Predicting Hearing Thresholds in Occupational Noise-Induced Hearing Loss by Auditory Steady State Responses Joseph Attias,1,2 Hanin Karawani,1 Rafi Shemesh,1 and Ben Nageris3 ASSR test is not, by itself, an appropriate tool for hearing screening in the general population.
Objectives: Currently available behavioral tools for the assessment of noise-induced hearing loss (NIHL) depend on the reliable cooperation of the subject. Furthermore, in workers’ compensation cases, there is considerable financial gain to be had from exaggerating symptoms, such that accurate assessment of true hearing threshold levels is essential. An alternative objective physiologic tool for assessing NIHL is the auditory steady state response (ASSR) test, which combines frequency specificity with a high level of auditory stimulation, making it applicable for the evaluation of subjects with a moderate to severe deficit. The primary aim of the study was to assess the value of the multifrequency ASSR test in predicting the behavioral warble-tone audiogram in a large sample of young subjects with NIHL of varying severity or with normal hearing. The secondary goal was to assess suprathreshold ASSR growth functions in these two groups.
Key words: Auditory steady state response, Noise-induced hearing loss, Specificity and sensitivity, Suprathreshold. (Ear & Hearing 2014;35;330–338)
INTRODUCTION Occupational noise-induced hearing loss (NIHL) is considered one of the most prevalent work-related health hazards, with significant medical and economic ramifications. It has long been recognized that exposure to occupational noise is associated with a characteristic audiometric pattern of selective high-frequency, bilateral, and symmetrical loss known as the “boilermakers’ notch.” The noticeable dip appears in the audiogram at 4000 Hz after a few years of noise exposure and may deteriorate to high-frequency hearing loss after decades of noise exposure (Taylor et al. 1965). With continued chronic exposure, the loss may extend to the low frequencies (Riley et al. 1961) including the frequencies encompassing the speech spectrum. Early diagnosis of NIHL with appropriate intervention can reduce the risk of deterioration. At present, the diagnosis of NIHL is based primarily on behavioral subjective measures that require the complete and continuous cooperation of the subject. When cooperation is limited, an accurate objective tool is needed to assess the true hearing threshold with frequency specificity. Such a tool would also be beneficial in cases of workers’ compensation for NIHL in which there is considerable financial gain to be had from exaggerating symptoms of hearing loss. Estimated rates of exaggerated hearing threshold levels among workers with NIHL range from 16 to 30% (Alberti et al. 1987; Barrs et al. 1994; Rickards & De Vidi 1995; Hone et al. 2003). The auditory steady state response (ASSR) test may serve as a good alternative tool for objectively predicting hearing thresholds in infants, young children, and individuals who are unable or unwilling to cooperate during conventional behavioral testing. Responses are elicited by continuous tones modulated in amplitude or frequency rather than by transient stimuli (Kuwada et al. 1986; Rickards et al. 1994), thereby combining frequency specificity and a high level of auditory stimulation for testing moderate to severe hearing losses. ASSRs can be evoked by stimuli modulated at rates in the 30 to 50 Hz range (“40 Hz ASSR”) as well as in the 70 to 110 Hz range (“80 Hz ASSR”). Although 80 Hz ASSRs are two to five times smaller than 40 Hz ASSRs (Herdman et al. 2002), they are preferred for threshold estimations in sleeping or sedated infants and adults because they are little affected by the state of arousal (Rickards et al. 1994). In addition, because ASSRs are analyzed in the frequency domain,
Design: The study group included 157 subjects regularly exposed to high levels of occupational noise, who attended a university-associated audiological clinic for evaluation of NIHL from 2009 through 2011. All underwent a behavioral audiogram, and on the basis of the findings, were divided into those with NIHL (108 subjects, 216 ears) or normal hearing (49 subjects, 98 ears). The accuracy of the ASSR threshold estimations for frequencies of 500, 1000, 2000, and 4000 Hz was compared between groups, and the specificity and sensitivity of the ASSR test in differentiating ears with or without NIHL was calculated using receiver operating characteristic analysis. Linear regression analysis was used to formulate an equation to predict the behavioral warble-tone audiogram at each test frequency using ASSR thresholds. Multifrequency ASSR amplitude growth as a function of stimulus intensity was compared between the NIHL and normal-hearing groups for 1000 Hz and 4000 Hz carrier frequencies. Results: In the subjects with NIHL, ASSR thresholds to various frequencies were significantly and highly correlated with the behavioral warbletone thresholds; Pearson correlation coefficients ranged from 0.6 to 0.8 over the four frequencies tested. Differences between thresholds ranged from 10 to 13 dB. The configuration of the ASSR waveforms closely approximated the behavioral audiogram. The sensitivity for screening hearing thresholds was 92%; by frequency, sensitivity ranged between 92.7 and 98.4%, but specificity was lower, especially at the low frequencies. ASSR accurately predicted moderate and severe NIHL. The mean ASSR growth amplitude to increasing stimulus level for 1000 and 4000 Hz was significantly steeper in the NIHL than in the normal-hearing group, with no significant difference between frequencies. Conclusions: The ASSR test has a high sensitivity to detect moderate to severe hearing loss in subjects with NIHL. Its use can facilitate the early identification of noise-exposed workers with NIHL. It may also serve an important medico-legal function in cases of workers’ compensation. The Department of Communication Sciences & Disorders, University of Haifa, Haifa, Israel; 2Institute for Audiology and Clinical Neurophysiology, Schneider Children’s Medical Center of Israel, Petach Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; and 3Department of Otolaryngology, Head and Neck Surgery, Rabin Medical Center, Petach Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. 1
0196/0202/13/353-0330/0 • Ear & Hearing • Copyright © 2014 by Lippincott Williams & Wilkins • Printed in the U.S.A. 330
ATTIAS ET AL. / EAR & HEARING, VOL. 35, NO. 3, 330–338
statistical tests can be used for the objective determination of the absence or presence of responses rather than visual interpretation of the waveforms (Picton et al. 1987). Another advantage is the significant reduction in test duration when multiple modulating tones are presented to both ears simultaneously in patients with sensorineural hearing loss (SNHL), with no loss of the test’s predictive efficacy (Picton et al. 2009). Studies have reported a difference of 10 to 20 dB between the physiologic (ASSR) and behavioral (pure-tone) thresholds under these conditions (Picton et al. 2003). The correlations have been found to be better at higher frequencies and in patients with moderate to severe hearing loss (Lins et al. 1996). Sloping losses have prompted concerns that ASSR thresholds obtained with higher-frequency stimuli may be affected by the responses of neighboring low-frequency regions, leading to underestimation of the behavioral threshold at the test frequency. In addition, noise exposure may affect the accuracy of the ASSR in predicting both normal and impaired auditory behavioral thresholds. However, the ASSR has proved very useful for evaluating patients with high-frequency SNHL and candidates for cochlear implants (Attias et al. 2006). To date, research in subjects with NIHL also indicates high reliability of the ASSR test and a high power of prediction of the behavioral audiograms (Hsu et al. 2003, 2010), with differences between thresholds ranging from 20 dB at 500 Hz to 10 dB at 4000 Hz (Hsu et al. 2010). Similar findings were noted for multifrequency ASSRs (Hsu et al. 2011). Previous research has shown a difference in growth function between subjects with normal hearing (Lins & Picton 1995) and SNHL (Lenarz et al. 1986; Rodriguez et al. 1986; Dimitrijevic et al. 2002). Dimitrijevic et al. (2002) reported that the mean amplitude of ASSRs as a function of sensation level was larger in subjects with SNHL than controls, with a trend for steeper amplitude growth functions in the SNHL group. However, comparison of these studies is limited owing to differences in stimuli, recording, and analysis conditions as well as the small sample sizes. In addition, no growth ASSR functions were reported for noise-exposed subjects. The primary aim of the present study was to extend previous investigations and create a more suitable match with which to compare ASSR thresholds. By including a large sample size (314 ears), we were able to calculate the sensitivity and specificity of the ASSR in the assessment of NIHL. To exclude agerelated hearing loss, the sample consisted of only participants less than 56 years old. The secondary aim of the study was to evaluate the ASSR growth function in a large group of subjects with NIHL or normal hearing.
SUBJECTS and METHODS Subjects The study was conducted in 157 subjects (total 314 ears), aged 25 to 55 years (mean 23.5, standard deviation [SD] 1.67), regularly exposed to high levels of occupational noise. The subjects were randomly selected from 1200 workers who attended the Audiological Clinical Center of the University of Haifa in 2009–2011 as part of their periodic evaluation for NIHL. All were enrolled in the annual national hearing-loss prevention program at the recommendation of their occupational doctors. Participants in the study provided written informed consent in accordance with the guidelines of the Institutional Review
331
Board of the University of Haifa, Israel. Exclusion criteria were presence of an ear infection by detailed history or otoscopic examination, mixed or conductive hearing loss, and fluctuating, rapidly progressive, or retrocochlear loss by history, and auditory brainstem response test. Normal middle ear function was confirmed in all subjects on the day of testing by tympanometry or otoacoustic emissions or both. All subjects underwent behavioral standard (steady state) pure-tone assessment, and on the basis of the results, were divided into two groups: normal hearing (control) and occupational NIHL. The normal-hearing group (n=98) was defined as having a threshold up to 20 dB HL for 250 Hz to 8000 Hz (ANSI 1989). The NIHL group included only patients presenting with an audiogram of bilateral and symmetrical cochlear hearing loss of 25 dB HL or more at multiple frequencies in the range of 250 to 4000 Hz (n=126) or subjects showing a bilateral notch at 3000 to 6000 Hz (n=90). All audiological tests were carried out by skilled audiologists at the clinic.
Behavioral Warble-Tone Threshold For the purpose of this study, the behavioral tone thresholds of each subject were retested in a double-walled, sound-attenuated room with calibrated AC-40 audiometers (Interacoustic A/S, Assens, Denmark). The stimulus consisted of a warble tone that deviated ±5% in frequency, delivered at a rate of 5 Hz. This type of stimulus was used to have a more suitable match to the ASSR stimuli (warble-tone threshold). Air conduction thresholds were measured in each ear by using insert earphones (EARTONE® 5A; Auditory Systems, Indianapolis, IN) at 250 to 8000 Hz; bone conduction thresholds were measured using a Radioear vibrator B-71 at 250 to 4000 Hz. Threshold levels were determined with a 10dB-down/5dB-up approach (Carhart & Jerger 1959). Clinical masking was used if necessary or warranted by the clinical audiologic data. The findings were included in the study only if the behavioral audiograms were considered reliable by the audiologist.
Auditory Steady State Responses ASSRs were assessed in a sound-attenuated room with the subjects in the supine position. (They were told to relax, and most of them slept through the test.) The Bio-logic MASTER Version 2.02 (Biologic System Corp., Mundelein, IL) was used in all cases. Potentials were collected from scalp electrodes located at Cz (active) and referenced to the midline posterior neck (about 7 cm below the inion), with the right earlobe as ground. Electrode impedance was 5 kohm or less at 20 Hz for each electrode, and the difference between the electrodes was 3 kohm or less. Carrier frequencies were matched to the audiometric frequencies; 500, 1000, 2000, and 4000 Hz. In some cases, the behavioral threshold was sufficiently low to allow for ASSR testing at 6000 Hz. The modulation frequencies (amplitude modulation [AM]/frequency modulation [FM]) for those audiometric frequencies were 82, 84, 87, and 89 Hz for the left ear, respectively, and 91, 94, 96, and 98 Hz for the right ear. These high modulation rates were used because they are less affected by the state of subject’s arousal, and because higher-frequency modulators tend to show less of an interaction between carriers (Picton et al. 2009) and exhibit lower growth function slopes (Lins & Picton 1995). Each auditory stimulus consisted of a sinusoidal tone with a carrier frequency that was 25% frequency-modulated and 100% amplitude-modulated.
332
ATTIAS ET AL. / EAR & HEARING, VOL. 35, NO. 3, 330–338
A relative phase of 270 degrees between the AM and FM components was chosen so as to elicit the largest combined response (John et al. 2002). All stimuli were routed through an Amplaid audiometer (Amplaid 319, Amplifon, Via Donizetti, Milan, Italy) and presented to both ears simultaneously through Bio-logic insert earphones (580-SINSER; Biologic System Corp.) with foam earplugs. The ASSR stimuli delivered through the insert phones were calibrated at dB HL in the MASTER setup by using the reference values reported by Wilber et al. (1988). The authors had performed data analysis of five different studies that used puretone hearing threshold measurements of normal listeners at 125 to 8000 Hz frequency range and calibration by 2 cm3 coupler. In the present study, stimulus intensity was also measured with the Bruel & Kjaer (Naerum, Denmark) 2613 amplifier and a 2120 frequency processor coupled to a 4153 microphone. All measured intensities were well correlated (±2 dB) to the normative dB SPL calibration data listed in the MASTER software. In the dichotic multiple frequency test, the initial stimulation level was 80 dB HL. If clear responses were recorded, the level was reduced in 20 dB increments until no responses were recorded, and then increased in 10 dB increments to determine the threshold at which a significant response was elicited. When there was no response to a level of 80 dB HL in the multicarrier condition, a single specific modulated tone of up to 132 dB SPL was presented. This minimized the possibility of undesired interactions between the carriers at levels above 80 dB HL.
ASSR Recordings A sampling rate of 1200 Hz at 16-bit resolution was used. The EEG responses were amplified using a gain of 10,000 at a filter band pass of 3 to 300 Hz (12 dB/octave). The data were recorded in epochs containing 1024 points; one response epoch lasted 0.853 seconds. Sixteen data epochs were collected and linked together to form one sweep with an overall duration of 13.653 seconds. Epochs that contained electrophysiological activity exceeding ±20µV were rejected, and the next acceptable epoch was used to build the sweep. Once completed, each sweep was averaged in the time domain and subsequently analyzed by fast Fourier transform to yield an amplitude spectrum with a resolution of 0.0732 Hz. To determine whether the response at the modulation frequency was significantly different from the background EEG activity in neighboring frequencies, we used the F ratio to compare the amplitude at the modulation frequency with the average amplitude of the noise in 120 adjacent frequency bins (60 bins above and 60 below the modulation frequency, i.e., ±4.4 Hz). A response was accepted as present if the F ratio, calculated against the critical values for F at 2 df and 240 df, was significant at p
Objectives: Currently available behavioral tools for the assessment of noise-induced hearing loss (NIHL) depend on the reliable cooperation of the subject. Furthermore, in workers’ compensation cases, there is considerable financial gain to be had from exaggerating symptoms, such that accurate assessment of true hearing threshold levels is essential. An alternative objective physiologic tool for assessing NIHL is the auditory steady state response (ASSR) test, which combines frequency specificity with a high level of auditory stimulation, making it applicable for the evaluation of subjects with a moderate to severe deficit. The primary aim of the study was to assess the value of the multifrequency ASSR test in predicting the behavioral warble-tone audiogram in a large sample of young subjects with NIHL of varying severity or with normal hearing. The secondary goal was to assess suprathreshold ASSR growth functions in these two groups.
Key words: Auditory steady state response, Noise-induced hearing loss, Specificity and sensitivity, Suprathreshold. (Ear & Hearing 2014;35;330–338)
INTRODUCTION Occupational noise-induced hearing loss (NIHL) is considered one of the most prevalent work-related health hazards, with significant medical and economic ramifications. It has long been recognized that exposure to occupational noise is associated with a characteristic audiometric pattern of selective high-frequency, bilateral, and symmetrical loss known as the “boilermakers’ notch.” The noticeable dip appears in the audiogram at 4000 Hz after a few years of noise exposure and may deteriorate to high-frequency hearing loss after decades of noise exposure (Taylor et al. 1965). With continued chronic exposure, the loss may extend to the low frequencies (Riley et al. 1961) including the frequencies encompassing the speech spectrum. Early diagnosis of NIHL with appropriate intervention can reduce the risk of deterioration. At present, the diagnosis of NIHL is based primarily on behavioral subjective measures that require the complete and continuous cooperation of the subject. When cooperation is limited, an accurate objective tool is needed to assess the true hearing threshold with frequency specificity. Such a tool would also be beneficial in cases of workers’ compensation for NIHL in which there is considerable financial gain to be had from exaggerating symptoms of hearing loss. Estimated rates of exaggerated hearing threshold levels among workers with NIHL range from 16 to 30% (Alberti et al. 1987; Barrs et al. 1994; Rickards & De Vidi 1995; Hone et al. 2003). The auditory steady state response (ASSR) test may serve as a good alternative tool for objectively predicting hearing thresholds in infants, young children, and individuals who are unable or unwilling to cooperate during conventional behavioral testing. Responses are elicited by continuous tones modulated in amplitude or frequency rather than by transient stimuli (Kuwada et al. 1986; Rickards et al. 1994), thereby combining frequency specificity and a high level of auditory stimulation for testing moderate to severe hearing losses. ASSRs can be evoked by stimuli modulated at rates in the 30 to 50 Hz range (“40 Hz ASSR”) as well as in the 70 to 110 Hz range (“80 Hz ASSR”). Although 80 Hz ASSRs are two to five times smaller than 40 Hz ASSRs (Herdman et al. 2002), they are preferred for threshold estimations in sleeping or sedated infants and adults because they are little affected by the state of arousal (Rickards et al. 1994). In addition, because ASSRs are analyzed in the frequency domain,
Design: The study group included 157 subjects regularly exposed to high levels of occupational noise, who attended a university-associated audiological clinic for evaluation of NIHL from 2009 through 2011. All underwent a behavioral audiogram, and on the basis of the findings, were divided into those with NIHL (108 subjects, 216 ears) or normal hearing (49 subjects, 98 ears). The accuracy of the ASSR threshold estimations for frequencies of 500, 1000, 2000, and 4000 Hz was compared between groups, and the specificity and sensitivity of the ASSR test in differentiating ears with or without NIHL was calculated using receiver operating characteristic analysis. Linear regression analysis was used to formulate an equation to predict the behavioral warble-tone audiogram at each test frequency using ASSR thresholds. Multifrequency ASSR amplitude growth as a function of stimulus intensity was compared between the NIHL and normal-hearing groups for 1000 Hz and 4000 Hz carrier frequencies. Results: In the subjects with NIHL, ASSR thresholds to various frequencies were significantly and highly correlated with the behavioral warbletone thresholds; Pearson correlation coefficients ranged from 0.6 to 0.8 over the four frequencies tested. Differences between thresholds ranged from 10 to 13 dB. The configuration of the ASSR waveforms closely approximated the behavioral audiogram. The sensitivity for screening hearing thresholds was 92%; by frequency, sensitivity ranged between 92.7 and 98.4%, but specificity was lower, especially at the low frequencies. ASSR accurately predicted moderate and severe NIHL. The mean ASSR growth amplitude to increasing stimulus level for 1000 and 4000 Hz was significantly steeper in the NIHL than in the normal-hearing group, with no significant difference between frequencies. Conclusions: The ASSR test has a high sensitivity to detect moderate to severe hearing loss in subjects with NIHL. Its use can facilitate the early identification of noise-exposed workers with NIHL. It may also serve an important medico-legal function in cases of workers’ compensation. The Department of Communication Sciences & Disorders, University of Haifa, Haifa, Israel; 2Institute for Audiology and Clinical Neurophysiology, Schneider Children’s Medical Center of Israel, Petach Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; and 3Department of Otolaryngology, Head and Neck Surgery, Rabin Medical Center, Petach Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. 1
0196/0202/13/353-0330/0 • Ear & Hearing • Copyright © 2014 by Lippincott Williams & Wilkins • Printed in the U.S.A. 330
ATTIAS ET AL. / EAR & HEARING, VOL. 35, NO. 3, 330–338
statistical tests can be used for the objective determination of the absence or presence of responses rather than visual interpretation of the waveforms (Picton et al. 1987). Another advantage is the significant reduction in test duration when multiple modulating tones are presented to both ears simultaneously in patients with sensorineural hearing loss (SNHL), with no loss of the test’s predictive efficacy (Picton et al. 2009). Studies have reported a difference of 10 to 20 dB between the physiologic (ASSR) and behavioral (pure-tone) thresholds under these conditions (Picton et al. 2003). The correlations have been found to be better at higher frequencies and in patients with moderate to severe hearing loss (Lins et al. 1996). Sloping losses have prompted concerns that ASSR thresholds obtained with higher-frequency stimuli may be affected by the responses of neighboring low-frequency regions, leading to underestimation of the behavioral threshold at the test frequency. In addition, noise exposure may affect the accuracy of the ASSR in predicting both normal and impaired auditory behavioral thresholds. However, the ASSR has proved very useful for evaluating patients with high-frequency SNHL and candidates for cochlear implants (Attias et al. 2006). To date, research in subjects with NIHL also indicates high reliability of the ASSR test and a high power of prediction of the behavioral audiograms (Hsu et al. 2003, 2010), with differences between thresholds ranging from 20 dB at 500 Hz to 10 dB at 4000 Hz (Hsu et al. 2010). Similar findings were noted for multifrequency ASSRs (Hsu et al. 2011). Previous research has shown a difference in growth function between subjects with normal hearing (Lins & Picton 1995) and SNHL (Lenarz et al. 1986; Rodriguez et al. 1986; Dimitrijevic et al. 2002). Dimitrijevic et al. (2002) reported that the mean amplitude of ASSRs as a function of sensation level was larger in subjects with SNHL than controls, with a trend for steeper amplitude growth functions in the SNHL group. However, comparison of these studies is limited owing to differences in stimuli, recording, and analysis conditions as well as the small sample sizes. In addition, no growth ASSR functions were reported for noise-exposed subjects. The primary aim of the present study was to extend previous investigations and create a more suitable match with which to compare ASSR thresholds. By including a large sample size (314 ears), we were able to calculate the sensitivity and specificity of the ASSR in the assessment of NIHL. To exclude agerelated hearing loss, the sample consisted of only participants less than 56 years old. The secondary aim of the study was to evaluate the ASSR growth function in a large group of subjects with NIHL or normal hearing.
SUBJECTS and METHODS Subjects The study was conducted in 157 subjects (total 314 ears), aged 25 to 55 years (mean 23.5, standard deviation [SD] 1.67), regularly exposed to high levels of occupational noise. The subjects were randomly selected from 1200 workers who attended the Audiological Clinical Center of the University of Haifa in 2009–2011 as part of their periodic evaluation for NIHL. All were enrolled in the annual national hearing-loss prevention program at the recommendation of their occupational doctors. Participants in the study provided written informed consent in accordance with the guidelines of the Institutional Review
331
Board of the University of Haifa, Israel. Exclusion criteria were presence of an ear infection by detailed history or otoscopic examination, mixed or conductive hearing loss, and fluctuating, rapidly progressive, or retrocochlear loss by history, and auditory brainstem response test. Normal middle ear function was confirmed in all subjects on the day of testing by tympanometry or otoacoustic emissions or both. All subjects underwent behavioral standard (steady state) pure-tone assessment, and on the basis of the results, were divided into two groups: normal hearing (control) and occupational NIHL. The normal-hearing group (n=98) was defined as having a threshold up to 20 dB HL for 250 Hz to 8000 Hz (ANSI 1989). The NIHL group included only patients presenting with an audiogram of bilateral and symmetrical cochlear hearing loss of 25 dB HL or more at multiple frequencies in the range of 250 to 4000 Hz (n=126) or subjects showing a bilateral notch at 3000 to 6000 Hz (n=90). All audiological tests were carried out by skilled audiologists at the clinic.
Behavioral Warble-Tone Threshold For the purpose of this study, the behavioral tone thresholds of each subject were retested in a double-walled, sound-attenuated room with calibrated AC-40 audiometers (Interacoustic A/S, Assens, Denmark). The stimulus consisted of a warble tone that deviated ±5% in frequency, delivered at a rate of 5 Hz. This type of stimulus was used to have a more suitable match to the ASSR stimuli (warble-tone threshold). Air conduction thresholds were measured in each ear by using insert earphones (EARTONE® 5A; Auditory Systems, Indianapolis, IN) at 250 to 8000 Hz; bone conduction thresholds were measured using a Radioear vibrator B-71 at 250 to 4000 Hz. Threshold levels were determined with a 10dB-down/5dB-up approach (Carhart & Jerger 1959). Clinical masking was used if necessary or warranted by the clinical audiologic data. The findings were included in the study only if the behavioral audiograms were considered reliable by the audiologist.
Auditory Steady State Responses ASSRs were assessed in a sound-attenuated room with the subjects in the supine position. (They were told to relax, and most of them slept through the test.) The Bio-logic MASTER Version 2.02 (Biologic System Corp., Mundelein, IL) was used in all cases. Potentials were collected from scalp electrodes located at Cz (active) and referenced to the midline posterior neck (about 7 cm below the inion), with the right earlobe as ground. Electrode impedance was 5 kohm or less at 20 Hz for each electrode, and the difference between the electrodes was 3 kohm or less. Carrier frequencies were matched to the audiometric frequencies; 500, 1000, 2000, and 4000 Hz. In some cases, the behavioral threshold was sufficiently low to allow for ASSR testing at 6000 Hz. The modulation frequencies (amplitude modulation [AM]/frequency modulation [FM]) for those audiometric frequencies were 82, 84, 87, and 89 Hz for the left ear, respectively, and 91, 94, 96, and 98 Hz for the right ear. These high modulation rates were used because they are less affected by the state of subject’s arousal, and because higher-frequency modulators tend to show less of an interaction between carriers (Picton et al. 2009) and exhibit lower growth function slopes (Lins & Picton 1995). Each auditory stimulus consisted of a sinusoidal tone with a carrier frequency that was 25% frequency-modulated and 100% amplitude-modulated.
332
ATTIAS ET AL. / EAR & HEARING, VOL. 35, NO. 3, 330–338
A relative phase of 270 degrees between the AM and FM components was chosen so as to elicit the largest combined response (John et al. 2002). All stimuli were routed through an Amplaid audiometer (Amplaid 319, Amplifon, Via Donizetti, Milan, Italy) and presented to both ears simultaneously through Bio-logic insert earphones (580-SINSER; Biologic System Corp.) with foam earplugs. The ASSR stimuli delivered through the insert phones were calibrated at dB HL in the MASTER setup by using the reference values reported by Wilber et al. (1988). The authors had performed data analysis of five different studies that used puretone hearing threshold measurements of normal listeners at 125 to 8000 Hz frequency range and calibration by 2 cm3 coupler. In the present study, stimulus intensity was also measured with the Bruel & Kjaer (Naerum, Denmark) 2613 amplifier and a 2120 frequency processor coupled to a 4153 microphone. All measured intensities were well correlated (±2 dB) to the normative dB SPL calibration data listed in the MASTER software. In the dichotic multiple frequency test, the initial stimulation level was 80 dB HL. If clear responses were recorded, the level was reduced in 20 dB increments until no responses were recorded, and then increased in 10 dB increments to determine the threshold at which a significant response was elicited. When there was no response to a level of 80 dB HL in the multicarrier condition, a single specific modulated tone of up to 132 dB SPL was presented. This minimized the possibility of undesired interactions between the carriers at levels above 80 dB HL.
ASSR Recordings A sampling rate of 1200 Hz at 16-bit resolution was used. The EEG responses were amplified using a gain of 10,000 at a filter band pass of 3 to 300 Hz (12 dB/octave). The data were recorded in epochs containing 1024 points; one response epoch lasted 0.853 seconds. Sixteen data epochs were collected and linked together to form one sweep with an overall duration of 13.653 seconds. Epochs that contained electrophysiological activity exceeding ±20µV were rejected, and the next acceptable epoch was used to build the sweep. Once completed, each sweep was averaged in the time domain and subsequently analyzed by fast Fourier transform to yield an amplitude spectrum with a resolution of 0.0732 Hz. To determine whether the response at the modulation frequency was significantly different from the background EEG activity in neighboring frequencies, we used the F ratio to compare the amplitude at the modulation frequency with the average amplitude of the noise in 120 adjacent frequency bins (60 bins above and 60 below the modulation frequency, i.e., ±4.4 Hz). A response was accepted as present if the F ratio, calculated against the critical values for F at 2 df and 240 df, was significant at p
