Whole Stimulus DPOAE Analysis

Area of Science:

• Audiology
• Biomedical Engineering
• Signal Processing

Background:

• Distortion Product Otoacoustic Emissions (DPOAE) are crucial for hearing diagnosis but are complicated by interacting components.
• Current signal processing methods for separating DPOAE components are imperfect, often involving trade-offs between frequency precision and signal-to-noise ratio.

Purpose of the Study:

• To present a novel signal processing method for simultaneously estimating the entire spectra of distortion (D) and reflection (R) components in DPOAE.
• To overcome limitations of existing time-windowing methods and improve the accuracy of cochlear mechanism assessment.

Main Methods:

• The proposed method simultaneously estimates the full spectra of D and R components using a linear convex problem formulation.
• Assumptions include the sum of R and D measurements as convolutions with the stimulus, smoother frequency representation of D than R, and available noise level estimates.
• The method incorporates stimulus signal models to minimize interference and efficiently handles data with measurement artefacts.

Main Results:

• The new approach offers advantages including no compromise between frequency precision and signal-to-noise ratio.
• It effectively incorporates stimulus signal models and efficiently utilizes data corrupted by artefacts.
• The method shows improvement over existing techniques, particularly in reducing the contamination of D components by R components.

Conclusions:

• The developed method provides a more robust and accurate way to separate D and R components in DPOAE.
• This advancement has the potential to improve the diagnostic capabilities of DPOAE for assessing cochlear functionality.
• The method's adaptability allows for tracking DPOAE changes in response to various treatments.