Department of Audiology Faculty Articles

Effects of Filtering and Re-Windowing on TEOAEs

ORCID

0000-0001-6770-4377

Publisher

Wolters Kluwer Health, Inc.

ISSN

2333-6218

Publication Date

9-2005

Keywords

audiology, hearing loss, transient-evoked otoacoustic emissions (TEOAEs)

Abstract

The purpose of this study was to compare two TEOAE protocols. The default ILO protocol (FullScreen) uses a 20-msec window and a 50-Hz stimulation rate. The QuickScreen protocol uses a 12.5-msec window, an 80-Hz stimulation rate, and activation of an 800-Hz high-pass filter. Previous comparisons of the two protocols have attributed differences in test performance to window length or a combination of window length and stimulation rate. This study was conducted to examine the contribution of filtering, re-windowing, and the combined effects of filtering and re-windowing in TEOAE test performance.

The research design was a retrospective analysis of clinical outcomes. Electronic waveforms using a stimulus level of 80 dB ±3, a minimum of 50 samples, and a stability of 70% or greater were considered valid. Valid TEOAEs were matched to clinical records in which same-day tympanograms and monaural pure-tone thresholds were available. We collected data for 240 ears of 135 participants of the FullScreen protocol, and for 84 ears of 49 participants for the QuickScreen protocol. FullScreen TEOAEs were (1) filtered, (2) re-windowed, and (3) filtered and re-windowed for comparative purposes. Approximately 20%-25% of each group consisted of ears with hearing loss. TEOAE response level and noise level were compared with the “gold standard” of pure-tone thresholds at three octave frequencies (1000, 2000, and 4000 Hz).

FullScreen data were subjected to a one-way ANOVA for repeated measures to determine if the manipulated conditions were significantly different from the original recording. Results revealed a significant difference at 1000 and 4000 Hz for TEOAE response level and a significant difference at all three frequencies for TEOAE noise level. We compared FullScreen filtered and re-windowed data with the QuickScreen data in a one-way MANOVA. There was a significant difference at 4000 Hz for TEOAE response level and a significant difference in noise level for all frequencies.

Post hoc analysis of the FullScreen data revealed that activation of the 800-Hz filter was most effective in improving the SNR at 1000 Hz. Post hoc re-windowing was equally effective across the three frequencies. Comparisons between the two protocols suggest that the differences in test performance are due in part to a reduction in the upward spread of masking.

Results support the idea that post hoc analysis of TEOAE FullScreen waveforms should be explored as a diagnostic application of TEOAE evaluation. Future research should focus on same-subject comparisons of the filtered and re-windowed FullScreen protocol with the QuickScreen protocol. Since test performance has been demonstrated to be superior for the FullScreen TEOAE protocol, automated post hoc filtering and re-windowing should be explored as an option to further enhance diagnostic applications.

DOI

10.1097/01.HJ.0000292836.44494.29

Volume

57

Issue

9

First Page

12

Last Page

12

Disciplines

Communication Sciences and Disorders | Medicine and Health Sciences | Speech and Hearing Science | Speech Pathology and Audiology

Link to Full Text

Share

COinS