Main

BI, referring to those aspects of hearing for which two ears are necessary, is the substrate for psychoacoustic functions of sound localization and lateralization that are very important for the development of language and learning skills. This phenomenon is thought to represent electrophysiologic activity of the neural elements in the brain subserving binaural processing and is produced by binaurally influenced neurons central to the cochlear nucleus. The neural BI can be examined from BAER, an objective tool reflecting functional integrity of the auditory brainstem pathway. Since BI in the BAER was first described in the cat by Jewett(1) in 1970, this phenomenon has been examined in human adults by a number of investigators(2–17), but very few studies have been conducted in infants(6, 12). There is little information available regarding the time course of BI development. Stimulus conditions are known to have significant effects on the BAER. However, whether stimulus conditions such as intensity level and repetition rate have significant effects on the infant BI is to be clarified.

The BI depends on functional integrity of binaurally innervated neurons along the central auditory pathway. Development of BI in the BAER relies on the maturation of binaurally innervated neurons in the brainstem. Alterations in the BI during early childhood may impair speech development and comprehension. Animal experiments have suggested that early auditory experience affects the development of binaural processing in the brain(18). Binaural hearing is more susceptible to alterations in sound experience than other types of hearing. The period of normal development and susceptibility to altered sound experience is more protracted for binaural processing than for other aspects of auditory perception. An animal experiment has shown that conductive hearing loss caused marked alteration in the maturation of BI components in the BAER(19). In a human study of the long-term effects of early conductive hearing loss on auditory electrophysiology, Gunnarson and Funitzo(20) found a significant difference between the affected children and controls in the presence of the BI. These observations suggest that the BI is likely to be a sensitive index for exploring altered development or neural abnormalities in central auditory pathways.

The latency range of BI components in man has been reported to occur between the latencies of BAER wave V to wave VI(3, 5, 7, 9). In term infants, Hosford-Dunn et al.(6) described only one BI component that appeared shortly before BAER wave V, whereas McPhersonet al.(12) reported that BI wave form occurred during BAER waves IV, V, and VI. In the adults we have observed that the BI wave form occurs beyond these latency range and is composed of seven components that coincide consistently with the latency range of BAER waves IV through VII. Wave DVII, which has never been described before, is the most prominent BI component at high stimulus intensity levels. In this study we sought to ascertain whether the BI components found in the adults exist in term neonates and to examine the effects of stimulus intensity on the BI to gain insights into the maturational properties of binaural processing in human infants at birth.

METHODS

Subjects. Fifteen full-term neonates were recruited into this study. Their BAER thresholds met the following two criteria: 1) monaural response threshold in the infants better than 20 dB HL and2) BAER threshold interaural difference less than 5 dB. All neonates had a birth weight above the 10th percentiles of normal population and ranged in conceptional age (gestational age plus chronologic age) between 37 and 42 wk. The gestational age was determined by maternal history or, if there were no adequate obstetric data, by Dubowitz criteria(21). All infants were judged to be stable and healthy in the nursery with an Apgar score greater than 8 at 5 min. Infants with any perinatal complications such as asphyxia, periventricular hemorrhage and intrauterine infection or growth retardation, hearing loss, or congenital abnormality of CNS were excluded. Seven adults with normal hearing and no known neurologic disorders served as controls.

BAER recording and BI deriving. The test was undertaken between 1 and 11 d after birth in an electrically shielded, darkened, and soundproofed room. Infants lay supine comfortably in their cribs. The BAER was recorded during unsedated sleep after breast or milk feeding. Three silver cup electrodes, filled with commercial electrode paste, were placed along the midline of the scalp with an active electrode (positive) at the vertex, an inactive electrode (negative) at the nape of the neck, and a common (ground) electrode at the upper forehead.

Acoustic stimuli to evoked BAER were alternate clicks that were generated by passing 0.1-ms square pulses through an attenuator to a pair of matched TDH-39 earphones encased in cushions. Brain responses were band passed(100-2000 Hz) and averaged by a signal processor. For each recording, the responses to 2048 stimulus repetitions were averaged. BAER to monaural and binaural stimulation were recorded separately. The first session of recording was made in the sequence of left monaural, right monaural, and binaural stimulation. The second session was simply a replication of the recording but in reverse sequence. Two or more runs were made for each recording condition. The sequence for BAER recording was the responses to left, right, and binaural stimulation. Replicate BAER recordings were made in reverse order. The clicks were started at the intensity of 70 dB HL and the repetition rate of 10/s. Previous studies have shown that BI wave form could be contributed by ACT at high stimulus intensity, and this effect can be reduced or avoided by decreasing stimulus intensity to moderate and low levels(3, 5). The maximum level to avoid ACT is 46 dB above the threshold of the opposite ear. Because BAER usually cannot be detected until it is at least 5 dB above threshold, a level of around 51 dB may be still safe(5). Thus, to examine and reduce the effect of ACT on BI components, subjects were tested at both 50 and 70 dB HL. To observe the intensity effect in more detail, subjects were also tested at 30 and 10 dB HL clicks and, if the subjects kept sleep, the test was continued at 80, 40, and 20 dB HL clicks. Some subjects were also tested at faster stimulus repetition rates: 20, 40, and 60/s.

The two monaural traces were digitally summed to yield a single trace,i.e. summed monaural response. The BI was derived by subtracting the binaurally evoked trace from the summed monaurally evoked trace;i.e. (left monaural + right monaural) - binaural or, in short, BI =(L + R) - B.

RESULTS

BI wave form. In adult controls seven BI components, three upward and four downward, were identified, which coincided consistently with the latency range of BAER waves IV through VII. Based on the latencies, these BI components were designated in association with the corresponding BAER waves (Fig. 1). Waves DV, DVI, and DVII occurred in the region of down slope of BAER waves V, VI, and VII, respectively. Waves DV′, DVI′, and DVII′ were the negative troughs immediately after the three upward BI components in sequence and wave DIV′ was the trough between BAER waves IV and V. Here, D referred to the difference between the summed monaural response and the binaural response. Waves DV and DVII were the two most prominent, consistent and reproducible BI components in the adults. The time between various BI components, especially waves DV, DVI, and DVII, were fairly constant. Thus, three interpeak intervals of the BI,i.e. DV-DVII, DV-DVI and DVI-DVII, were developed.

Figure 1
figure 1

Sample recordings of summed monaural response(dashed line) and derived BI wave form (solid line) showing seven BI components and their latencies in association with the corresponding BAER components (labeled with italic numerals). The BI is derived by subtracting the binaurally evoked response from the sum of the monaurally evoked responses, i.e. BI = (L + R) - B.

Full size image

As shown in Figure 1, most of the BI components seen in the adults could be identified in the neonates. In general, the later BI components, i.e. those with longer latency, were less differentiated than the earlier ones, i.e. those with shorter latency. The most consistent and reproducible BI component was wave V. A marked difference between the neonatal and adult BI wave form was that neonatal wave DVII was much smaller relative to adult wave DVII, especially at high click intensity levels. Wave DVI in the neonatal BI wave form was relatively more prominent. These differences were largely related to the fact that the amplitude of BAER waves VII was much smaller than wave VI in the neonates, whereas the two amplitudes were similar in the adults. Wave DV in the neonates, as in the adults, was the most reliable BI component at all stimulus conditions used here and tended to be relatively more distinguished at moderate click intensities than at high and low intensities. Wave DVI usually was the most prominent BI component at high stimulus intensity levels, but diminished markedly at moderate levels. These characteristics of wave DVI were similar to those of wave DVII observed in the adults. Wave DVII in the neonates was very small even at high intensity levels. Waves DIV′ and DV′ could be identified in most neonates. Waves DVI′ and DVII′ were usually poorly differentiated. At the stimulus condition of 10/s and 70 dB HL, the seven BI components from DIV′ to DVII′ in sequence were identified in 14, 15, 11, 12, 8, 10, and 6 of the 15 neonates, respectively.

The latencies of all BI components in the neonates were significantly longer than those in the adults. This was associated or concomitant with the longer BAER wave latencies in the monaural and binaural responses (Fig. 1). Tables 1 and 2 present, respectively, latencies and interpeak intervals of the BI components and the corresponding BAER waves recorded at 70 dB HL clicks with a repetition rate of 10/s.

Table 1 Means and standard deviations for latencies (ms) of BI components and relevant BAER components in the summed monaural and binaural responses at 70 dB HL and 10/s clicks
Full size table
Table 2 Means and SD for interpeak intervals (ms) of BI at 70 dB HL and 10/s clicks
Full size table

Compared with those in the adults, the amplitudes of BI components were very small in the neonates. This difference was roughly in proportion to the difference in BAER wave amplitudes between the neonates and the adults (Fig. 1). Wave DV amplitude in the neonates was equal to 28% (range of 17-36%) of the amplitude of BAER wave V in the summed monaural response at 70 dB HL. This was slightly greater than that in the adults (24%, range of 16-28%). The amplitude of wave DVI in the neonates accounted for 36%(range of 18-62%) of the amplitude of BAER wave VI.

Change in click repetition rate had a significant effect on BI wave form morphology. Figure 2 shows sample BI traces from a neonate and an adult at various click rates. At the click rate of 10/s, the BI components were generally well differentiated. When click rate increased to above 20/s, the morphology of BI wave form tended to be altered or deteriorated. Some BI components were poorly differentiated in the neonates and difficult to reliably identify from the baseline. In the adults, nevertheless, most BI components were still reliably identifiable. The latencies of BI components increased slightly with increasing click rate, which was more significant for the later components than for the earlier ones. There was also a tendency for a slight increase in interpeak intervals at faster click rates. The amplitudes of BI components usually reduced slightly with increasing repetition rate, although some amplitudes were occasionally increased slightly.

Figure 2
figure 2

Derived BI trances from a neonate and an adult at the click repetition rates between 10 and 60/s. Click intensity is 70 dB HL. Thedashed line at the top of neonatal BI traces is the summed monaural recording and the italic numerals are labels of BAER waves. At high click rates, BI wave form morphology tended to be altered. This is more significant in the neonate than in the adult.

Full size image

Neonatal BI wave form was particularly easy to be contaminated by muscle artifacts, making it difficult to reliably identify BI components. In the infants who were restless during the test, it was extremely difficult to get satisfactory recordings. Therefore, it is critically important to keep the subject thoroughly relaxed or asleep during the test.

Change in BI components with stimulus intensity. Click intensity levels had a significant effect on BI wave form. At high intensities(80-60 dB HL), most BI components could be clearly identified. As the intensity decreased, the later BI components, particularly wave DVI, diminished to a greater extent than the earlier ones, so that the earlier components such as waves DIV′ and DV became relatively more prominent at moderate intensities (50-30 dB HL). When the clicks were reduced to below 30 dB HL, most BI components became indistinct from the baseline. Only wave V still could be identified at the low levels in some neonates.

Figure 3 illustrates typical BI wave forms at various click intensities in a neonate. The shift of the BI wave latencies was more significant at lower intensities than at higher intensities. The amplitudes of BI components generally reduced with decreasing intensity. The DV-DVII interval was essentially constant, whereas the shorter intervals of DV-DVI and DVI-DVII were relatively variable. The change in the latencies with intensity in the neonate was more significant than that in the adults.

Figure 3
figure 3

Sample BI traces at different click intensity levels in a neonate. The click rate is fixed at 10/s. The dashed curves are the latency-intensity functions for waves DV, DVI, and DVII, respectively, in an adult. To show more clearly the comparison of the corresponding functions between the neonate and adult, the adult functions are adjusted to the appropriate latencies to fit the peaks of waves DV, DVI, and DVII at 70 dB HL to those in the neonate (top BI traces). Note that the slopes of latency-intensity functions in the neonate were steeper that those in the adults.

Full size image

Figures 4 and 5 present graphically the latencies and interpeak intervals of the three major BI components,i.e. waves DV, DVI, and DVII, at the click intensities of 70, 50, and 30 dB HL in the neonates and show comparison with those in the adults. Most data at the click intensity levels less than 30 dB in the neonates were omitted from these figures, because the latencies and intervals could not be reliably measured at these low intensities in most recordings. The latencies of BI components increased as a function of decreasing click intensity. All latencies shifted by roughly the same amount. The latency-intensity functions were slightly steeper at lower intensity levels than at higher intensity levels. Regression analysis showed that the slopes of these functions in the neonates were significantly steeper than those in the adults (all p< 0.01), suggesting that the shift of the latencies is greater in the neonate than in the adults. This also could be seen clearly in Figure 4. Between 70 and 30 dB HL clicks, the latency shifted in 0.056, 0.059, and 0.061 ms/dB for waves DV, DVI, and DVII, respectively, in the neonates and, in 0.033, 0.038, and 0.037 ms/dB, respectively, in the adults. The interpeak intervals of BI components were essentially constant at various click intensity levels, although the DV-DVII interval showed a tendency to increase slightly with decreasing intensity (Fig. 5).

Figure 4
figure 4

Latency-intensity functions for BI components at the click intensities of 70, 50, and 30 dB HL with a fixed repetition rate of 10/s. The slopes in the neonates (solid lines) are steeper than those in the adult controls.

Full size image
Figure 5
figure 5

Means of the interpeak intervals of BI components at the same stimulus conditions as in Figure 3. The interval of DV-DVII tends to increase slightly with intensity decreasing.

Full size image

DISCUSSION

The present study revealed that most of the BI components in the BAER seen in the adults could be identified in term neonates, but these components, particularly the later ones such as wave DVII, are immature. Changes in the BI components with stimulus intensity and rate in the neonates were more significant compared with the adults. These findings suggest that neural connections in the human brainstem subserving binaural processing are established at birth but, especially at higher levels of the brainstem, are immature(22, 23), and are in accordance with the behavioral observation that human neonates are able to orient toward the source of a novel sound, but their accuracy of orientation is poorer than that of older infants.

Immaturity of BI components at birth. The BI components in the neonates, as in the adults, occur consistently between the latency range of BAER wave IV through wave VII, which is beyond the range described previously in the literature. It appears that there are binaurally neuronal interactions that generate in the brainstem structures rostral to the generators for BAER wave V.

The BI in humans was reported to be mature at or soon after birth, changing only inasmuch as the absolute latencies of the wave form peaks change with development(6). Hosford-Dunn et al.(6) reported that the gross response properties of BI were similar in neonates and adults. McPherson et al.(12) also claimed that BI which occurred at the time of BAER wave V was comparable in term infants and adults. The present study revealed, however, that although most of the BI components observed in the adults could be identified at birth, some of these components, especially the later ones, such as wave DVII, were apparently underdeveloped in wave form. These findings indicate an immaturity in neuronal responses contributing to these BI components. The longer latencies and smaller amplitudes of the BI components in the neonates were associated or concomitant with the longer latencies and smaller amplitudes of the BAER waves in the monaural and binaural responses. It is thought that age-related wave form changes in the BAER reflect development of synaptic and dendritic properties, whereas latency changes primarily reflect myelination(24). The underdeveloped BI wave form and lower wave amplitudes in the neonates reflect an immaturity in synaptic and dendritic properties of binaurally innervated neurons in the brainstem and the longer wave latencies is primarily due to the uncompleted myelination along the auditory brainstem pathway. The finding that the later BI components are particularly immature compared with the earlier ones in the neonates is in accord with the well established centripetal sequence of brain development. It appears that the generators of or neuronal responses contributing to wave DVII are particular immature at birth.

Our findings indicate that although the neuronal connections in the human brainstem for the BI are established and functioning at birth, the binaural processes intrinsic to the brainstem, especially at higher levels, are immature. Similarly, in the young ferrets we observed that, despite some differences, the overall anatomical topography of the projection from the brachium of the inferior colliculus to the superior colliculus is essentially the same as that in the adults(25, 26). However, the neuronal responses for sound location in the superior colliculus are far from mature. It seems that the immature functional representation of auditory space in the young ferrets is unlikely due to an immaturity in the gross innervation pattern from the brachium of the inferior colliculus, but, more likely, reflects an immaturity in the functional properties of those neurons. It is speculated that postnatal maturation or refinement of the neural connections subserving binaural processing is influenced substantially by sound experience or acoustic inputs from the two ears after birth.

The change in the BI components with varying click rate was more significant for the later BI components than for the earlier ones. This may be produced by the accumulation of stimulus rate effect from the earlier components on to the later ones. The finding that rate effect was more significant in the neonates than in the adults suggests that the activity of binaurally innervated neurons in the neonates is more vulnerable to fast acoustic stimulation than in the adults. This provides further evidence for functional immaturity of neural elements in the brainstem in general and binaurally innervated neurons in particular.

The morphology of BI wave form tended to be altered or deteriorated at faster click rates. Most BI components are difficult to reliably identify at the stimulus rates of above 20/s due to the alteration in the wave form produced by fast rate stimulation. This makes the measurement of latency and amplitude difficult. Slower stimulus rates, such as 10 or 20/s, could produce a satisfactory BI wave form and are therefore preferred for the routine test of the BI.

Muscle artifacts have a substantial influence upon the BI components. This is because BI components are much smaller compared with the BAER components and only slightly larger than the background noise. Such a problem is particularly significant in infants. To obtain satisfactory and reproducible BI wave form, it is essential to keep the infants in a comfortable position so that the muscles, especially in the head and neck, thoroughly relaxed and, if possible, to keep the infants asleep during the test. Recording should be interrupted whenever there appear too many muscle activities on the monitor.

ACT and the later BI components in the neonates. ACT may have a significant effect upon BI wave form at high intensity stimulation. The BAER may include not only the responses of the ipsilateral ear to the direct stimuli but also those of the contralateral ear to the ACT. Consequently, BI component produced by ACT rather than by neural interaction will be present in the BI wave form at high intensity stimulation. Levine(5) claims that the effect of ACT can be avoided at and below the intensity of 50 dB HL. It is therefore important to assess the BI at the stimulus levels both above and at or below 50 dB HL to distinguish neural BI from the effect of ACT.

In the adults, wave DVII was the most prominent BI component at high intensity stimulation, but was reduced markedly at moderate intensity levels, suggesting that ACT accounts for at least part of DVII wave form at high intensity stimulation. This component still could be clearly identified at the intensity levels below any possibility of ACT (50 dB HL and less). Thus, it is clear that wave DVII has neurogenic elements in origin and is not entirely attributed to ACT. Because the head size in neonatal infants is much smaller than that in adults, the ACT effect ought to be more significant in neonates than in adults. Any contributions to the BI wave form by the ACT effect should be more pronounced in neonates, and so wave DVII ought to be more prominent if this component is produced primarily by the ACT effect. In fact, wave DVII in the term neonates is much less prominent relative to that in the adults, even at the high intensity stimulation. We have made preliminary observation on the BI in preterm infants and noted that this component is hardly identifiable up to 35-36 wk of gestation. This provides further evidence that wave DVII is neural in origin, although ACT may have an effect on this component at high stimulus intensity levels.

The present work revealed that wave DVI usually was the most prominent BI component at high intensity stimulation but was reduced markedly with decreasing intensity from high to moderate levels, implying that ACT may partially contribute to this component at high intensity stimulation. Because wave DVI can be identified at high, moderate, and, in some cases, low intensity stimulation, this component, like wave VII, is most likely to be neural in origin.

Effect of stimulus intensity on BI components. We have previously observed that the latencies and amplitudes of the BAER in children change as a function of the stimulus intensity(27, 28). All latencies shifted by roughly the same amount so that the latency-intensity functions for all BAER components appeared to parallel each other, and the interpeak intervals changed little at various intensity levels. Thus, the changes in wave latencies with stimulus intensity may be transferred originally from the intensity-dependent change in wave I latency,i.e. the peripheral auditory system. Because the slopes of intensity-functions of BI components also parallel each other and the interpeak intervals were essentially constant, the intensity effect on the BI may occur before wave DV and is most likely to originate from the peripheral auditory system.

There was an age-related difference in the latency-intensity functions of BI components. The slopes of the functions were steeper in the neonates than in the adults, suggesting that the shift of the latencies with stimulus intensity is more significant in the neonate than in the adults. This difference may occur before the peak of BAER wave I in origin or at the peripheral part of the auditory pathways because the latency-intensity functions parallel each other for all of the BAER and BI components and the interpeak intervals change little with varying stimulus intensity in both the neonates and the adults. Thus, the age-related difference in the slope of latency-intensity functions primarily reflect the immaturity of the peripheral auditory mechanisms in response to sound stimulation at different intensity levels early in life.

Future study on the development and potentially clinical usefulness of BI. Although there have been sporadic reports on the BI in term neonates, little is known about the time course of development of the human BI. Until now, we do not know when various BI components emerge and reach adult-like. Thus, future study of BI development during the preterm and postneonatal period is particularly needed. In view of the small amplitude of BI wave form and the instability and immaturity of BAER in very young infants, it is likely that there will be some difficulty in deriving reliable BI in preterm infants. Although the neural connections in the brainstem subserving the BI are established at birth, refinements in these connections would occur over a long postneonatal period. Such findings will enable us to have better knowledge of the development of human BI and to establish the baseline data for the assessment of BI in clinical setting.

Whether the BI test has the potential to be a useful tool in the study of normal and abnormal neural development and in the detection of defects in the brain remains to be shown. Because BI is produced by binaurally innervated neurons, which account for more than 80% of the auditory neurons, it is possible that the BI, like the BAER, may show some changes associated with developmental abnormalities or disorders that involve the brainstem. For instance, in infants after birth asphyxia, we have revealed that the major abnormalities in the central BAER components was amplitude reduction of wave V, indicating that asphyxia has a major effect on the rostral brainstem, a crucial place for generating the BI(29). It is likely that there will be some changes in the functional properties of binaural innervated neurons in these infants, which may be demonstrated by the BI test. In view of the fact that the auditory brainstem pathway is close to many other neural centers in the brainstem and the BI is a unique aspect of the BAER, it is also possible that the BI test may provide novel information with respect to functional status of other neural centers in the brainstem and to detect some defects in neural function and central sensory processing that may not be disclosed by the conventional BAER test or other approaches. For instance, preterm infants experience inappropriate auditory stimulation during the preterm period. In a previous study we failed to find any significant abnormality in the conventional BAER test in the low risk preterm infants up to 6 y of age(30). Nevertheless, we do not rule out the possibility that there might be some subtle degree of functional alteration in the developing auditory brainstem such as aberrant binaural hearing that cannot be detected by conventional BAER test. If so, the finding is very important for us to formulate and implement strategies for modifying the preterm auditory experience.