This investigation addressed four factors affecting transient-evoked otoacoustic emission (TEOAE) reliability: 1) The effect of evoking-stimulus level, 2) the effect of analyzing bandwidth, 3) the effect of slight-mild hearing loss, and 4) the effect of variability in the stimulus spectrum.
TEOAEs at 80, 74, 68, and 62 dB pSPL evoking-stimulus levels were measured in 25 ears spanning a range of hearing levels from normal to mild hearing loss for a minimum of 10 test sessions. Reliability was assessed for 1/6-, 1/3- 1/2-, and 1-octave analyzing bandwidths.
Evoking-stimulus level, hearing loss, and center frequency did not significantly affect reliability. With decreasing analyzing bandwidth, reliability decreased. Intrasubject test-retest standard deviations were 1.2 dB for a broadband analyzing bandwidth and 1.4, 1.5, 1.6, and 1.8 dB for 1-, 1/2-, 1/3-, and 1/6-octave analyzing bandwidths, respectively. Stimulus variability within narrower bandwidths was to sufficient magnitude to influence test-retest reliability, and attempts to correct for the variations in stimulus spectrum were unsuccessful. Slopes of the input-output functions differed across frequencies, with shallower slopes at higher frequencies.
In general, TEOAE amplitude is highly reliable. For those individuals in this study who were more variable, the variability was at low frequencies or across the entire frequency spectrum. For clinical applications, the choice of analyzing bandwidth should be based on consideration of both frequency specificity (where narrow analyzing bandwidths are optimal) and reliability(where wide analyzing bandwidths are optimal).