Establishing Normal Hearing with the Dichotic Multiple-frequency Auditory Steady-State Response Compared to an Auditory Brainstem Response Protocol

Acta Otolaryngol 2004; 124: 62/68 Establishing Normal Hearing with the Dichotic Multiple-frequency Auditory Steady-State Response Compared to an Auditory Brainstem Response Protocol DEWET SWANEPOEL, DUNAY SCHMULIAN and RENÉ HUGO From the Department of Communication Pathology, University of Pretoria, Pretoria, South Africa Swanepoel D, Schmulian D, Hugo R. Establishing normal hearing with the dichotic multiple-frequency auditory steady-state response compared to an auditory brainstem response protocol. Acta Otolaryngol 2004; 124: 62/68. Objective*/To determine the clinical usefulness of the dichotic multiple-frequency (MF) auditory steady-state response (ASSR) technique for estimating normal hearing compared to a 0.5-kHz tone burst and broadband click auditory brainstem response (ABR) protocol in a sample of adults. Material and Methods*/A comparative experimental research design was selected in order to compare estimations of normal hearing obtained with the dichotic ASSR technique at 0.5, 1, 2 and 4 khz with a 0.5-kHz tone burst and broadband click ABR protocol. The recording times required for each procedure were also compared. Normal-hearing subjects (n/28) were selected according to immittance values within normal limits and pure-tone behavioural thresholds of B/25 db HL across frequencies. Results*/The dichotic MF ASSR estimated normal hearing to be, on average, 30/34 db HL across the range 0.5/4 khz. The mean estimate of normal hearing for 0.5 khz using tone burst ABRs was 30 db nhl and the mean click ABR threshold was 16 db nhl, i.e. 14/18 db better than the ASSR thresholds. The dichotic MF ASSR technique recorded 8 thresholds (4 in each ear) in a mean time of 23 min. The ABR protocol recorded 4 thresholds (2 in each ear) in a mean time of 25 min. Conclusion */Both the dichotic MF ASSR and ABR protocols provided a time-efficient estimation of normal hearing. There was no significant difference between the tone burst ABR and MF ASSR techniques in terms of estimation of normal hearing at 0.5 khz. The dichotic MF ASSR technique proved more time-efficient by determining more thresholds in a shorter time compared to the ABR protocol. Key words: ambient noise, amplitude-modulated tones, auditory-evoked response, carrier frequency, modulation frequency, multiple-stimulus technique, tone burst. INTRODUCTION The field of clinical objective audiometry has recently gained a new technique that promises to be a valuable addition to the auditory-evoked response (AER) test battery. The auditory steady-state response (ASSR), which is evoked by continuous amplitude-modulated tones, demonstrates unique characteristics developed primarily to address many of the limitations of the most widely used AER, the auditory brainstem response (ABR). Unlike ABRs obtained with brief transient stimuli, ASSRs are evoked using sustained continuous tones. These modulated tones are frequency-specific because spectral energy is contained only at the frequency of the carrier tone, plus and minus the frequency of modulation (1). While the 40- Hz response initially kindled interest, its application has been limited by its susceptibility to the state of consciousness (2, 3). A faster modulation rate of 75 /110 Hz is not significantly affected by sleep or sedation, and represents essentially the same generators as the ABR (4). These higher rates are suitable for audiometric purposes across populations (5, 6). According to Lins et al. (7), steady-state responses are, first of all, frequency-specific because the continuous tones do not suffer the spectral distortion problems associated with brief tone bursts or clicks (8). Secondly, as the response is periodic, it can best be represented in the frequency domain, simplifying the procedure. The response is determined at the frequency of modulation by a computer using wellestablished statistical procedures, inferring that no subjective judgement by an interpreter is necessary (5, 7, 9). Thirdly, research seems to indicate that amplitude-modulated tones probably represent a better technique for evaluating hearing aid performance than transient stimuli (10). Hearing aids and cochlear implants process continuous stimuli with less signal distortion because these tones do not present with abrupt changes over time, characteristic of transient stimuli (5, 11). According to Rance et al. (8) a fifth advantage of the ASSR lies in the continuous nature of the amplitude-modulated tones, which offers a presentation-level advantage over transient stimuli and allows for the investigation of ears with minimal amounts of hearing. The ASSR has demonstrated itself to be a reliable technique for estimating behavioural hearing thresholds in normal adults, healthy babies and hearingimpaired subjects using tones modulated in amplitude between 70 and 110 Hz (4/6, 10, 12, 13). Rickards et al. (6) recorded ASSR thresholds in healthy babies within the first 4 days of life that were similar to those obtained using unmasked tone bursts in sleeping adults. These promising applications of the ASSR technique may still, however, prove to be very timeconsuming if each frequency for both ears is explored separately (9). # Taylor & Francis 2004. ISSN 0001-6489 DOI 10.1080/00016480310015902

Acta Otolaryngol 124 MF ASSR versus ABR for estimating normal hearing 63 A more recent development of the technique has enabled the simultaneous evaluation of multiple frequencies per ear and is referred to as multiplefrequency (MF) ASSR (4). This implies that distinct modulation rates, more than one octave apart, are used for carrier tones at different frequencies. The modulated tones are added into a complex acoustic stimulus capable of simultaneous activation of different frequency regions within the cochlea. The technique is further optimized if two differently modulated MF stimuli are presented simultaneously to the left and right ears. In such a case multiple frequencies are explored simultaneously in both ears (9). This multiple-stimulus technique is therefore able to significantly decrease the time required to evaluate thresholds at multiple audiometric frequencies dichotically (4, 5). According to Lins et al. (7), using the dichotic MF ASSR to estimate pure-tone behavioural thresholds could be several times more efficient than using an ABR protocol, with the added advantage of presenting the results in the form of a conventional audiogram. This is an important characteristic, especially for the paediatric population for whom available assessment time is often very limited. The MF ASSR technique, and more specifically the dichotic multiple-stimulation condition, still, however, require clinical validation (14). Promising results for the estimation of normal hearing have only been reported for relatively small samples of adults and infants (4, 5, 9, 10, 14). In 1995, Lins and Picton (4) used the dichotic MF ASSR technique in normalhearing adult subjects (n/8) to evaluate two frequencies in both ears simultaneously. The results indicated average ASSR thresholds of 31 db HL for 0.5 khz and 25 db HL for 2 khz. In a subsequent study by Lins et al. (5), 4 frequencies were monaurally evaluated simultaneously in 15 normal-hearing subjects. The results indicated ASSR thresholds of 39, 29, 29 and 31 db HL for frequencies of 0.5, 1, 2 and 4 khz, respectively, with SDs varying between 10 and 15 db. More recently, Picton et al. (10), using the MF ASSR technique, reported thresholds for 10 subjects varying between 21 and 26 db HL (SD 7/15 db) at frequencies of 0.5, 1, 2 and 4 khz. In 2001, Perez-Abalo et al. (9) and Herdman and Stapells (14) reported results using the dichotic MF ASSR technique to evaluate four frequencies simultaneously in normal-hearing adults. Perez-Abalo et al. (9) evaluated 40 subjects and Herdman and Stapells (14) 10 subjects at 0.5, 1, 2 and 4 khz. Threshold results reported by Perez-Abalo et al. (9) varied between 25 and 30 db HL, with SDs of 9/10 db. Herdman and Stapells (14), however, reported thresholds of 7/14 db HL with SDs varying between 7 and 9 db. This variability which is evident in ASSR results reported for normal-hearing subjects can be attributed to several factors, such as the method of threshold determination, the size of the intensity step, the state of the subject and internal and external noise levels (10, 14, 15). The question that arises is whether one AER technique is able to provide all the necessary information to infer a hearing acuity profile in a clinically viable way. Picton (16) specifies five criteria for the perfect AER in estimating behavioural auditory thresholds. Firstly, the response must provide a reasonably accurate assessment of hearing threshold. Secondly, the response should be easily recorded during different states, and changes in arousal. Thirdly, the response must be easily recognizable at all ages. Fourthly, the response should be present at all frequencies of the conventional audiogram. According to Picton (16), the aim of objective procedures should remain identical to those of traditional testing, i.e. to obtain an audiogram if not at all frequencies, then at least between 0.5 and 4 khz. Fifthly, the stimulus used must evoke responses that measure thresholds specific to different frequencies. The issue in this case is not the response, but the stimulus used to elicit the response. A sixth criterion for the perfect AER, which is not mentioned by Picton (16) but is by other authors (17, 18), is the time required to obtain this information. Objective test procedures must be performed as quickly as possible, especially in the paediatric population. These criteria supply a framework with which to view emerging AER techniques and to provide comparisons with existing techniques, such as the ABR, in order to determine the advantages and limitations of each. The emergence of the ASSR has necessitated that it be validated alongside existing techniques such as the ABR. The aim of this paper was therefore to investigate, comparatively, the dichotic MF ASSR technique and an ABR protocol in the light of two criteria, namely the accuracy of estimating normal hearing and the recording time required. MATERIAL AND METHODS Subjects The study sample comprised 28 normal-hearing subjects (12 females, 16 males; age range 17/38 years). A total of 35 subjects were studied but only 28 completed all the tests and provided reliable results. Subjects were required to have behavioural pure-tone thresholds of 5/25 db HL at octave frequencies of 0.5/4 khz. The majority of pure-tone thresholds for normal-hearing subjects were 5/10 db HL and were obtained in a double-walled soundproof booth using 10 db down and 5 db up intensity steps.

64 D. Swanepoel et al. Acta Otolaryngol 124 All subjects underwent a thorough audiological test battery, including otoscopy, tympanometry and puretone audiometry, before any auditory-evoked responses were performed. Normal middle-ear compliance was a prerequisite for including any subject. Stimuli ASSRs were recorded using MF amplitude-modulated tones presented dichotically through TDH 39 supraaural earphones. Each ear was stimulated with MF stimuli consisting of 4 carrier frequencies (0.5, 1, 2 and 4 khz) modulated in amplitude (95% depth) at 77, 85, 93 and 101 Hz for the left ear and 81, 89, 97 and 105 Hz for the right ear. The MF stimuli were calibrated separately for each frequency using pure tones, as per the AS 1591.2 standard. All measurements were made with a sound-level meter (Investigator 2260; Brüel & Kjaer), an artificial ear (type 4152; Brüel & Kjaer) and a microphone (type 4144; Brüel & Kjaer). The MF stimuli were automatically adjusted to ensure that the overall acoustic energy corresponded to the nominal SPL value specified in the software. ABR thresholds were obtained using a broadband click and a 0.5-kHz tone burst (TB) stimulus. A 100-ms acoustic broadband click and a 0.5-kHz TB with a 6-ms period (rise/ fall: 2 ms; plateau: 2 ms) produced with a Blackman envelope were used to illicit responses. The click and 0.5-kHz TB stimuli were calibrated using a Larson Davis 824 instrument connected to an IEC 318 artificial ear simulator. Peak equivalent methods were used to calibrate the stimuli. The nhls for the click and TB stimuli were established by testing a group of 20 normal-hearing adults. Recording procedures MF ASSR and ABR thresholds were recorded in a single-walled soundproof booth using the Audix system (Neuronic S.A., Havana, Cuba). Ambient noise levels for the room were 16.6, 22, 18.2 and 13.1 db SPL for frequencies of 0.5, 1, 2 and 4 Hz, respectively. Subjects reclined on a bed and were encouraged to relax and sleep if possible. The electrophysiological thresholds for normal-hearing subjects were obtained using a 10-dB intensity step. Dichotic MF ASSRs were recorded with electrode discs of Ag/AgCl fixed with electrolytic paste to the scalp at Cz (positive), Oz (negative) and Fpz (ground) positions. Impedance values were kept below 3000 V. Stimulation was initially presented dichotically at a supra-threshold intensity of 50 db HL. The bioelectric activity was amplified with a gain of 100 000 and analogue-filtered between 30 and 300 Hz with the notch filter switched on at 50 Hz. No less than 10 and no more than 40 epochs of 8192 samples (digitized with a sampling period of 1.37 ms) each were averaged in a response. A fast Fourier transform was calculated online for each long epoch to average the response spectra continuously. The presence of a response was determined by using the F-test for hidden periodicity to test the amplitude of the spectrum at each modulation frequency against the 120 neighbouring bins for significant amplitude difference. The significance level for statistical detection of a signal was set at p B/0.05. Artefact rejection was carried out with shorter epoch sections of 512 points and a rejection level of 50 mv was specified. The minimum response level (or lowest obtained response) for each frequency in each ear, obtained with electrical noise levels of B/0.1 mv, was considered a threshold. ABR recordings were performed on the same day as the MF ASSR procedure. Electrode discs of Ag/AgCl were fixed with electrolytic paste to the scalp at Cz (positive), A1 and A2. A1 and A2 were switched between reference and ground depending on the test side. Impedance values were kept below 3000 V and stimulation commenced monotically at a supra-threshold intensity of 50 db nhl. The bioelectric activity was amplified with a gain of 100 000 and analoguefiltered between 100 and 3000 Hz for the click ABR and between 30 and 3000 Hz for the TB ABR. The notch filter switched on at 50 Hz. A maximum of 2000 recordings were averaged per trial. Two replications were made near and at minimum response levels. The minimum response level was considered as the lowest intensity at which a repeatable wave V was replicated. The click and 0.5-kHz ABR thresholds were determined firstly in the left ear, followed by the right ear. Three clinicians experienced in the interpretation of ABR waveforms judged the minimum response level as the threshold. The software recorded the test data and the time at which each procedure commenced and terminated. This allowed for exact measurement of the time required for each procedure for each subject. RESULTS Thresholds The mean9/sd of the ABR and MF ASSR thresholds is shown in Table I. The mean MF ASSR thresholds demonstrated a consistent SD of 9/11 db across all frequencies. The majority (62%) of the dichotic MF ASSR thresholds (224) were 5/30 db HL, whilst only 26% were 5/20 db HL. Of the MF ASSR thresholds at 20 db HL, 81% represented behavioural thresholds between 0 and 5 db HL, and of the MF ASSR thresholds at 30 db HL, 76% represented behavioural thresholds between 0 and 5 db HL. Fig. 1 presents the frequency distribution of dichotic MF ASSR thresholds. The mean 0.5-kHz TB and broadband click ABR thresholds differed considerably

Acta Otolaryngol 124 MF ASSR versus ABR for estimating normal hearing 65 Table I. Dichotic MF ASSR and ABR thresholds (mean9/sd) obtained for 56 normal-hearing ears, together with dichotic MF ASSR thresholds (mean9/sd) obtained in 2 previously published studies Stimulus (khz) ABR (db nhl) MF ASSR (db HL) Perez-Abalo et al. (9) Herdman and Stapells (14) 0.5 309/16 339/11 309/11 149/9 1 349/11 289/9 119/8 2 329/11 259/10 79/7 4 309/11 279/11 139/9 Broadband click 169/7 single signals (2 in each ear) in a mean time of 25 min with an SD of 9/8 min. Fig. 1. Frequency distribution of ASSR (0.5, 1, 2 and 4 khz) thresholds for 56 normal-hearing ears. The percentage of ASSR thresholds acquired at different intensity levels is represented. The majority of thresholds were acquired at 30 db HL, except for 0.5 khz where the majority of responses were acquired at 40 db HL. from each other. Almost a half (48%) of the 0.5-kHz TB thresholds were 5/20 db nhl, and 61% were 5/30 db nhl. Of the 0.5-kHz TB thresholds at 20 db nhl, 67% represented behavioural thresholds between 0 and 5 db HL, and of the thresholds at 30 db nhl, 73% represented behavioural thresholds between 0 and 5 db HL. Only 9% of the click ABR thresholds were measured above 20 db nhl and 48% were equal to 10 db nhl. The click ABR thresholds at 10 db represented 74% of pure-tone averages between 0 and 5 db HL and the click ABR thresholds at 20 db represented 71% of pure-tone averages between 0 and 5 db HL. The click ABR was selected for this study in order to represent common clinical practice for evaluating high-frequency hearing. In retrospect a high-frequency TB would have proved a more suitable comparison against the ASSR technique. Recording time The recording time required by the dichotic MF ASSR technique and the ABR protocol is another important measured variable. The dichotic MF ASSR technique recorded thresholds using 8 simultaneous signals (4 in each ear) in a mean time of 23 min with an SD of 9/8 min. The ABR protocol recorded thresholds using 4 DISCUSSION Dichotic MF ASSR thresholds The ASSR thresholds for the normal-hearing subjects obtained in the current study are generally higher than, or similar to, the ASSR thresholds previously reported in the literature (7, 9, 10, 12, 14). Mean ASSR thresholds obtained in the current study ranged between 30 and 34 db HL for carrier frequencies between 0.5 and 4 khz. Aoyagi et al. (12) reported ASSR thresholds for normal-hearing adults of 29 db HL for single stimuli with a carrier frequency of 1 khz amplitude-modulated at 80 Hz. Using the multiplestimulus technique to obtain thresholds for 4 stimuli (2 in each ear) in normal-hearing subjects, Lins and Picton (4) reported mean thresholds of 31 db HL for 0.5 khz and 25 db HL for 2 khz. In a subsequent study by Picton et al. (10), ASSRs were measured for normal-hearing subjects using the multiple-stimulus technique with four stimuli presented simultaneously to one ear. The mean ASSR thresholds obtained at 0.5, 1, 2 and 4 khz were :/ 20, 29, 19 and 17 db HL (converted from db SPL), respectively. Results in normal-hearing subjects using this same dichotic MF ASSR technique were reported by Herdman and Stapells (14) and by Perez-Abalo et al. (9). Mean thresholds and SDs reported by these authors are contrasted with those obtained in the current study in Table I. Threshold data were converted into db HL according to American National Standard Institute (19) specifications in order to allow comparisons. Herdman and Stapells (14) reported mean MF ASSR thresholds for 10 normal-hearing subjects. Perez-Abalo et al. (9) reported MF ASSR thresholds for 40 normal-hearing subjects that were higher than those reported by Herdman and Stapells (14), and the thresholds obtained in the current study are slightly (3/7 db) higher than those of Perez-Abalo et al. (9). The SDs from the mean thresholds are consistent between studies. The differences in thresholds for these studies can be attributed to several factors and three methodological variations are discussed below.

66 D. Swanepoel et al. Acta Otolaryngol 124 The first concerns the threshold-seeking procedure utilized by Perez-Abalo et al. (9), in the current study and by Herdman and Stapells (14). Perez-Abalo et al. (9) and the present authors used a threshold-seeking procedure that was only sensitive to intensity steps of 10 db, in contrast to the threshold-seeking procedure that was sensitive to intensity steps of 5 db used by Herdman and Stapells (14). This discrepancy in threshold-seeking procedures can account for up to 5 db of the difference between the thresholds obtained by Herdman and Stapells (14) and those obtained by Perez-Abalo et al. (9) and the present authors. An important influence affecting the results of each study is the level of acoustic ambient noise (10, 14). The ASSR thresholds reported by Herdman and Stapells (14) were recorded in a double-walled, sound-attenuated booth with low levels of acoustic ambient noise between 10 and 12 db SPL across the octave bands centred at 0.5, 1, 2 and 4 khz. The thresholds obtained by Herdman and Stapells (14) were 10 db better than those reported by Lins and Picton (4) for simultaneous presentation of two amplitude-modulated tones to both ears. The authors suggest that this improvement in threshold sensitivity can probably be attributed to the lower ambient acoustic noise levels in the test environment. The slightly smaller range of normal deviation reported by Herdman and Stapells (14) is commensurate with the lower thresholds obtained in the low levels of acoustic ambient noise. Perez-Abalo et al. (9) obtained ASSR thresholds in a sound-treated room with acoustic ambient noise levels greater than permissible (20) at 30, 30, 27 and 21 db SPL for frequencies of 0.5, 1, 2 and 4 khz. ASSR recordings in the current study were made in a single-walled, sound-attenuated booth with ambient noise levels of 17, 22, 18 and 13 db SPL for octave bands centred at 0.5, 1, 2 and 4 khz. A third aspect that must be pointed out is the difference in averaging procedures used in the various studies. The time taken to complete one electroencephalogram sweep in the current study and that of Perez-Abalo et al. (9) was 11.14 s. Herdman and Stapells (14) took slightly longer at 16.38 s per single sweep. The number of sweeps required to identify a significant response varied between a minimum and maximum number of sweeps, with higher numbers of sweeps usually required for low-intensity signals. A non-response could only be determined after the maximum number of sweeps was averaged. Herdman and Stapells (14) had the highest maximum value of sweeps, along with the longest duration for a single sweep. According to Picton et al. (10), longer periods of averaging might show responses closer to behavioural thresholds, as is the case in prolonged averaging of the ABR. Thus the results reported by Herdman and Stapells (14) might show reduced differences between behavioural and ASSR thresholds when compared to the current study because of the longer duration of the sweeps and the higher quantity of averaged sweeps. These differences are reflected in the average time taken to complete an ASSR recording as reported in these three studies. The average recording time reported by Herdman and Stapells (14) is almost three times longer than those reported by Perez-Abalo et al. (9) and the present authors. ABR thresholds The mean click ABR threshold obtained in the current study was 16 db nhl. In general the click ABR is a very good predictor of average hearing loss, and there is excellent correspondence between pure-tone average hearing loss and click ABR threshold (21, 22). The mean 0.5-kHz TB threshold obtained in the current study (30 db nhl) is well within the range of TB results reported in the literature for normal-hearing adult subjects, which vary between 8.8 and 41.7 db nhl (23). The literature meta-analysis of Stapells (23) indicates that TB thresholds in normal-hearing adult subjects are typically evoked between 10 and 20 db nhl but that 0.5-kHz TB thresholds are usually elevated by 5 /10 db compared to higher-frequency TB thresholds. The SD (16 db) of the ABR thresholds evoked by 0.5-kHz TBs in the current study is at the upper limit of the range specified in the literature of 3.5/15.6 db (23). The increased SD along with the mean thresholds at the upper limit of the specified range may relate to the use of only 2000 averages and 2 replications performed near threshold. More averaging and replications could have contributed to better (lower) threshold detection. A shorter 0.5-kHz TB stimulus (e.g. a period of 4 ms with a 2 ms rise and 2 ms fall) may also have been more efficient in obtaining better responses. Comparison of ASSR and ABR thresholds Limited data exists comparing TB ABR and ASSR estimations of hearing sensitivity (24). Comparison between the mean 0.5-kHz TB ABR threshold and the 0.5-kHz MF ASSR threshold favoured the TB threshold slightly by 4 db. Sinninger and Cone-Wesson (21) refer to an unpublished study by Kosmider (25) comparing TB and ASSR thresholds at 0.5 and 4 khz for 10 normal-hearing subjects. There was no statistically significant difference in threshold estimates for ASSR and ABR detected by Fsp at either test frequency, although visual detection of the ABR yielded the lowest threshold estimates. The authors conclude that for normal-hearing subjects TB ABRs and ASSRs can generally be detected within 20 db of the behavioural threshold. The TB ABR and MF

Acta Otolaryngol 124 MF ASSR versus ABR for estimating normal hearing 67 ASSR thresholds for this study were in agreement with these results, indicating slightly better TB thresholds at 0.5 khz determined by means of visual inspection than the ASSR determined by means of statistical analysis. The mean click ABR threshold was 14/18 db better than all the MF ASSR thresholds. The click ABR was selected as representative of common clinical practice, and unfortunately a frequency-specific ABR measure was not recorded for the high frequencies to compare against the MF ASSR results. Although the click ABR evaluates the high-frequency region between 1 and 4 khz, and more specifically between 2 and 4 khz (26), it does not provide a frequency-specific threshold measure comparable to the MF ASSR results. The larger neural response elicited by the broadband click stimuli is able to provide more time-efficient responses at lower levels of stimulation, accounting for the lower threshold results compared to the ASSR thresholds in this study. Recording time The average recording times for the dichotic MF ASSR and ABR protocol thresholds did not differ significantly, although the average dichotic MF ASSR recording time was 2 min shorter. What is significant to note, however, is that the dichotic MF ASSR technique evaluated auditory sensitivity at four frequencies in each ear in less time than the ABR protocol required to evaluate two frequency regions in each ear. According to Bachmann and Hall (18), the recording time is of fundamental importance, especially for paediatric populations, because it invariably affects the amount of information that can be acquired regarding an individual s auditory status. Being able to obtain more information regarding auditory sensitivity in a shorter period of time is an important advantage, especially for difficult-to-test subjects for whom assessment time is limited (18). Herdman and Stapells (14), as well as Perez-Abalo et al. (9), reported average recording times for the dichotic MF (four frequencies/ear) ASSR technique. Results from the current study and those of Perez- Abalo et al. (9) indicate similar recording times, whilst the recording time reported by Herdman and Stapells (14) differs significantly. There is a difference of :/60 min between the recording time of Herdman and Stapells (14) and those reported in the other 2 studies. Reasons for these differences in recording time have already been mentioned. The averaging procedure used by Herdman and Stapells (14) utilized longer sweeps and a larger maximum number of sweeps than either of the other studies, and theirs was the only study to determine thresholds in a more thorough but time-consuming manner using intensity steps of 10 db down and 5 db up. A related reason for the extended recording times is the low level of stimulation at which ASSR thresholds were recorded by Herdman and Stapells (14). MF ASSR thresholds were recorded at intensities significantly lower than those in the other 2 studies, with mean response threshold intensities of 13.5, 10, 6 and 12.5 db HL (converted from db SPL) at 0.5, 1, 2 and 4 khz, respectively. According to Picton et al. (10), at near-threshold intensities there is probably too much latency jitter in the ASSR to allow averaging to detect a response and the authors suggested that no recognizable ASSR responses below 10 db HL could be recorded. The mean ASSR thresholds reported by Herdman and Stapells (14), however, were recorded at or near this level. The averaging process involved in determining MF ASSR thresholds at such low intensities requires the maximum amount of averaging to determine a significant ASSR. This means a maximum amount of time is necessary because of the inherent difficulty in determining a significant ASSR at these low intensities. CONCLUSION The dichotic MF ASSR and ABR protocols provided a time-efficient estimation of normal hearing. There was no significant difference between the TB ABR and MF ASSR techniques in terms of the prediction of normal hearing at 0.5 khz. The click ABR, although not frequency-specific, provided a significantly closer approximation of normal hearing than the ASSR thresholds. The dichotic MF ASSR technique provided four thresholds for each ear in a more timeefficient manner than the ABR protocol took to establish two thresholds for each ear. ACKNOWLEDGEMENTS We thank the Cuban Neuroscience Centre, and especially Arquimedes Montoya and Maria Abalo-Perez, for their collaborative effort in bringing this project to fruition. Some of these data were presented at the 17th Biennial Symposium of the International Evoked Response Audiometry Study Group, Vancouver, B.C., 22 /27 July, 2001. REFERENCES 1. Van der Reijden CS, Mens LHM, Snik FM. Comparing signal to noise ratios of amplitude modulation following responses from four EEG deprivations in awake normally hearing adults. Audiology 2001; 40: 202/7. 2. Hall III JW. Handbook of auditory evoked responses. Boston, MA: Allyn & Bacon, 1992.

68 D. Swanepoel et al. Acta Otolaryngol 124 3. Hood LJ. Clinical applications of the auditory brainstem response. San Diego, CA: Singular Publishing Group, 1998. 4. Lins OG, Picton TW. Auditory steady-state responses to multiple simultaneous stimuli. Electroencephalogr Clin Neurophysiol 1995; 96: 420 /32. 5. Lins OG, Picton TW, Boucher BL, et al. Frequencyspecific audiometry using steady-state responses. Ear Hear 1996; 17: 81/96. 6. Rickards FW, Tan LE, Cohen LT, Wilson OJ, Drew JH, Clark GM. Auditory steady state evoked potentials in newborns. Br J Audiol 1994; 28: 327/37. 7. Lins OG, Picton PE, Picton TW, Champagne SC, Durieux-Smith A. Auditory steady-state responses to tones amplitude-modulated at 80 /110 Hz. J Acoust Soc Am 1995; 97: 3051/63. 8. Rance G, Dowell RC, Rickards FW, Beer DE, Clark GM. Steady state evoked potential and behavioral hearing thresholds in a group of children with absent click evoked auditory brainstem response. Ear Hear 1998; 19: 48/61. 9. Perez-Abalo MC, Savio G, Torres A, Martin V, Rodriguez E, Galan L. Steady state responses to multiple amplitude modulated tones: an optimized method to test frequency specific thresholds in hearing impaired children and normal subjects. Ear Hear 2001; 22: 200 / 11. 10. Picton TW, Durieux-Smith A, Champagne S, et al. Objective evaluation of aided thresholds using auditory steady-state responses. J Am Acad Audiol 1998; 9: 315 / 31. 11. Collet L, Gallégo S, Durrant JD, Truy E. Electrically evoked multiple auditory steady-state responses recorded in digisonic cochlear-implanted patients. Proc 17th Biennial Symposium of the International Evoked Response Audiometry Study Group, Vancouver, B.C., 22 /27 July, 2001. 12. Aoyagi M, Kiren T, Furuse H, et al. Pure-tone threshold prediction by 80 Hz amplitude modulation following response. Acta Otolaryngol Suppl (Stockh) 1994; 504: 7/14. 13. Rance G, Rickards FW, Cohen LT, De Vidi S, Clark GM. The automated prediction of hearing thresholds in sleeping subjects using auditory steady-state evoked potentials. Ear Hear 1995; 16: 499 /507. 14. Herdman AT, Stapells DR. Thresholds determined using monotic and dichotic multiple auditory steady state response technique in normal-hearing subjects. Scand Audiol 2001; 30: 41/9. 15. Dimitrijevic A, John MS, Van Roon P, et al. Estimating the audiogram using multiple auditory steady-state responses. J Am Acad Audiol 2002; 13: 205/24. 16. Picton TW. Clinical usefulness of auditory evoked potentials: a critical evaluation. J Speech Lang Pathol Audiol 1991; 15: 3/18. 17. Arnold SA. The auditory brain stem response. In: Roeser RJ, Valente M, Hosford-Dunn H, eds. Audiology diagnosis. New York: Thieme Medical Publishers, 2000: 451/70. 18. Bachmann KR, Hall JW. Pediatric auditory brainstem response assessment: the cross-check principle twenty years later. Semin Hear 1998; 19: 41/60. 19. ANSI. American national standard specification for audiometers, ANSI S3.6. New York: American National Standards Institute (ANSI), 1996. 20. ANSI. Maximum permissible ambient noise levels for audiometric test rooms, ANSI S3.1. New York: American National Standards Institute (ANSI), 1999. 21. Sinninger, Cone-Wesson. In: Katz J, ed. Handbook of clinical audiology, 5th edn. London: Williams & Wilkins, 1994: 3/5. 22. Gorga MP, Kaminski JR, Beauchaine KL, Bergman BM. A comparison of auditory brain stem response thresholds and latencies elicited by air- and boneconducted stimuli. Ear Hear 1993; 14: 85/94. 23. Stapells D. Threshold estimation by the tone-evoked ABR: a literature meta-analysis. J Speech Lang Pathol Audiol 2000; 24: 74 /83. 24. Aoyagi M, Yamazaki Y, Yokota M, et al. Frequency specificity of 80 Hz amplitude modulation following response. Acta Otolaryngol Suppl (Stockh) 1996; 522: 6/10. 25. Kosmider D. The auditory brainstem response and the steady state evoked potential as predictors of the behavioural audiogram [unpublished master s thesis]. Melbourne, Vic.: Department of Otolaryngology, Audiology and Speech Sciences, University of Melbourne, 1997. 26. Gorga MP. Predicting auditory sensitivity from auditory brainstem response measurements. Semin Hear 1999; 20: 29/43. Submitted April 3, 2002; accepted August 7, 2003 Address for correspondence: DeWet Swanepoel Department of Communication Pathology University of Pretoria Pretoria 0002 South Africa Tel.: /27 12 4205152 Fax: /27 12 4203517 E-mail: dswanepoel@postino.up.ac.za