Periodogram based tests for distortion product otoacoustic emissions

Area of Science:

• Audiology
• Otoacoustic Emissions
• Hearing Science

Background:

• Distortion product otoacoustic emissions (DPOAEs) are a nonbehavioral measure of cochlear function, analogous to pure-tone audiograms.
• DPOAEs arise from nonlinearities in the cochlea and are detected using periodogram-based tests in the Fourier domain.
• Previous statistical analysis of DPOAE tests relied on receiver operator characteristics curves.

Purpose of the Study:

• To explicitly derive the null distribution of the order-statistic based DPOAE test.
• To evaluate the power of the DPOAE test using simulations mimicking real-ear noise.
• To compare the DPOAE test's power and computational ease with an alternative statistical test.

Main Methods:

• Explicit derivation of the null distribution for the order-statistic based DPOAE test.
• Simulations were conducted to mimic nonwhite, real-ear noise characteristics.
• The power of the DPOAE test was compared against a local F test.

Main Results:

• The null distribution of the DPOAE test was explicitly derived.
• Simulations demonstrated the power of the DPOAE test under realistic noise conditions.
• A local F test was found to be more powerful and computationally simpler than the DPOAE test.
• Test power increases with more bins, but improvement is negligible beyond four bins.
• Test power decreases at lower distortion product frequencies.

Conclusions:

• The local F test offers a more powerful and practical alternative for analyzing DPOAE data.
• Careful consideration of the number of bins and DP frequency is necessary to optimize test power.
• Findings contribute to a better understanding of statistical methods for evaluating cochlear function via DPOAEs.