Comparison of different auto-detection methods for Wave V with wireless Automated Auditory Brainstem Response (AABR) measurement system

The wireless automatic auditory brainstem response measurement system developed in previous research uses Kalman filter with an exponential weight averaging method (Kalman filter with EWA) to filter signals and used the differential method to detect the Wave V of the ABR. However, the signals are too noisy to be accurately determined. Therefore, this study compares the Wavelet Kalman filter and Moving average to the Kalman filter with EWA, and the fitted parametric peak (FPP) to the differential method, respectively. The simulation results showed that the signal-to-noise ratio of ABR is the highest after Wavelet Kalman filtering, and the audiologist can mark the waves III and V of ABR the most. The latency of the wave V detected by FPP was used to calculate the Pearson product-moment correlation coefficient with the audiologist’s manual mark. The two latencies are highly correlated, and the correlation coefficient is higher than the differential method. Therefore, the Wavelet-Kalman filter and FPP algorithm were implemented in the AABR measurement system. To evaluate the accuracy of the algorithm in this study, subjective experiments were conducted with four normal hearing subjects. The Mann-Whitney U test was used to test the difference between the average value of automatic detection and reproducibility.