Brainstem auditory evoked potentials (BAEP) are an essential tool in audiological diagnosis, especially in young children. The electrophysiological response to stimuli such as clicks, or tone bursts is widely used in clinical practice to evaluate neural integrity and to predict auditory thresholds.1,2

Brainstem auditory evoked potentials in response to complex sounds, such as speech sounds, investigate the neural representation of these sounds at subcortical levels, and faithfully reflect the stimulus characteristics. However, there are still few studies that use this type of stimulus, and the understanding of these auditory responses could help elucidate of the effects of age on the development of the auditory system, in addition to being useful in the evaluation of communication difficulties, learning deficits, peripheral auditory deficits, auditory neuropathies or the indication for cochlear implants.3

The characteristics and maturation of electrophysiological responses to speech sounds during the first year of life are still unclear.4 The organ of Corti develops before birth,5 but the maturation of the auditory pathway continues up to adolescence.6,7

Some studies suggest that the perception of speech sounds is strongly influenced by innate factors; soon after birth neonates are able to detect differences between sounds, including sounds they have never heard before.8 It can also be observed that they have a similar pattern of sound perception, regardless of the language environment to which they have been previously exposed.8–11 It is believed that the latency values and response morphology for speech stimuli reach the adult pattern at around 5 years of age.12

The auditory response evoked by speech stimulus can be subdivided into two parts:

Transient portion: constituted by the components of the onset response (beginning of the stimulus); this occurs in the first 10ms and is similar to the response to a click stimulus, with millisecond precision.

Sustained portion: constituted by the Frequency Following Response (FFR), which occurs between 18 and 40ms. It reflects the harmonic structure of the stimulus and provides information about the integrity of the response to the stimulus.3,13,14 The FFR is an electrophysiological measure that provides an insight into the mechanisms of subcortical processing of the stimulus in the auditory system. This term was first used by Worden and Marsh (1968) to describe a phase-locked response to the frequency components of the stimulus in cats.15

A study investigated the FFR characteristics in response to speech stimuli in American and Chinese neonates and adults.16 The authors studied the evoked potentials generated by speech stimuli according to the language (Chinese and English) and age (neonates and adults). The results showed maturation of vocal pitch processing in neonates one to three days after birth, and a significant effect of language experience in the neural processing of speech stimuli in the brainstem. These findings highlight the need to better understand neural responses in the brainstem.

As mentioned before, the perception of human speech is strongly influenced by innate factors. However, the specific language environment to which children are exposed is also crucial for the perception of speech sounds. Exposure to a particular language early in life results in a reduced ability to perceive differences between the speech sounds of other languages.16

Studies have shown that the brainstem auditory evoked potential with speech stimulus also showed variations in children with language disorders. A recent study17 investigated brainstem auditory evoked potentials in response to speech stimuli in children considered to be good readers and dyslexic children. The authors reported that children with dyslexia showed greater variability in responses than the group of children who were considered good readers. The results suggested that good readers have a stable neural representation of sound, while those who show impaired reading skills have inconsistent neural responses.

It was verified that the brainstem auditory evoked potential in response to speech stimuli may vary with age. In a study comparing responses obtained from young and old individuals,18 significantly higher latencies of brainstem auditory evoked potentials for speech stimuli were observed in the elderly, compared to those obtained in younger subjects, even after considering the differences in auditory thresholds between the groups. These results are consistent with reduced neural synchrony for transient speech components in the elderly.

However, the sustained portion of brainstem potentials in response to speech stimuli did not change significantly after adjustment for hearing loss, suggesting that maturity may affect response more than auditory sensitivity and other peripheral alterations.

For this potential to be used in clinical practice it is necessary to establish standards of normality through studies in populations of different age groups, exposed to different languages and without complaints.

Therefore, the aim of the present study was to analyze the latencies and amplitudes of the waves obtained from the tracings of brainstem auditory evoked potentials by speech stimuli in Brazilian neonates with normal hearing and without auditory risk factors.

Twenty-one neonates were assessed, of whom 11 were females and 10 males, aged between 2 and 38 days of life. Only neonates without hearing risk were selected, according to the risk criteria of the Joint Committee on Infant Hearing (2007).19 All included subjects had a normal hearing screening, with normal results for transient evoked otoacoustic emissions and automated cortical evoked auditory potentials present at 35 dBHL.20 Additionally, only neonates who showed auditory evoked potentials for the click stimulus at the presentation level of 80 dBHL, with latencies within the expected for the age, were included.1 The present study was approved by the Research Ethics Committee (CEP 1243/11).

Brainstem auditory evoked potentials were assessed with speech stimulus during natural sleep, at the level of 80 dBHL. To undergo the test, the newborn was placed in the cradle or on the mother's lap. Three response samples were obtained, with the presentation of 1000 stimuli each.21 The syllable chosen for the stimulus was the synthesized syllable /da/ with a 40 milliseconds (ms) duration,22 available with the Navigator Pro-Biologic equipment, as proposed by Kraus et al.23 The sampling rate was 10.9 stimuli per second and the analysis window was 74.67ms. The stimuli were presented through an Etymotic ER-3 Insert Earphones. After cleansing the skin with Nuprep abrasive paste, the surface electrodes were fixed with Ten20 Conductive Paste for the recording of responses, with the reference electrode being positioned on the ipsilateral earlobe, the active electrode placed on the high forehead, and the ground electrode on the low forehead. The setting included one channel. The stimuli were presented in both ears, and the starting ear (right or left) was chosen randomly.

The three tracings obtained were compared to each other and were only included if a reproducibility >60% was observed, according to the equipment software. For that purpose, we used the software tool that allows overlapping waves, performing a mathematical comparison and statistically demonstrating the correlation between them. Based on this analysis, the responses of three neonates were disregarded, as they did not show reproducibility of responses in both ears, and three ears of three other neonates were also disregarded, comprising a total of 33 valid ears. The waves were analyzed as follows:

Transient portion of the response – Peaks I, III, V and A were evaluated, determining the frequency of occurrence, latency and amplitude of each peak;

Sustained portion of the response (FFR) – Valleys B, C, D, E, F and O were evaluated by determining the frequency of occurrence, latency and amplitude of each valley.

In the transient portion, the wave marking occurred as it is usually done with the click stimulus, and in the sustained portion, the marking of the valleys was performed according to the baseline, beginning with valley A that follows Wave V, as shown in Fig. 1.

Figure 1.

Example of the representation of valleys corresponding to the sustained portion (frequency-following response [FFR]) of the response to BAEP-speech.

(0.1MB).

Initially, we analyzed the occurrence of observable waves for the auditory evoked potentials for the syllable /da/ in the transient and sustained portions. In the transient portion, we observed responses in 100% of all analyzable waves (Wave I, Wave III, Wave V and Wave A). In the sustained portion, the frequency of identification of Wave B was 53.12%; of Wave C, 75%; of Wave D, 90.62%; of Wave E, 96.87%; of Wave F, 87.5%; and of Wave O, 87.5%. It was observed that all waves, except Wave B, were identifiable in >70% of subjects. However, none of the analyzed waves was present in 100% of the subjects. Table 1 shows the appearance of each of the analyzable waves for the brainstem auditory evoked potentials by speech stimulus.

The latency and amplitude analyses were initially performed by separating the values found for the right and left ears. Student's t test was applied to evaluate whether there were statistically significant differences, and no differences were found between the ears (Table 2); therefore, the analysis was performed with the pooled values. When analyzing the response latency for the transient portion waves, it was observed that, as for the click stimulus, the group of Waves I, III and V appear in the first 10ms after the stimulus. The latencies for the speech stimulus were higher than those expected for the click stimulus; however, the neonates’ responses were very similar.

For the click stimulus, we observed a mean latency of wave V appearance at 6.75ms with a standard deviation of 0.23ms, while for speech stimulus we obtained a mean latency of wave V appearance at 7.67ms with a standard deviation of 0.35ms. The statistical parameters of latencies and amplitudes for each analyzed wave with the grouped ears are shown in Table 3 and in Figs. 2–5.

Figure 2.

Amplitude of the waves from the transient portion to the BAEP-speech in neonates.

(0.11MB).

Figure 3.

Latency of the waves from the transient portion to the BAEP-speech in neonates.

(0.05MB).

Figure 4.

Amplitude of the waves from the sustained portion to the BAEP-speech in neonates.

(0.1MB).

Figure 5.

Latency of the waves from the sustained portion to the BAEP-speech in neonates.

(0.09MB).

The analysis of latencies and amplitudes showed greater variation in response amplitude values than in latency. Tables 2 and 3 and Figs. 2–5 show these differences. A greater similarity was observed in the latencies of responses and greater amplitude variation in the studied group.

The summary of the characteristics of the studied population and the latency results observed for each wave are shown in Table 4.

It was possible to observe waves with typical morphology for brainstem evoked potentials for speech stimulus in neonates. The latencies of waves I, III and V for the speech stimulus were longer than those generated by a click stimulus.

The four waves that describe the transient portion (up to 10ms) were detected in all neonates.

The waves of the sustained portion showed rates of occurrence >70% for all waves, except for Wave B that was present in 53.12% of the neonates; the initial latency of the sustained response ranged from 10 to 52.16ms.

The response to speech stimuli in neonates is evidence for the “biological capacity model,” indicating that neonates are born with similar innate abilities of pitch coding at the subcortical level.

The authors declare no conflicts of interest.