Auditory Evoked Potentials Based Detection of Hearing Threshold Using Pso Neural Network

Hearing loss has been the most prevalent sensory disability throughout the world. In Malaysia, National and Hearing Disorder Survey has reported that the prevalence of hearing disorder is about 17.14% with an estimated population of 28.8 million and 1.5% newborns have the onset of hearing impairment. A conventional hearing screening test’s applicability is limited as it requires a feedback response from the subject under test. In this paper, a simple method based on auditory evoked potential (AEP) is proposed to classify the hearing level of a subject. AEP signals are unilaterally recorded with monaural acoustical stimulus from the normal hearing and abnormal hearing subjects and the correlation between the brain dynamics and the hearing threshold response has been observed. From the recorded AEP signals, independent power spectral features (delta, theta, alpha, beta, and gamma) are extracted. A particle swarm optimization (PSO) learning algorithm is applied to a dynamic feedback neural network and a neural network model was developed to associate features with the normal and abnormal hearing levels. The PSO neural network employed with the independent spectral gamma band features has a maximum classification accuracy of 83.45% and 85.77% for the left and right ears across 20 subjects. This study indicates that the gamma band has a higher power value for hearing loss participants while compared with the normal hearing participants. The gamma power establishes a specific sensory code and exhibits a high sensitivity level to discriminate the normal and abnormal hearing level.