Otoacoustic emissions and improved pass/fail separation using wavelet analysis and time windowing

Area of Science:

• Audiology
• Signal Processing
• Biomedical Engineering

Background:

• Transient evoked otoacoustic emissions (TEOAEs) are crucial for hearing screening.
• Current TEOAE screening methods face challenges in accurately separating pass/fail results.
• Improving the specificity of TEOAE screening is essential for reliable hearing assessment.

Purpose of the Study:

• To develop and evaluate a novel method for enhancing pass/fail discrimination in TEOAE hearing screening.
• To improve the specificity of TEOAE screening while maintaining high sensitivity.
• To provide a more accurate assessment of hearing levels using TEOAE signals.

Main Methods:

• Signal decomposition using discrete wavelet transform.
• Application of non-linear denoising techniques.
• Scale-dependent time windowing of TEOAE signals.
• Utilizing the cross-correlation coefficient between processed TEOAE signals as a criterion.
• Assessing performance using receiver operating characteristics (ROC) analysis.

Main Results:

• The proposed method significantly improves specificity compared to the conventional wave reproducibility parameter.
• Specificity increased from 68% to 83% at a 90% sensitivity level.
• Performance was validated on a large database of 5,214 individuals.

Conclusions:

• The novel signal processing method offers superior performance for TEOAE hearing screening.
• This approach enhances the reliability and accuracy of hearing loss detection in screening programs.
• The improved specificity reduces false positives, leading to more efficient audiological referrals.