Title: Analyzing transient-evoked otoacoustic emissions by concentration of frequency and time

Abstract: The linear part of transient evoked (TE) otoacoustic emissions (OAEs) are thought to be generated via coherent reflection near the characteristic place of constituent wave components. Because of the tonotopic organization of the cochlea, high frequency emissions should return earlier than low frequencies; however, due to the random nature of coherent reflection, the frequency-latency relation is monotonic only when averaged across many ears. For individual ears, the instantaneous frequency (IF) and amplitude envelope of TEOAEs both fluctuate and are difficult to analyze via linear transformations. In this paper, we propose to analyze TEOAE signals by a nonlinear, synchrosqueezing-based technique called concentration of frequency and time (ConceFT). In our simulation, the present method is able to capture the IF function at near 0 dB signal-to-noise ratio. Validity of the present method is discussed in the context of coherent reflection based on its formulation in both the time-frequency and the space-wavenumber domains.