High-resolution analysis of transient evoked otoacoustic emissions

A new decomposition and time-frequency method applied to the estimation of transient evoked otoacoustic emissions (TEOAEs) is presented. TEOAEs are low-level signals in response to auditory stimuli that are used to monitor the functionality of the cochlea. TEOAE parameter estimation is challenging due to the short duration of the signal (about 20ms) and its highly nonstationary nature. Existing time-frequency methods are incapable of providing accurate estimates of the components within the TEOAE signal. The proposed decomposition and time-frequency method is based on the effective tracking of nonstationary sinusoidal components using a nonlinear iterative sinusoid tracking algorithm. The presented technique improves upon existing techniques by offering a high time-frequency resolution and is capable of providing accurate estimates of components within a TEOAE signal. The structure of the proposed decomposition technique as well as the underlying algorithm are presented. Simulation results are presented to show the performance of the proposed technique. Its performance is also demonstrated on a clinically recorded signal.