Methods to reduce stimulus artifact in the detection of transient evoked otoacoustic emissions (TEOAEs)

This paper introduces four methods to reduce the stimulus artifact in the detection of TEOAEs based on the derived nonlinear response (DNLR) method. The details and performances of these methods, i.e., the DNLR combined with window function, bi-directional filtering (BF), predictor-subtraction-restorer filter (PSR) and discrete wavelet transform (DWT), are discussed and compared in this work. The analysis of the real TEOAEs signal shows that these proposed methods not only eliminate artifact more effectively at a higher stimulus level, but also expand time region of the TEOAEs uncontaminated by artifact. Furthermore, they help to capture more reliably the TEOAEs with higher-frequency and shorter-latency overlapping with artifact and reserve more the information which is useful in the clinical diagnosis. The efficiency and limitation of each method are discussed