Sources of variability in distortion product otoacoustic emissions.

The goal of this study was to determine the extent to which the variability seen in distortion product otoacoustic emissions (DPOAEs), among ears with normal hearing, could be accounted for. Several factors were selected for investigation, including behavioral threshold, differences in middle-ear transmission characteristics either in the forward or the reverse direction, and differences in contributions from the distortion and reflection sources. These variables were assessed after optimizing stimulus parameters for individual ears at each frequency. A multiple-linear regression was performed to identify whether the selected variables, either individually or in combination, explained significant portions of variability in DPOAE responses. Behavioral threshold at the f(2) frequency and behavioral threshold squared at that same frequency explained the largest amount of variability in DPOAE level, compared to the other variables. The combined model explained a small, but significant, amount of variance in DPOAE level at five frequencies. A large amount of residual variability remained, even at frequencies where the model accounted for significant amounts of variance.

Influence of primary-level and primary-frequency ratios on human distortion product otoacoustic emissions.

The combined influence of primary-level differences (L1-L2) and primary-frequency ratio (f2/f1) on distortion product otoacoustic emission (DPOAE) level was investigated in 20 normal-hearing subjects. DPOAEs were recorded with continuously varying stimulus levels [Neely et al. J. Acoust. Soc. Am. 117, 1248-1259 (2005)] for the following stimulus conditions: f2= 1, 2, 4, and 8 kHz and f2/f1=1.05 to 1.4; various L1-L2, including one individually optimized to produce the largest DPOAE. For broadly spaced primary frequencies at low L2 levels, the largest DPOAEs were recorded when L1 was much higher than L2, with L1 remaining relatively constant as L2 increased. As f2/fl decreased, the largest DPOAEs were observed when L1 was closer to L2 and increased as L2 increased. Optimal values for L1-L2 and f2 f1 were derived from these data. In general, average DPOAE levels for the new L1-L2 and f2/f1 were equivalent to or larger than those observed for other stimulus combinations, including the L1-L2 described by Kummer et al. [J. Acoust. Soc. Am. 103, 3431-3444 (1998)] and those defined by Neely et al. in which L1-L2 was evaluated, but f2/f1 was fixed at 1.2.

Using a combination of click- and tone burst-evoked auditory brain stem response measurements to estimate pure-tone thresholds.

DESIGN: A retrospective medical record review of evoked potential and audiometric data were used to determine the accuracy with which click-evoked and tone burst-evoked auditory brain stem response (ABR) thresholds predict pure-tone audiometric thresholds. METHODS: The medical records were reviewed of a consecutive group of patients who were referred for ABR testing for audiometric purposes over the past 4 yrs. ABR thresholds were measured for clicks and for several tone bursts, including a single-cycle, Blackman-windowed, 250-Hz tone burst, which has a broad spectrum with little energy above 600 Hz. Typically, the ABR data were collected because the patients were unable to provide reliable estimates of hearing sensitivity, based on behavioral test techniques, due to developmental level. Data were included only if subsequently obtained behavioral audiometric data were available to which the ABR data could be compared. Almost invariably, the behavioral data were collected after the ABR results were obtained. Because of this, data were included on only those ears for which middle ear tests (tympanometry, otoscopic examination, pure-tone air- and bone-conduction thresholds) indicated that middle ear status was similar at the times of both tests. With these inclusion criteria, data were available on 140 ears of 77 subjects. RESULTS: Correlation was 0.94 between click-evoked ABR thresholds and the average pure-tone threshold at 2 and 4 kHz. Correlations exceeded 0.92 between ABR thresholds for the 250-Hz tone burst and low-frequency behavioral thresholds (250 Hz, 500 Hz, and the average pure-tone thresholds at 250 and 500 Hz). Similar or higher correlations were observed when ABR thresholds at other frequencies were compared with the pure-tone thresholds at corresponding frequencies. Differences between ABR and behavioral threshold depended on behavioral threshold, with ABR thresholds overestimating behavioral threshold in cases of normal hearing and underestimating behavioral threshold in cases of hearing loss. CONCLUSIONS: These results suggest that ABR thresholds can be used to predict pure-tone behavioral thresholds for a wide range of frequencies. Although controversial, the data reviewed in this paper suggest that click-evoked ABR thresholds result in reasonable predictions of the average behavioral thresholds at 2 and 4 kHz. However, there were cases for which click-evoked ABR thresholds underestimated hearing loss at these frequencies. There are several other reasons why click-evoked ABR measurements were made, including that they (1) generally result in well-formed responses, (2) assist in determining whether auditory neuropathy exists, and (3) can be obtained in a relatively brief amount of time. Low-frequency thresholds were predicted well by ABR thresholds to a single-cycle, 250-Hz tone burst. In combination, click-evoked and low-frequency tone burst-evoked ABR threshold measurements might be used to quickly provide important clinical information for both ends of the audiogram. These measurements could be supplemented by ABR threshold measurements at other frequencies, if time permits. However, it may be possible to plan initial intervention strategies based on data for these two stimuli.

A validation and potential clinical application of multivariate analyses of distortion-product otoacoustic emission data.

OBJECTIVE: To test the generalizability of multivariate analyses of distortion-product otoacoustic emission (DPOAE) data. Previously published multivariate solutions were applied to a new set of data to determine if test-performance improvements, evident in previous reports, are retained. An additional objective was to provide an alternative approach for making multivariate dichotomous decisions of hearing status in the clinic, based on DPOAE measurements. DESIGN: DPOAE level and noise were obtained in 345 ears of 187 subjects. Approximately one third of the subjects had normal hearing, whereas the remainder had hearing loss, ranging from 25 to more than 120 dB HL. DPOAE data were collected at each of nine frequencies. After data collection, clinical decision theory, in combination with univariate (DPOAE level and signal-to-noise ratio [SNR]) and multivariate (logistic regression) analyses, was used to construct relative operating characteristic (ROC) curves and to generate ROC curve areas. In addition, test performance was assessed by fixing the false-alarm rate and comparing different approaches to analyses in terms of their failure rates as a function of magnitude of hearing loss. The DPOAE test results were compared with either single-frequency or multifrequency gold standards. The multivariate solutions were taken from previously published work (Dorn et al., 1999; Gorga, et al., 1999). RESULTS: DPOAE level and SNR resulted in roughly equivalent test performance (ROC curve areas and failure rates among ears with hearing loss), although DPOAE level performed better for frequencies above 1 kHz, and SNR performed better for frequencies at 0.75 and 1 kHz. Multivariate analyses resulted in better test performance for nearly all conditions, compared with the univariate approaches that used either DPOAE level or SNR. The improvements in test performance were greatest for the frequencies at which the univariate analyses performed poorest (0.75 kHz, 1 kHz, and 8 kHz). Less difference was observed between univariate and multivariate approaches when multifrequency gold standards were used; however, even for the multifrequency cases, multivariate analyses generally resulted in better performance. An approach that might facilitate the interpretation of multifrequency DPOAE measurements in the clinic is described. CONCLUSIONS: Previously described multivariate analyses were robust in that they improved test performance when applied to an entirely new set of DPOAE data. This, in turn, suggests that the previously described multivariate solutions may have clinical utility in that they are expected to improve test performance at no additional cost in terms of data-acquisition or data-analysis time. In addition to demonstrating that these solutions generalized to new data, an alternative approach to interpreting multifrequency DPOAE measurements is provided that includes the advantages of using multivariate analyses. This new metric may be useful when DPOAEs are used for screening purposes.

Stimulus-frequency otoacoustic emissions measured with amplitude-modulated suppressor tones (L).

Stimulus-frequency otoacoustic emissions (SFOAEs) are typically derived as the difference in sound pressure in the ear canal with and without a suppressor tone added to the probe tone. A novel variation of this method applies a sinusoidal amplitude modulation (AM) to the suppressor tone, which causes the SFOAE to also be modulated. The AM-SFOAE can be separated from the probe frequency using spectral methods. AM-SFOAE measurements are described for four normal-hearing subjects using 6-Hz AM. Because the suppressor modulation is at a higher rate, the AM-SFOAE technique avoids the confounding influence of heartbeat, which also modulates the probe tone.