Modified cyclic shift tree denoising technique with fewer number of sweep for wave V detection

Abstract

Nowadays, in developing countries Newborn Hearing Screening (NHS) has become one of the most important recommendations in modern pediatric audiology due to the important of early detection for newborn as the first six month of age are the critical period for learning communication. Auditory Brainstem Response (ABR) is an electrophysiological response in the electroencephalography generated in the brainstem in response to the acoustical stimulus. The conventional method used previously was accurate, but it is time consuming especially with the presence of noise interference. The objective of this research is to reduce screening time by implementing enhanced signal processing method and also to reduce the influence of noise interference. This thesis applies Wavelet Kalman Filter (WKF), Cyclic Shift Tree Denoising (CSTD) and Modified Cyclic Shift Tree Denoising (MCSTD) to overcome these problems. The modified approach MSCTD is a modification from CSTD where it is a combination of the wavelet, KF and CSTD. The modified approach was compared to the averaging, WKF and CSTD to analyze an effective wavelet method for denoising that can give the rapid and accurate extraction of ABRs. Results show that the MCSTD outperform the other methods and giving the highest SNR value and able to detect wave V until reduce sweeps number of 512 and 1024 respectively for chirp and click stimulus.