normally-hearing and sensorineurally hearing-impaired adults

The present study provides comparative evaluation of the ABR and MLR to 1 kHz brief tones in two groups of hearing-impaired subjects (noise-masked normally-hearing; and sensorineurally hearing-impaired adults), as well as a normally-hearing control group. Tones were presented at intensities from threshold to 80 /90 dB nHL. The results of this study show that: (1) the ABR and MLR to these low-frequency (1 kHz) tones are equally accurate in estimating hearing threshold, (2) at suprathreshold levels, there are differences in the ABRs and MLRs for subjects with decreased hearing sensitivity resulting from cochlear pathology, compared to those obtained from adults with simulated hearing loss due to broadband masking, and (3) supra-threshold stimuli produce differential effects on the latency and amplitude characteristics of the ABR and MLR in listeners with true sensorineural hearing impairments. Possible physiologic explanations are offered for this differential pattern of results. Sumario El presente estudio permite una evaluación comparativa de las ABR y las MLR con tonos breves de 1 kHz en dos grupos de sujetos con impedimentos auditivos (normoyentes con enmascaramiento por ruido y adultos con hipoacusias sensorineurales) ası́ como en un grupo control de normoyentes. Se presentaron los tonos a intensidades que fueron desde el nivel de umbral hasta 80 /90 dB nHL. Los resultados de este estudio muestran que:(1) Las ABR y MLR a estos tonos de baja frecuencia (1 kHz) son igualmente precisas para estimar los umbrales auditivos. (2) en niveles supraliminares hay diferencias en las ABRs y las MLRs en sujetos con sensibilidad auditiva disminuida que resulta de patologı́a coclear, comparada con las respuestas obtenidas en adultos con pérdida auditiva simulada por enmascaramiento de banda ancha y (3) los estı́mulos supraliminares producen efectos diferenciales en las caracterı́sticas de latencia y amplitud de las ABR y las MLR en sujetos con verdaderos impedimentos sensorineurales. Se ofrecen posibles explicaciones fisiológicas para este patrón diferencial de resultados. The auditory brainstem (ABR) and middle latency (MLR) responses have been advocated for use in estimating the puretone audiogram (for reviews, see: Cacace & McFarland, 2002; Stapells et al, 1994). Within the literature, however, are conflicting reports about whether or not one of these evoked potential measures is superior to the other for estimating thresholds, especially in response to low-frequency stimuli (i.e., .5 and 1 kHz) (e.g., Laukli et al, 1988; McGee & Kraus, 1996; Musiek & Geurkink, 1981; Stapells, 1984). How these two measures compare in this regard is an important consideration for the clinician when the choice of appropriate (or alternative) electrophysiologic measure for a given patient population and/or assessment protocol must be made. Many researchers have shown good response detectability and estimation of hearing sensitivity using tone-evoked ABRs for .5, 1, 2, and 4 kHz tones in normally-hearing and hearing-impaired adults, children, and infants (e.g., Munnerley et al, 1991; Sininger et al, 1997; Stapells et al, 1995; Suzuki et al, 1984; and see Stapells, 2000b for review). A few researchers, however, have reported that ABR recordings to .5 and/or 1 kHz tones yield unacceptably elevated thresholds (i.e., >25 dB re: behavioral threshold), which could preclude their use as an objective audiometric measure to these low-frequency tones (e.g., Beattie et al, 1996; Gorga et al, 1988; Laukli et al, 1988; Scherg & Volk, 1983). The controversy about using tone-evoked brainstem response recordings to estimate low-frequency hearing sensitivity led several researchers to investigate and propose the use of the MLR as an objective electrophysiologic measure of lowfrequency hearing (Kavanagh et al, 1984; Kraus & McGee, 1990; McGee & Kraus, 1996; Musiek & Geurkink, 1981; Scherg & Volk, 1983). Several researchers have shown good threshold prediction from MLR recordings to .5 kHz tones in normally-hearing adults and children (Barajas et al, 1988; ISSN 1499-2027 print/ISSN 1708-8186 online DOI: 10.1080/14992020500060891 # 2005 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society Jo Manette K. Nousak Department of Communication Sciences and Disorders, ML 0379, University of Cincinnati, Cincinnati, OH 45267-0379, USA. E-mail: [email protected] Received: April 27, 2004 Accepted: October 21, 2004 Kavanagh et al, 1984; Kavanagh et al, 1988). Only a few studies, however, have reported on the accuracy of the toneevoked MLR for threshold estimation to frequencies other than .5 kHz in either normally-hearing subjects (Stapells, 1984; Suzuki et al, 1981) or in hearing-impaired subjects (Stapells, 1984; McFarland et al, 1977). Studies of MLR estimates of hearing threshold at other frequencies important for communication (e.g., 1, 2, and 4 kHz) using appropriate acquisition parameters and in various hearing-impaired groups (by degree, type, and configuration of loss) are needed for clinical application of this response. Previous research that compared the detectability rate of the ABR vs MLR by obtaining both responses from the same subjects give conflicting conclusions. For example, Musiek & Geurkink (1981) obtained ABRs and MLRs to clicks from 15 normally-hearing adults and suggested that the MLR was superior to the ABR for threshold estimation. Conversely, Kavanagh et al (1988) recorded ABRs and MLRs to clicks from 48 children and judged the ABR to be superior to the MLR in response detectability. Stapells (1984) obtained both responses from normally-hearing and hearing-impaired adults, and reported that the ABR yielded lower thresholds (i.e., better accuracy) than the MLR. In all of these studies, however, ABR and MLR recordings were made separately, usually using different EEG filter settings (e.g., ABR: 1.5 /3 kHz vs MLR: .01 /.1 kHz), stimuli, presentation rates, and/or a different number of trials. Thus, the conflicting conclusions drawn from these studies may reflect the differences within each study in the acquisition parameters used, rather than differences in the ABR vs MLR regarding their accuracy in threshold estimation. The ABR and MLR may be recorded to low-frequency tones using the same filter settings, stimuli, and presentation rate (Lichtenstein & Stapells, 1996; Scherg & Volk, 1983; Stapells, 2000a; Suzuki & Horiuchi, 1977; Suzuki et al, 1984); and direct comparison of their response properties is best made by recording these responses simultaneously (Suzuki et al, 1981). Although simultaneous recording of the ABR/MLR has been recommended for routine clinical use (Hood, 1995; Kraus & McGee, 1990), very few studies exist that describe ABR/MLR amplitudes and latencies at various intensities, or ABR/MLR threshold estimates (Scherg & Volk, 1983; Suzuki et al, 1981). In addition, these few reports provide results for a limited number of stimulus intensities (Scherg & Volk, 1983; Suzuki et al, 1981) or from a heterogeneous (by configuration, type and degree of hearing loss) hearing-impaired group (Suzuki et al, 1981). Thus further examination of the ABR/MLR is needed. One means of assessing and comparing the ABR/MLR for their accuracy in threshold estimation and detectability in subjects with decreased hearing sensitivity, is by recording these responses simultaneously from a group of normally-hearing subjects with tones presented alone and in the presence of a masking noise. Obtaining masked behavioral and ABR/MLR thresholds from a group of subjects who serve as their own control likely will reduce some of the variability in the results, and provide for consistent threshold elevations, allowing for assessment of the similarities and/or differences between the ABR and MLR. The effects of simultaneously-presented broadband ipsilateral masking noise on the click-evoked ABR are reasonably wellknown. In general, with increasing noise level, the latency of wave V increases and its amplitude decreases (e.g., Beattie et al, 1994; Burkard & Hecox, 1983a,b; Picton et al, 1979; Owen & Burkard, 1991; Stapells, 1984). Few studies, however, have reported on the effects of masking on the transient MLR (Beattie & Boyd, 1985; Gott & Hughes, 1989). Gott & Hughes (1989) reported little or no change in MLR latencies and amplitudes with increasing masker level. Beattie & Boyd (1985) reported similar rates of detectability for the ABR and MLR to .5, 1, and 2 kHz tones presented simultaneously with white noise. However, they presented the tones at a rate of 31.1/s, which is too high for obtaining the transient MLR (Stapells, 1984, 1994), and their recordings and analyses (25 ms) did not include wave Pa, the most robust positive peak of the MLR (Kraus et al, 1994). There are no studies that have assessed the accuracy of the tone-evoked ABR or transient MLR in estimating threshold from normally-hearing subjects whose behavioral threshold has been artificially shifted by broadband masking noise. Further, no studies have compared ABR/MLR latencies and amplitudes in the presence of noise masking, or have compared noisemasked results with those from subjects with sensorineural hearing loss. In light of the above review indicating (i) the need for further studies directly comparing the ABR and MLR for their accuracy in estimating behavioral thresholds to low-frequency tones, (ii) the need for studies that include results from hearing-impaired subjects, and (iii) the lack of information concerning the effects of noise masking on the tone-evoked MLR, the present study was undertaken to examine the ABR vs MLR recorded simultaneously to 1 kHz tones from noise-masked normallyhearing adults and from a group of subjects with sensorineural hearing impairment. Behavioral thresholds were shifted in the normally-hearing group to levels that were similar to those of the sensorineurally-impaired subjects. Comparative evaluation of ABR/MLR response properties included analyses of recordings to tones presented over a range of stimulus intensities, including threshold levels, in order to investigate the changes in ABR/MLR amplitudes and latencies with increasing