Denoising of Electroretinogram signals using empirical mode decomposition

Abstract

Clinical Electrophysiologic tests derived from human eyes are the tests that use to review whole visual pathways and they are important for ophthalmology and neuro ophthalmology. Electroretinographies is one of the electrophysiological tests often used to investigate the electrical response of the retinal layers from retinal pigment epithelium up to the occipital cortex. ERG signals have two important amplitudes that are used to diagnose diseases by doctors. These are negative a wave and positive b wave. Implicit times of the a and b waves are also meaningful to diagnose. ERG signals have small amplitudes (about µV). Because of this reason it is significant to separate the signal from the noise and interference that occurs as a result of movement. In this study, we propose using a new technique, called the empirical mode decomposition to denoised ERG responses. The Empirical Mode Decomposition is a signal processing method for analyzing nonlinear and nonstationary signals. ERG signals which are nonstationary signals are decomposed into a series of Intrinsic Mode Functions and then noise and interference are eliminated. Finally ERG signals which have signal to noise ratio less or equal than 10dB are reconstructed. As a result we successfully obtained denoised ERG signals.