Electroencephalographic based hearing identification using back-propagation algorithm

• Published in: 2009 IEEE Toronto International Conference Science and Technology for Humanity pp. 991 - 995
• Main Authors: Sudirman, R, Seow, S C
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2009
• Subjects: Artificial intelligence; artificial intelligent; Artificial neural networks; Auditory system; back-propagation neural network; Biological neural networks; Electroencephalographic; Electroencephalography; Feedforward neural networks; Feedforward systems; Frequency; hearing threshold; Neural networks; Signal analysis
• ISBN: 9781424438778; 1424438772
• DOI: 10.1109/TIC-STH.2009.5444351

Electroencephalographic (EEG) based hearing identification using artificial intelligent is an application between human's cognitive ability (hearing), EEG technology and artificial intelligent. EEG signals which are produced when a subject listen to an audible sounds with particular frequency will be recorded using Neurofax EEG-9200 device for further analysis. The EEG signals are the sources for this research; used to train a 21 layers feed-forward back-propagation neural network (NN) in order to recognize the patterns of the brain wave. The EEG signals are analyzed using Fast Fourier Transform (FFT) and filtering techniques available in Matlab. Furthermore, the well trained network can recognise the brain signal effectively. A graphic user interface (GUI) has been developed to display the digitalised brain signal and identification result. The result showed that the NN algorithm was able to process the EEG data to identify the sound frequency perceived by the subjects.