Design Of Brainstem Auditory Evoked Potential Detection System Based On Wavelet Information Entropy

Brainstem auditory evoked potential signal is a series of potential changes performance between the head and the mastoid process about 10 ms when the auditory system stimulated by specific sound. The signal has lower amplitude and SNR. This method is widely used in clinical practice, such as hearing test and hearing screening. For the extraction algorithm of brainstem auditory evoked potential signal, superposed average method is mainly applied in the clinical application, but there are a disadvantage for stimulation times and cause subjects fatigue response, therefore, this paper mainly studies the single dynamic extraction algorithm of the BAEP signal. And design a cost-effective Brainstem auditory evoked potential signal detector according to this algorithm.This study compared the advantages and disadvantages in clinical applications of superposition average algorithm and wavelet analysis algorithm to extract brainstem auditory evoked potential signal, this paper proposes a wavelet entropy of brainstem auditory evoked potential extraction algorithm. Wavelet information entropy method maintaining the advantage of wavelet analysis, and combining with the theory of information entropy, using the different of brainstem auditory evoked potential signal and background noise in the wavelet domain statistical characteristics, through the threshold algorithm as far as possible to remove background noise components and retains the useful signal component, which compared to the traditional wavelet analysis method, wavelet information entropy add a weighted factor to improve threshold function, and according to the wavelet information entropy to determine the weighting factor, in order to obtain a better effect. In order to confirm the validity and accuracy of the theory that wavelet information entropy extracts brainstem auditory evoked potentials, the syntheticsignal simulation experiment and clinical trial are designed, using the method of superposition average, wavelet analysis and wavelet entropy to extract the synthetic brainstem auditory evoked potentials and the clinical brainstem auditory evoked potentials.The experimental results show that compared to the superposition averaging method, wavelet entropy reduces the number of tests, and completes a single dynamic extraction of evoked potential signal, at the same time compared to the wavelet analysis method, the algorithm retains the signal of feature point information(latency and amplitude)better. Meanwhile the effect of restraining signal suppression is better, the signal waveform is smooth, which verifies the validity and accuracy of the wavelet information entropy method proposed in this paper.