International Journal of Computer Applications (0975 – 8887) Volume 57– No.9, November 2012 35 EOG Artifact Correction from EEG Signals for Biomedical Analysis Suhas S. Patil Associate Professor, Department of Electronics, K.B.P. College of Engineering, Satara. Minal K. Pawar P.G. Student Department of Electronics Sinhgad College of Engineering, Pune ABSTRACT The electroencephalogram records the electrical activity of the brain and is the main resource of information for studying neurological disorders. Corruption of EEG signal is caused by occurrence of various artifacts like line interference, electro- oculogram, electrocardiogram, and muscle activity. These artifacts increase the difficulty in analyzing the EEG and obtaining clinical information. The ocular artifact detection and correction from EEG is of considerable significance for both the automatic and visual analysis of brainwave activity by neurologists for proper diagnosis. In this paper, a statistical method for removing ocular artifacts from EEG recordings through thresholding and correlation is proposed. EEG database of 325 samples from Colorado state university is used for experimentation. The mean, variance, standard deviation, and correlation are the performance metrics used. The results show that the proposed method significantly detects and removes the EOG and line frequency artifact without loss of important part of original EEG. General Terms Signal processing, Biomedical analysis. Keywords Electroencephalogram (EEG), artifacts, Electro-oculogram (EOG), correlation. 1. INTRODUCTION The human body is filled of electrical signals, which can be picked up and analyzed. Significance of the human body has never decreased and research on it has never been stopped since hundreds years ago. Among all physiological signals, EEG signals are accepted and productive in the application of mental state detection of a person. The EEG was first measured in humans by Hans Berger in 1929. EEG refers to the recording of brains spontaneous electrical activity over a short time period. Electrical impulses produced by nerve firings in the brain diffuse through the head and can be measured by electrodes positioned on the scalp. The EEG provides a coarse analysis of neural activity and has been utilized to non-invasively study cognitive processes and the composition of the brain. EEG utilized by neurologists for diagnosis of diseases such as, schizophrenia, Alzheimer, dementia and apnea. For proper diagnosis of any disease, analysis of the EEG signals should be correct. For appropriate analysis one must eliminate the noise due to facial muscle movements, eye blinking etc. [1] [2] The Electrical activity of the brain has amplitude in the microvolt range. In a practice, EEG is measured at the scalp using a surface electrode. Due to the variability of impedance and the potential for transmission of infectious disease, type electrode is no longer normally used. The most frequent electrode presently used is a gold-plated disc 10 mm in diameter. Electrode positions and names are specified by the International 10–20 system for most medical and research applications. This system assures that the naming of electrodes is consistent across laboratories. In most clinical applications, 19 recording electrodes (plus ground and system reference) are used. A smaller number of electrodes are typically used when recording EEG from neonates. The 10-20 system shown in Figure 1 was proposed to standardize the collection of EEG. The reference electrodes are positioned on non-active sites such as forehead or earlobes. EEG electrodes are placed on scalp. The recorded signal is achieved by subtracting signal measured below the eye from one measured above the eye. The data is stored up at 250 samples per seconds and digitized with 12 bits of precision sets of 10 sec as shown in figure 2. An illustration of four seconds of EEG data as shown in Figure 2, recorded at 250 samples per second from seven sites. The two spikes are the result of eye blinks on the recorded data. The electrical fluctuations originated by brain activity and recorded as EEG are generally in the range of – 50 to 50 μV. Eye blinks have higher amplitude and often have voltages of over 100 μV. [1] The seven sites of Electrodes are labeled according to the International 10-20 system. Eye Blink: The eye blink signal is very frequent in EEG data (see Figure 3). It creates a high amplitude signal that can be several times greater than the EEG signals of importance. Because of its high amplitude an eye blink can corrupt data on all electrodes, even those at the back of the head, these are often measured more directly in the EOG, pairs of electrodes positioned above and around the eyes. Unfortunately, these measurements are contaminated with EEG signals of interest and so simple subtraction is not a removal option even if a precise model of EOG diffusion across the scalp is exist. These noise, or unnecessary signals produced by patient, sources comprise: line noise from the power grid, eye blinks, eye movements, heartbeat, breathing, and other muscle activity. Some of these, such as eye blinks, produce voltage alterations of much higher amplitude than the endogenous brain activity. In this condition the data must be rejected unless that signals which are produced by patient can be removed from the data. 2. LITERATURE REVIEW A number of methods has been proposed for EOG artifact correction from EEG. Due to the limitation of EEG signal recording technology, physiological artifacts, especially those generated by eye (EOG) with EEG, may change the characteristics of the neurological event in EEG. The ocular artifact correction based on Linear Combination and