Variability in DPOAE Measurements and its Relation to the Magnitude and Frequency of Occurrence of Peaks in the DPOAE-gram

Abstract

This study investigated the magnitude, reliability, and frequency of occurrence of peaks in the fine structure of the distortion product otoacoustic emission audiogram (DPOAE-gram). Thirty-six normal-hearing women were tested. Fine structure was assessed using 1/15 octave intervals (1000-8000 Hz) with primary tones presented at L1 = 65 and L2 = 55 dB SPL. DPOAE amplitudes were obtained at five frequencies (f2 = 1500 Hz, 2000 Hz, 3031 Hz, 4000 Hz, and 6031 Hz) for three values: one-frequency, three-frequency, and five-frequency averages. Compared to the one-frequency condition, averaging DPOAE amplitudes across three frequencies resulted in an improvement in reliability of approximately 1.0 dB. Because of the improved reliability, the three-frequency condition may be preferred when monitoring the effects of noise or ototoxic drugs. The number and reliability of amplitude peaks was ascertained. The reliability of the peak amplitudes (maxima or minima) was assessed with the standard deviation of test-retest differences. This value was 2.9 dB, suggesting that approximately 95% of measurements revealed test-retest differences within (+ or -) 5.9 dB. Thus, when obtaining measurements in 1/15 octave steps, it is common to observe maxima/minima in the DPOAE-gram of <6 dB that reflect test-retest variability. Moreover, due to chance, peaks greater than 6 dB will be observed in approximately 5% of measurements. Individual data points in the DPOAEgram may fall below normative values because of hearing loss, unreliability, or because of fine structure minima. Therefore, clinicians should conduct repeat testing and/or fine resolution measurements to ascertain whether the abnormal test results are due to cochlear impairment or normal variability.