological differences in the visual cortex of blind and sighted subjects.

Using ERP (Kujala et al 1992; Rosler et al 1993) and magnetoencephalography (MEG), non-specific activation of the visual cortex has been found in blind subjects with auditory and/or tactile tasks.

Finally, the calibration of auditory localization may occur early in life. Ashmead et al (1998b) found similar results in blind children as young as 12 years of age as did Knudsen (1999) and Rauschecker and Kniepert (1993) in animal models where the plastic developing system permits the establishment of new connections and take-over of some functions. It is known that there is a need to recalibrate sound-localization cues during growth (Ashmead and McCarty 1991; Knudsen 1988.) This suggests an important role for continued plasticity in which the nonvisual processes are sufficient (Ashmead et al 1998b).

4.2 The effect of center frequency and ITD on IID.
For the most part, our subjects behaved as expected from studies by Blauert (1997), Bernstein and Trahiotis (1985b), Trahiotis and Bernstein (1986) and Simon et al (1994). For the 100-Hz bandwidth narrow band signals, both groups of subjects used significantly greater values of IID inserted into the pointer to match the position of a non-zero target at frequencies below 1500 Hz than they did in the higher frequencies. In other words, stimuli were perceived closer to midline as the center frequency of the non-zero ITD increased from 1250 Hz to 4000 Hz. Thus, both groups lateralized significantly more at the lowest frequency than at the mid-frequency region and significantly more at the mid-frequency region than at the highest frequency region. However, the Blind subjects lateralized significantly more than the Sighted subjects at the two lowest frequency regions, especially for the peripheral ITD targets. In fact, the Blind subjects lateralized as much in the mid-frequency region, where time becomes less of a cue for lateralization (Yost and Hafter 1987), as the Sighted subjects did in the lowest frequency region. In this sense, the Blind subjects were better at utilizing a diminishing and possibly more ambiguous cue.

Although not evident in the grouped data, all subjects showed greater variability at the higher frequencies and at larger ITDs (cf Sayers 1964; Simon et al 1994; Yost and Hafter 1987). In the low frequencies, our subjects, especially the Blind Group, required IIDs greater than the 12 dB found by other investigators to correspond to an image fully lateralized by ITD (at 400 µs).

4.3 The effect of level on IID.
The audiogram in Fig. 1 indicates that the mean high frequency thresholds of the Blind listeners were slightly poorer than those of the Sighted subjects. However, it is unlikely that this threshold difference explains the difference in performance of these two groups. If performance of the Blind were related to the poorer high frequency thresholds, we would expect to see poorer performance in this group of subjects at the lowest intensity levels (25 dB SL) and highest frequencies and certainly differences reflected at each ITD. This was not the case.

The significant effect of level and its significant two-way and three-way interactions are consistent with the findings of Simon et al (1994). In both studies it appears that the effect of level is discontinuous and thus only enabling above a certain threshold intensity, especially in the higher frequencies (Figure 7a, b). The interaction of the poor temporal resolution in the high frequencies coupled with the low energy signals at 25 dB SL (and at 60 dB SPL in Simon et al 1994) resulted in decreased discrimination of the ITD targets in comparison to the higher intensity signals. As Simon et al (1994) noted, most studies of lateralization have used presentation levels below 60 dB SPL. Since Häusler et al (1983) found that normal hearing subjects performed independently of overall intensity level above 20 dB SL using broadband signals, it was assumed that presentation level above threshold was not particularly important in lateralization. However, the results of the present study and the Simon et al study show that there may not be sufficient energy for the processor of ITD to function in lateralization below 70 dB SPL, at high frequencies.

The limit in processing the low intensity signals used in lateralization in this study and Simon et al (1994) may be due to the loss of phase-locking for this level of signal (Javel et al 1988). There is the possibility that at lower SPLs, such as used in these studies, the discharges elicited by the signal in the auditory nerve fibers do not cluster within the stimulus period at any point. It is more difficult to conjecture about why the 90-dB SPL signals at the lowest frequencies (Figure 7a) are less well discriminated. Although no subjects specifically complained, it is possible that the intensity in one ear was becoming uncomfortable when subjects were trying to match the 400 µs ITDs. For the 90 dB SPL presentation level, the IID on the lateralized side could be at high as 100 dB SPL for a fully lateralized signal, and the subjects may have opted not to use the pointer to achieve full lateralization if the signal was becoming uncomfortably loud.

However, Figure 5 and Figure 6a and 6b show that the Blind subjects lateralize significantly more than the Sighted subjects at 25 dB SL and at 90 dB SPL. At 70 dB SPL, which is loud conversational level, there is no significant difference in performance between the two groups. Across Center Frequency and ITD, the Blind subjects lateralize as fully at 25 dB SL as they do at 70 dB SPL, while the Sighted subjects lateralize significantly more at 70 dB SPL than at 25 dB and 90 dB SPL. As stated above, the Blind subjects show greater lateralization than the Sighted subjects, even when the task may be more difficult, either due to a signal that is may be too soft (25 dB SL) or too loud (90 dB SPL). Again, this may show that the Blind subjects make better use of a diminishing or uncomfortable auditory cue.

4.4. Conclusions
Plasticity, reorganization, and peripheral ability of the blind have been previously recognized (cf Rauschecker 1995). For all blind children, it is important to take advantage of this plasticity and sensory ability at an early age. In addition, it may be helpful to recognize this superior auditory ability when designing the next generation of sensory prostheses.

Acknowledgements: This research was supported by grants from NIDCD (R29DC00468-04) and NIDRR (H133G20048-94) as well as a grant from the Smith-Kettlewell Eye Research Institute (Dr. Simon and Mr. Lotze). Dr. Divenyi is supported by a grant from NIA (AG-07998) and by the VA Medical Research. Special thanks to Brennan McBride for software development, Albert B. Alden and Steven T. Chung for hardware development, Judith Paton for audiological evaluations, Jesse Canchola for statistical analysis and to Inna Aleksandrovsky, Elaine Goduti and Lani Hardage for assistance during data collection and analysis. We also wish to thank John Brabyn, Ph.D., Deborah Gilden, Ph.D., E. Wm. Yund, Ph.D., and the reviewers for their helpful comments and suggestions.

REFERENCES

Amenta C A, Trahiotis C, Bernstein L R, and Nuetzel J M, 1987 "Some physical and psychological effects produced by selective delays of the envelope of narrow bands of noise" Hearing Research 29 147-161.

ANSI, 1996 "American national specifications for audiometers (ANSI S3.6-1996)" American National Standards Institute.

Ashmead D, Wall R, Eaton S, Ebinger K, Snook-Hill M, Guth D, and Yang X, 1998a "Echolocation reconsidered: Using spatial variations in the ambient sound field to guide locomotion" The Journal of Visual Impairment and Blindness 9 615-632.

Ashmead D H and McCarty M E, 1991 "Postural sway of human infants while standing in light and dark" Child Development 62 1276-87.

Ashmead D H and Wall R S, 1999 "Auditory perception of walls via spectral variations in the ambient sound field" The Journal of Rehabilitation Research and Development 36 313-322.

Ashmead D H, Wall R S, Ebinger K A, Eaton S B, Snook-Hill M M, and Yang X, 1998b "Spatial hearing in children with visual disabilities" Perception 27 105-22.

Axelrod S, 1959 Effects of Early Blindness (New York: Am. Found. Blind).

Benedett L H and Loeb M A, 1972 "Comparison of auditory monitoring performance in blind subjects with that of sighted subjects in light and dark." Perception and Psychophysics 11 10-16.

Bernstein L R and Trahiotis C, 1985a "Lateralization of low-frequency, complex waveforms: The use of envelope-based temporal disparities" The Journal of the Acoustical Society of America 77 1868-1880.

Bernstein L R and Trahiotis C, 1985b "Lateralization of sinusoidally amplitude-modulated tones: Effects of spectral locus and temporal variation" The Journal of the Acoustical Society of America 78 514-523.

Blauert J, 1997 Spatial Hearing Revised Ed., (Cambridge, MA: MIT Press).

Brainard M S and Knudsen E I, 1998 "Sensitive periods for visual calibration of the auditory space map in the barn owl optic tectum" The Journal of Neuroscience 18 3929-3942.

Bross M and Borenstein M, 1982 "Temporal auditory acuity in blind as compared with sighted subjects: a signal detection analysis" Perceptual Motor Skills 55 963-966.

Buchel C, Price C, Frackowiak R S J, and Friston K, 1998 "Different activation patterns in the visual cortex of late and congenitally blind subjects." Brain 121 409-419.

Curtis J F and Winer D M, 1969 "The auditory abilities of the blind as compared to the sighted." Auditory Research 9 57-59.

Dehay C, Horsburgh G, Berland M, Killackey H, and Kennedy H, 1989 "Maturation and connectivity of the visual cortex in monkey is altered by prenatal removal of retinal input" Nature 337 265-7.

Domnitz R H and Colburn H S, 1977 "Lateral position and interaural discrimination" The Journal of the Acoustical Society of America 61 1586-1598.

Durlach N I, Thompson C L, and Colburn H S, 1981 "Binaural interaction in impaired listeners" Audiology 20 181-211.

Gabriel K J, Koehnke J, and Colburn H S, 1992 "Frequency dependence of binaural performance in listeners with impaired binaural hearing" The Journal of the Acoustical Society of America 91 336-347.

Hafter E R and Buell T N, 1985 "The importance of transients for maintaining the separation of signals in auditory space" Attention and Performance XI. (New Jersey: L. Erlbaum).

Henning G B, 1974 "Detectability of interaural delay in high frequency complex waveforms" The Journal of the Acoustical Society of America 55 84-90.

Hubel D H and Wiesel T N, 1977 "Ferrier lecture. Functional architecture of macaque monkey visual cortex [Review]." Proceedings Royal Soc Lond B. Biol Sci. 198 1-59.

Javel E, McGee J, Horst J W, and Farley G R, 1988 "Temporal mechanisms in auditory stimulus coding" In Auditory Function, ed. W.E. Gall G. M. Edelman, and W. M. Cowan. (New York: John Wiley and Sons) 515-558

Jones B, 1975 "Spatial perception in the blind" The British Journal of Psychology 66 461-472.

Killion M and Villchur E, 1989 "Comments on 'Earphones in audiometry'" The Journal of the Acoustical Society of America 85 1775-1778.

Knudsen E I, 1988 "Experience shapes sound localization and auditory unit properties during development in the barn owl." In Auditory Function: Neurobiological Bases of Hearing, Eds. G. M. Edelman, W. E. Gall, and W. M. Cowan. (New York: J. Wiley and Sons) 137-149.

Knudsen E I, 1999 "Mechanisms of experience-dependent plasticity in the auditory localization pathway of the barn owl" The Journal of Comparative Physiology 185 305-321.

Koehnke J, Besing J, and Goulet C, 1992 "Speech intelligibility, localization, and binaural detection with monaural and binaural amplification" The Journal of the Acoustical Society of America 92 2434 (A).

Koehnke J and Zurek P M, 1990 "Localization and binaural detection with monaural and binaural amplification" The Journal of the Acoustical of America 88 S169.

Kujala T, Alho K, Paavilainen P, Summala H, and Naatanen R, 1992 "Neural plasticity in processing of sound location by the early blind: an event related potential study" Electroencephalography and Clinical Neurophysiology 84 469-472.

Lessard N, Pare M, Lepore F, and Lassonde M, 1998 "Early-blind human subjects localize sound sources better than sighted subjects" Nature 395 278-80.

Littell R C, Freund R J, and Spector P C, 1991 SAS® System for Linear Models 3rd ed., (Cary, NC: SAS Institute Inc).

Loomis J M, Klatzky R L, and Golledge R G, 2001 "Navigating without vision: Basic and applied research" Optometry and Vision Science 78 282-289.

McGurk E, 1965 "Susceptibility to visual illusions" Journal of Psychology 61 127-43.

Middlebrooks J C and Green D M, 1991 "Sound Localization by Human Listeners" Ann. Rev. Psychol. 42 135-159.

Muchnik C, Efrati M, Nemeth E, Malin M, and Hildesheimer M, 1991 "Central auditory skills in blind and sighted subjects" Scandinavian Audiology 20 19-23.

Neville H J and Lawson D, 1987 "Attention to central and peripheral visual space in a movement detection study: An event-related potential and behavioral study. II. Congenitally deaf adults." Brain Research 405 268-283.

Niemeyer W and Starlinger I, 1981 "Do the blind hear better? Investigations on auditory processing in congenital or early acquired blindness. II. Central functions" Audiology 20 510-515.

Olsho L W, Harkins S W, and Lenhardt M L, 1985 "Aging and the auditory system" In Handbook of the Psychology of Aging, Eds. J. Birren and K. W. Schiver (New York: Van Nostrand Reinhold) 332-377.

Plenge G, 1974 "On the difference between localization and lateralization" The Journal of the Acoustical Society of America 56 994-951.

Rauschecker J and Kniepert U, 1993 "Auditory localization behavior in visually deprived cats" European Journal of Neuroscience 6 149-160.

Rauschecker J P, 1995 "Compensatory plasticity and sensory substitution in the cerebral cortex" Trends in Neuroscience 18 36-43.

Rauschecker J P and Harris L R, 1983 "Auditory compensation of the effect of visual deprivation in cats' superior colliculus" Experiments in Brain Research 50 63-83.

Rauschecker J P and Kniepert U, 1994 "Auditory localization behaviour in visually deprived cats" European Journal of Neuroscience 6 149-60.

Rayleigh L J W S, 1907 "On our perception of sound direction" Phil. Mag. 13 214-232.

Rice C E, 1969 "Perceptual enhancement in the early blind" The Psychological Record 19 1-14.

Rice C E, 1970 "Early blindness, early experience and perceptual enhancement." Research Bulletin of the American Foundation for the Blind 22 1-22.

Roder B, Teder-Salejarvi W, Sterr A, Rosler F, Hillyard S A, and Neville H J, 1999 "Improved auditory spatial tuning in blind humans" Nature 400 162-166.

Rosler F, Roder B, Heil M, and Hennighausen E, 1993 "Topographic differences of slow-event-related brain potentials in blind and sighted adult human subjects during haptic mental rotation." Brain Research 1 145-159.

Sadato N, Pascual-Leone A, Grafman J, Ibanez V, Deiber M P, Dold G, and Hallett M, 1996 "Activation of the primary visual cortex by Braille reading in blind subjects [see comments]" Nature 380 526-8.

Sayers B, 1964 "Acoustic-image lateralization judgments with binaural tones" The Journal of the Acoustical Society of America 36 923-933.

Schiano J L, Trahiotis C, and Bernstein L R, 1986 "Lateralization of low-frequency tones and narrow bands of noise" The Journal of the Acoustical Society of America 79 1563-1570.

Simon H J, Collins C C, Jampolsky A, Morledge D E, and Yu J, 1994 "The measurement of the lateralization of narrow-bands of noise: The effect of sound pressure level" The Journal of the Acoustical Society of America 95 1534-1547.

Starlinger I and Niemeyer W, 1981 "Do the blind hear better? Investigations on auditory processing in congenital or early acquired blindness. I. Peripheral functions" Audiology 20 503-509.

Stern R M, Zeiberg A S, and Trahiotis C, 1988 "Lateralization of complex binaural stimuli: A weighted-image model." The Journal of the Acoustical Society of America 84 156-165.

Strelow E R and Brabyn J A, 1982 "Locomotion of the blind controlled by natural sound cues" Perception 11 635-40.

Tonning F M, 1975 "Ability of the blind to localize noise: Practical consequences" Scandinavian Audiology 4 183-186.

Trahiotis C and Bernstein L R, 1986 "Lateralization of bands of noise and sinusoidally amplitude-modulated tones: Effects of spectral locus and bandwidth" The Journal of the Acoustical Society of America 79 1950-1957.

Trahiotis C and Stern R M, 1989 "Lateralization of bands of noise: Effects of bandwidth and differences of interaural time and phase" The Journal of the Acoustical Society of America 86 1285-1293.

Veraart C, De Volder A G, Wanet-Defalque M C, Bol A, Michel C, and Goffinet A M, 1990 "Glucose utilization in human visual cortex is abnormally elevated in blindness of early onset but decreased in blindness of late onset," Brain Research 510 115-121.

Wanet M C and Veraart C, 1985 "Processing of auditory information by the blind in spatial localization tasks" Perception and Psychophysics 38 91-96.Wanet-Defalque M C, Veraart C, De Volder A, Metz R, Michel C, Dooms G, and al. e, 1988 "High metabolic activity in the visual cortex of early blind human subjects." Brain Research 446 369-373.

Wightman F L and Kistler D J, 1992a "The dominant role of low-frequency interaural time differences in sound localization" The Journal of the Acoustical Society of America 91 1648-1661.

Wightman F L and Kistler D J, 1992b "Perceptual consequences of engineering compromises in synthesis of virtual auditory objects" The Journal of the Acoustical Society of America 92 2332A.

Wilber L A, Kruger B, and Killion M C, 1988 "Reference thresholds for the ER-3A insert earphone" The Journal of the Acoustical Society of America 83 669-676.

Yin T C and Chan J C, 1990 "Interaural time sensitivity in medial superior olive of cat" Journal of Neurophysiology 64 465-88.

Yost W A and Hafter E R, 1987 "Lateralization" In Directional Hearing, Eds. W.A. Yost and G. Gourevitch. (New York: Springer-Verlag) 49-84.