Towards a joint reflection-distortion otoacoustic emission profile: Results in normal and impaired ears

Abstract

Otoacoustic emissions (OAEs) provide salient information about cochlear function and dysfunction. Two broad classes of emissions, linear reflection and nonlinear distortion, arise via distinct cochlear processes and hence, appear to provide independent information about cochlear health and hearing. Considered in combination, these two OAE types may characterize sensory hearing loss most effectively. In this study, the level-dependent growth of stimulus-frequency OAEs (a reflection-type emission) and distortion-product OAEs (a distortion-type emission) were measured in ten normal-hearing ears and eight ears with slight-to-moderate sensorineural hearing loss. Metrics of OAE strength and compression were derived from OAE input/output functions and then considered in a combined fashion. Results indicate that SFOAEs and DPOAEs differ significantly in their strength and compression features. When SFOAE and DPOAE metrics are displayed together on a two-dimensional plot, relatively well-defined data clusters describe their normative relationship. In hearing-impaired ears, this relationship is disrupted but not in a uniform way across ears; ears with similar audiograms showed differently altered joint-OAE profiles. Hearing loss sometimes affected only one OAE or one more than the other. Results suggest a joint-OAE profile is promising and warrants study in a large group of subjects with sensory hearing loss of varied etiologies.

FIG. 1.

Audiograms for eight hearing-impaired subjects.

FIG. 2.

(Color online) (A) and (D) OAE spectra for stimulus-frequency (reflection) OAEs and the distortion-component of the DPOAE (distortion OAEs) respectively, in the same ear. Both spectra depict OAE magnitude at successively lower stimulus levels. In general, the systematically decreasing response magnitude indicates reduced stimulus levels (level is designated by color online). The noise floor is shown in light-gray. A vertical dashed line and arrow indicate one exemplar analysis frequency at 1.1 kHz. Input/Output functions were created at all analysis frequencies. (B) and (E) The input-output function, which plots OAE magnitude as a function of stimulus level, at 1.1 kHz for both SFOAE and DPOAE data. (C) and (F) The input/output function at 1.1 kHz transformed into a “gain function” by dividing OAE magnitude at each level by stimulus level. The gain function plots the OAE in dB re: stimulus level, and is a measure of OAE strength. A fit was applied to gain functions as shown by the black (SFOAE) and dark gray (DPOAE) lines superimposed on the data to estimate three parameters: PS, CT, and CS.

FIG. 3.

(Color online) (A) and (B) The same SFOAE and DPOAE gain function presented in Fig. 2 at one analysis frequency and the three parameters derived from this fit: PS, CT, and CS. The smaller inset above the main panels provides a visual depiction of the SF and DP OAE fits superimposed: the SFOAE fit is typically higher than the DPOAE fit, indicating greater PS; and the compression “knee” in the fit typically occurs at higher stimulus levels for the DPOAE than SFOAE.

FIG. 4.

Three metrics derived from SFOAE (open circles) and DPOAE (gray squares) gain functions are shown in normal-hearing ears. PS, CT, and CS are plotted as a function of analysis frequency. Linear regressions are plotted on the data for each OAE-type separately. PS was greater for SFOAEs, CT was higher for DPOAEs, and CS was not different between the two OAE types.

FIG. 5.

The two most informative metrics derived from OAE gain functions, PS and CT, are graphed jointly on an x-y plot; their position in this two-dimensional space elucidates the relationship between reflection and distortion emissions. PS in (A) falls mostly below the diagonal confirming that SFOAEs show greater strength than DPOAEs at this frequency-band (1–2.5 kHz). CT in (B) is above the diagonal for most ears, indicating that response compression occurs at higher stimulus levels for DPOAEs compared to SFOAEs. PS and CT have relatively well-defined clusters describing the normative space for each measure. The vertical and horizontal lines define the 75% interquartile ranges.

FIG. 6.

(A) OAE Level at CT (L@CT) plotted for SFOAEs (open circles) and DPOAEs (gray squares) across analysis frequency in normal-hearing subjects. A linear regression fit to each OAE separately shows no real difference between the two OAEs in the low-to-mid-frequencies but above 2 kHz, the SFOAE shows higher levels at CT than does the DPOAE. (B) L@CT in a two-dimensional x-y plot. The data cluster around the diagonal indicating that OAE levels normalized by CT are relatively equal for the two OAE types in this low-to-mid frequency range. Bars provide 75% interquartile ranges.

FIG. 7.

(Color online) SFOAE (lower panels) and DPOAE spectra (upper panels) shown for five of the eight hearing-impaired ears. Three of those displayed include spectra from both OAE-types in the same ear. The mean noise floor is shown as a dashed gray line. All ears with hearing-impairment had OAEs showing “islands” or frequency-segments of measureable response for at least three stimulus levels, thus allowing for a fit and the calculation of PS, CT, and CS.

FIG. 8.

OAE gain functions and fits (SFOAE – black, DPOAE – gray) are shown for four of the eight hearing-impaired ears. The normal template is provided as a referent (NH). Hearing loss did not disrupt the joint-OAE profile in a uniform manner. HI03 and HI04 continue to show a typical pattern of stronger SFOAEs than DPOAEs but with minimally compressive response growth; HI02 and HI05, in contrast, show equivalent or stronger DPOAEs compared to SFOAEs, which is an atypical pattern.

FIG. 9.

(Color online) (A) The joint-OAE profile for PS as a two-dimensional plot at one analysis frequency-band (1–2.5 kHz). Data from all 18 subjects are included. Hearing-impaired ears are designated by filled squares and the subject number is provided for association with audiograms in Fig. 1. (Normal-hearing ears are shown as open circles.) All but one hearing-impaired ear falls outside of the normal range of PS and incipient clusters appear to be present (e.g., HI01, HI04, HI08 and HI07) but too few data are available to confirm this. (B) The joint-OAE profile for CT from all 18 subjects. Four of the eight impaired ears fall outside of the nucleus of normal SFOAE-DPOAE CT. Two ears (HI04, HI06) show strongly elevated DPOAE CT but normal SFOAE CT; another shows the opposite pattern (HI02). HI01 shows elevated CT for both emission types. Not all sensory hearing losses exhibit the same pattern of OAE disruption even though audiograms are similar.

FIG. 10.

(Color online) OAE Level (dB SPL) at CT for SFOAEs and DPOAEs at one analysis frequency-band (1–2.5 kHz). In general, at low-to-mid frequencies, OAE L@CT is comparable for reflection and distortion emissions in normal ears. Four of these eight hearing-impaired ears show reduced L@CT, as noted by movement toward the lower left of the diagonal. However, the OAEs, though reduced in level, retain their relationship in impaired ears and hug the diagonal line. A different pattern of disruption in L@CT might be evident at higher frequencies.