Brainstem encoding of speech in normal-hearing individuals with absent acoustic reflex

The acoustic reflex test is an important tool for identifying auditory disorder from the middle ear to the superior olivary complex. Absence of acoustic reflexes is the early sign of many auditory disorders. Absence of acoustic reflex with normal hearing sensitivity may be an early sign of auditory neuropathy with poor encoding of speech at initial stage. Speech auditory brainstem response was recorded with /da/ (40 ms) stimuli in two groups of patients. The control group contained normal-hearing participants with presence of acoustic reflex, whereas the experimental group contained normal-hearing participants with absent acoustic reflexes. The peak latency, amplitude, and F0 and F1 mean amplitude were analyzed in both groups. MANOVA showed no significant difference in any parameter between the control and experimental group. Results of the current study showed that absence of acoustic reflexes in normal-hearing patients without auditory complaint is not sufficient by itself to diagnose the existence of auditory neuropathy. This study also highlighted that normal-hearing patients with absence of acoustic reflex have similar brainstem encoding of speech as that of patients with acoustic reflex.


Background
Loud sounds cause involuntary contraction of the stapedius muscle, called acoustic refl ex. It involves either one or both middle ear muscles. It can be recorded ipsilaterally and contralaterally. If it is being recorded ipsilaterally, then recording should be done on the same side of the stimulus. In contralateral recording, the recording should be done on the opposite side of the sound stimulus [1]. Th e acoustic refl ex is usually measured at 500, 1000, 2000, and 4000 Hz. At present, acoustic refl ex is one of the important tests in audiology, as it helps in the assessment of aff erent (sensorial) and eff erent (motor) systems. Th ese are the part of acoustic refl ex pathway.
Recording of the acoustic refl ex allows the professional to assess the middle ear up to the superior olivary complex. Th e acoustic refl ex has many functions, such as enhancement of localization ability and sense of sound direction with the help of binaural interaction; it helps in improving auditory attention for continuous sounds; it improves speech intelligibility by separating hearing signals from background noise; it helps to know the variation in intensity level above threshold; it helps in reducing noise produced during chewing and movement of the mandible during speech; it participates in vocalization; and it helps in understanding speech through frequency selectivity and improves speech discrimination at higher intensities [2][3][4][5][6][7].
Th e normal value of the acoustic refl ex threshold varies from 70 to 90 dBHL [8][9]. In clinical practice, if the acoustic refl ex is absent even above 110 dBHL, it indicates an absence of the acoustic refl ex [10]. In clinical practice we see patients without auditory complaints and with hearing threshold within the normal limit correlating with speech audiometry showing absent acoustic refl ex responses at values above those expected on an A-type tympanogram. A study by Pinotti et al. [12] showed that 10% of normal-hearing individuals had absent acoustic refl ex, without any ear complaints. Th ese fi ndings wherein absence of acoustic refl ex occurs alone, with normal pure tone audiometry and speech audiometry, normal tympanometry, and presence of otoacoustic emission within normal patterns have aroused interest in research in these areas. Absence of AR under the above-mentioned conditions may be an indication of auditory neuropathy at initial stage. Disorder in the auditory nerve in auditory neuropathy may be the reason for absent acoustic refl ex. In subjects with auditory neuropathies, because of lack of synchronous fi ring, the auditory nerve may not be able to process sound, which is required for the Brainstem encoding of speech in normal-hearing individuals with absent acoustic re ex Rajkishor Mishra a , Himanshu Kumar Sanju c , Preeti Sahu b The acoustic re ex test is an important tool for identifying auditory disorder from the middle ear to the superior olivary complex. Absence of acoustic re exes is the early sign of many auditory disorders. Absence of acoustic re ex with normal hearing sensitivity may be an early sign of auditory neuropathy with poor encoding of speech at initial stage. Speech auditory brainstem response was recorded with /da/ (40 ms) stimuli in two groups of patients. The control group contained normal-hearing participants with presence of acoustic re ex, whereas the experimental group contained normal-hearing participants with absent acoustic re exes. The peak latency, amplitude, and F0 and F1 mean amplitude were analyzed in both groups. MANOVA showed no signi cant difference in any parameter between the control and experimental group. Results of the current study showed that absence of acoustic re exes in normal-hearing patients without auditory complaint is not suf cient by itself to diagnose the existence of auditory neuropathy. This study also highlighted that normal-hearing patients with absence of acoustic re ex have similar brainstem encoding of speech as that of patients with acoustic re ex. activation of middle ear muscles to activate the acoustic refl ex [11]. One of the most important objective tests that help in diagnosing auditory neuropathy is auditory brainstem response (ABR) audiometry. Several studies have been conducted on such cases using ABR, which suggest the presence of an abnormal synchronization or absence of peak in the waveform. Th e study by Pinotti et al. [12] revealed that ABR audiometry using click stimuli in seven patients with no auditory complaints, normal hearing, and A-type tympanogram with absence of acoustic refl ex showed the presence of waves I, III, and V, with absolute latency values and interpeak latencies within the normal standard, which suggests that absence of stapedial refl exes alone in normal-hearing patients, without any auditory complaint, is not suffi cient by itself to diagnose the existence of auditory neuropathy. Th e studies revealed poor processing to speech sound in case of auditory neuropathy [13][14][15][16]. Speech ABR in patients with no auditory complaints, with normal hearing, and with an A-type tympanogram with the absence of acoustic refl ex has not been documented yet. Such a study would give insight into the processing capability of speech sound in patients with auditory neuropathy. Th ere is a need to study brainstem encoding of speech sound in individuals with absent acoustic refl ex with normal pure tone audiometry correlating well with speech audiometry, as well as normal tympanometry and presence of optoacoustic emission.

Participants
Th e study consisted of 24 age-matched individuals (48 ears) aged 18-26 years who were divided into two groups. Group I comprised 12 healthy adults (24 ears) (seven males and fi ve females) having A-type tympanogram with presence of refl ex, and group II consisted of 12 adults (24 ears) (seven males and fi ve females) having A-type tympanogram with absence of acoustic refl ex.

Inclusion criteria
Patients aged 18-26 years with no otological and auditory complaints, normal hearing sensitivity below 15 dBHL for octaves from 250 to 8000 Hz, correlated SRT with PTA, and presence of otoacoustic emission were eligible for the study. Th e above audiological testing was carried out by experienced audiologists. All participants had normal otological and otoscopic examination fi ndings, with no middle ear pathology, evaluated by an experienced otolaryngologist. All had ' A' type tympanogram but the only diff erence between group I and group II was the presence of ipsilateral and contralateral refl exes in both ears in group I, which was absent in group II. All participants were well informed about the study and written consent was taken before the study.

Instrumentation
An AC-40 Interacoustics (Drejervaenget 8, DK-5610 Assens) dual-channel, clinical audiometer was used for measuring pure tone thresholds. Live speech and tones were presented through earphones (Telephonics, TDH Hz. ECHO Scan screening MAICO OAE instrument (MAICO Diagnostics, 10393 West 70St, Eden Prairie, MN 55344) was used for Distortion product evoked otoacoustic emissions elicited using an 80 dB SPL peak equivalent (peak) level click, present in fi ve successive trials, where L1-65 dB and L2-55 dB was set and a signalto-noise ratio of at least 6 dB with a reproducibility score of at least 70% , in frequency band of 2 to 5 kHz. Biologic Navigator Pro was used for recording speech ABR.

Procedure
Pure tone thresholds were obtained using the modifi ed version of the Hughson and Westlake procedure to establish normal hearing sensitivity. Air conduction thresholds were obtained for octave frequencies between 250 Hz and 8 kHz, whereas bone conduction thresholds were measured at octave frequencies between 250 Hz and 4 kHz. SRT and SIS were obtained using spondee words and phonetically balanced words, respectively. Th ese tests were administrated in participants' native language using live monitored speech. To rule out any middle ear pathology, immittance audiometry and refl exometry were carried out. Immittance audiometry was performed using a 226 Hz probe tone, whereas refl exometry included obtaining ipsilateral and contralateral acoustic refl exes at 500, 1000, 2000, and 4000 Hz with the same 226 Hz probe tone. Biologic Navigator pro (version 7) (Natus Medical Incorporated, One Bio-logic Plaza, Mundelein, IL 60060) was used to record click-evoked ABR to check the integrity of the neural pathway at the levels of the brainstem before measuring the speech-evoked ABR. Speech-evoked ABR was recorded for all participants with speech stimuli /da/ of 40 ms duration produced using a KLATT synthesizer (Klatt, 1980) [17]. Silver chloride electrodes were used to record the responses. Recording was done monaurally and ipsilaterally with electrodes at the vertex (noninverting), at the ipsilateral mastoid (Inverting), and at the contralateral mastoid (ground). Th e absolute electrode impedance and interelectrode impedance were maintained below 5 and 2 kΩ, respectively. At least two recordings of 3000 sweeps to rarefaction polarity at a rate of 10.9/s were collected. Th e responses were amplifi ed 100 000 times and weighted. Th e sum of the two recordings was taken for analysis. Time window of 64 ms, including 10 ms prestimulus time, was used. Th e responses were band pass-fi ltered online between 100 and 3000 Hz. Th e stimuli were presented through Etymotic ER-3A (Etymotic Research, Inc. 61 Martin Lane Elk Grove Village, IL 60007) earphones and the intensity level was 80 dB SPL (Table 1).

Test environment
All audiological evaluations were carried out in sound-treated rooms (ANSI S3. 1,1999). Pure tone audiometry and speech audiometry were carried out in a two-room audiological setup, whereas immittance audiometry and click-evoked and speech-evoked ABR were carried out in a single-suit audiological setup.

Stimulus and recording parameters
Th e speech stimulus /da/ was used to record speechevoked ABR, which is a synthesized speech syllable of 40 ms produced using the KLATT synthesizer. Th is stimulus has fast temporal information and broad spectral characteristic of stop consonants. Th is stimulus also has spectrally reaching formant transition between vowel and consonants. Th e voicing for this stimulus begins at 5 ms, with onset noise burst during the fi rst 10 ms. Th e fundamental frequency of /da/ stimulus is from 103 to 125 Hz, increasing in a linear manner. Th e fi rst formant of /da/ (F1) increases from 220 to 720 Hz, whereas the second formant (F2) of /da/ decreases from 1700 to 1240 Hz over the duration of the stimulus. Th e third formant (F3) falls slightly from 2580 to 2500 Hz, whereas the fourth (F4) and fi fth formants (F5) remain the same at 3600 and 4500 Hz, respectively. Figure 1 shows the time domain waveform and Fig. 2 shows the spectral waveform of /da/ stimulus used in the present study.

Measurement of fundamental frequency (F0) and rst formant frequency (F1)
Frequency following response has energy mainly at its harmonics and at fundamental frequency. We performed Fast Fourier analysis of waveform from 11.5 to 46.5 ms. We measured activity occurring between 103 and 121 Hz, which corresponds to fundamental frequency (F0) of speech, and between 220 and 279 Hz, which corresponds to fi rst formant frequency (F1). Th e subject response should be above noise fl oor to enable the analysis. Th is was ascertained by comparing the spectral magnitude from prestimulus condition to poststimulus condition -that is, response. If the magnitude of the fundamental frequency (F0) or fi rst formant frequency (F1) was greater than or equal to 1, response was considered to be present.
Time domain waveform of /da/ stimulus.

Results
Th e present study aimed to evaluate and compare the speech-evoked brainstem responses between groups I and II. Mean, SD, and signifi cance values of the data in each group were calculated. Th e data obtained were subjected to statistical analysis using SPSS software (Version 17, Chicago, USA). Independent t-test was applied to fi nd out the statistical diff erence, if any.
Th e response measures that were considered for analysis were V, Measurement of amplitude of F0 and F1 indicated that the mean amplitude of F0 was greater than the mean amplitude of F1 (

Discussion
Results of the current study showed no signifi cant diff erence between control and experimental groups in terms of the latency and amplitude measures of speech ABR. Th e result of the current study showed that there was no signifi cant diff erence in the amplitude Latency of different peaks of speech auditory brainstem response in the control (AR present) and experimental (AR absent) group.

Graph 1
Graph of F0 and F1 mean amplitude of the control (AR present) and experimental (AR absent) group.  of fundamental frequency (F0) between the control and experimental group. Similar fi ndings were seen for amplitude of fi rst format frequency (F1) also. Th ere was no signifi cant diff erence in the coding of F1 between control and experimental groups. Result of the current study also showed that normal-hearing subjects with absence of acoustic refl ex have similar brainstem encoding of speech as that of patients with acoustic refl ex. In a study by Berlin et al. [18] of 136 patients with auditory neuropathy, none showed normal refl exes at any frequency. Refl exes were either absent or observed above 100 dBHL, which was not congruent with their normal optoacoustic emissions.

Graph 2
Pinotti et al. [12] also reported that ABR audiometry using click stimuli in seven patients with no auditory complaints, normal hearing, and an A-type tympanogram with the absence of acoustic refl ex showed presence of waves I, III, and V, with absolute latency values and interpeak latencies within the normal standards, which suggests that the absence of stapedial refl exes in normal-hearing patients without auditory complaint is not suffi cient by itself to diagnose the existence of auditory neuropathy.
Th e fi nding of the current study also suggests that normal-hearing subjects with absence of acoustic refl ex, without any auditory complaint, have similar encoding of speech at the brainstem level, and absence of acoustic refl ex in normal-hearing patients without auditory complaint is not suffi cient by itself to diagnose the existence of auditory neuropathy.

Conclusion
Th e present study showed that brainstem encoding of speech is similar in patients with absent acoustic refl ex without any auditory complaint when compared with normal-hearing subjects with acoustic refl ex.
Sample waveform of the control group (acoustic re ex present). Waveform of speech auditory brainstem response of the two groups.

Figure 3
Sample waveform of the experimental group (acoustic re ex absent). Waveform of speech auditory brainstem response of the two groups.