Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name



Speech-Language-Hearing Sciences


Glenis R Long

Committee Members

Brett A Martin

Valarie Shafer

Subject Categories

Speech and Hearing Science


Stimulus frequency otoacoustic emissions, cochlear mechanics, traveling waves, phase, hearing thresholds


Otoacoustic emissions (OAEs) are sounds that originate in the cochlea and are measured in the ear canal. OAEs provide a noninvasive tool for investigating cochlear mechanics. Stimulus-frequency OAEs (SFOAEs) are evoked by presenting a single frequency tone, called a probe tone, which have an advantage over other OAEs because they are the least influenced by cochlear nonlinearities. However, because the SFOAE are generated in the cochlea with the same frequency as the stimulus, additional techniques, such as the use of suppressor tones are needed to enable separation of the probe tone from the SFOAE.

The primary goal of this investigation was to explore individual differences in SFOAE phase gradient delays. These delays were hypothesized to improve estimates of cochlear health, inferred from hearing thresholds. Efficient measures of phase gradient delays can be obtained using frequency swept tones analyzed with time-frequency filtering, such as the least squares (LS) fit. The least squares fit is a time-frequency filter because the LS fit estimates coefficients for a subset of the total signal which are then used to separate and estimate signals of interest. However, the limitations of the frequency swept tone procedure and LS fit for estimating SFOAEs are not well understood. This investigation first focused on identifying limitations of such SFOAE and refining the LS fitting procedure. It was determined that including a suppressor was necessary for obtaining optimal SFOAE estimates, and the investigation shifted from further refining the LS fitting procedure to exploration of alternative time-frequency analyses which permit clearer characterization of the various latency contributions to suppressor based SFOAEs estimates. The use of a fast, continuous filtered wavelet transform provided a unique perspective on the distribution of SFOAE energy in the time-frequency domain and confirmed that SFOAEs are a sum of both long and short latency contributions. The distributions of long and short SFOAE energy explain some the discrepancies between discrete tone and swept tone SFOAEs procedures.

Predicting behavioral thresholds from SFOAE phase, magnitude, or phase and magnitude combined may be misleading when the analysis is not focused around the SFOAE latency contributions from the region where SFOAEs are most affected by cochlear damage. It was revealed that more focus should be given to understanding the best ways to separate the long and the short latencies for different stimulus parameters and individuals, in order to improve sensitivity to cochlear health.