Abstract— Sleep is a natural periodic state of rest for the body, in which the eyes usually close and consciousness is completely or partially lost. Consequently, there is a decrease in bodily movements and responsiveness to external stimuli. Slow wave sleep is of immense interest as it is the most restorative sleep stage during which the body recovers from weariness. During this sleep stage, electroencephalographic (EEG) and electro-oculographic (EOG) signals interfere with each other and they share a temporal similarity. In this investigation we used the EEG and EOG signals acquired from 10 patients undergoing overnight polysomnography with their sleep stages determined by certified sleep specialists based on RK rules. In this pilot study, we performed spectral estimation of EEG signals by Autoregressive (AR) modeling, and then used Itakura Distance to measure the degree of similarity between EEG and EOG signals. We finally calculated the statistics of the results and displayed them in an easy to visualize fashion to observe tendencies for each sleep stage. We found that Itakura Distance is the smallest for sleep stages 3 and 4. We intend to deploy this feature as an important element in automatic classification of sleep stages. Keywords— Sleep staging; EEG signal processing; EOG signal processing; feature extraction; AR modeling; biomedical signal processing; Itakura Distance; sleep apnea. I. INTRODUCTION Using the criteria of Rechtschaffen and Kales (RK), human sleep is usually divided into two phases called rapid eye movement (REM) and non-rapid eye movement (NREM) sleep on the basis of pattern changes in the electroencephalogram (EEG), eye movements (EOG) and chin electromyogram (EMG) [1]. NREM is subdivided into four stages that represent progressive stages in depth of sleep; referred to by S1 thorough S4. Wakefulness, characterized by beta and alpha frequencies, is replaced by S1 NREM at sleep onset when the majority of the EEG tracing contains theta (2-7 Hz) range activity and is usually followed within a few minutes by S2 when spindles and/or K-complexes occur. As S2 sleep progresses, high amplitude delta waves (<3 Hz and >75 micro-Volts) develop. When they occupy more than 20 to 50% of an epoch, S3 is scored and when the majority of the trace is delta activity, S4 is scored. Stages 3 and 4 are combined in the terms of slow wave (SWS) or delta sleep. By 90 minutes this cycle has reversed itself back to S2 and then there is a transition to REM sleep. It is associated with vivid dream reports, contains rapid conjugate eye movements and loss of muscle tone [2]. The EEG tracing is very similar to S1 (low voltage mixed frequencies) and there is an absence of spindles and K- complexes. Cycles of NREM and REM sleep continue across the night with 1/2-3/4 activity usually absent in the last half of the night and the amplitude of its delta waves diminish with age. Conjugate eye movements are present in wake and REM sleep but slow rolling eye movements are associated with sleep onset and are generally absent thereafter in NREM sleep. The “restorative theory of sleep” postulates that sleep in general and slow wave sleep in particular is associated with recovery from fatigue by restoring energy levels and reversing the tiring effects of wakefulness [3]. The high amplitude electrical waves are picked up in both EEG and EOG leads. Thus, there is some cross talk between the recording sites. Most of this cross talk involves delta waves of NREM sleep, REM during wakefulness, REM sleep, and k-complexes in S2 [4]. Figure 1 shows the similarity of the EOG and EEG epochs for a SWS stage (Stage 4). In contrast, Figure 2 shows the dissimilarity of EEG and EOG for non-SWS stage (Stage 1). II. MATERIALS AND METHODS TABLE I SUBJECTS INFORMATION Sleep Stages Subject 1 Subject 2 Subject 3 Subject 4 Subject 5 Subject 6 Subject 7 Subject 8 Subject 9 Subject 10 Awake epochs 185 55 121 329 80 62 87 101 188 82 Stage1 epochs 52 16 127 87 110 68 77 104 175 148 Stage2 epochs 314 484 391 282 342 474 503 446 345 399 Stage3 epochs 84 64 107 65 110 58 3 49 0 24 Stage4 epochs 128 3 63 52 81 0 0 2 0 0 REM epochs 86 65 100 107 123 111 131 101 138 94 Total (hrs) 7.08 5.73 7.58 7.68 7.05 6.44 6.68 6.69 7.05 6.23 Itakura Distance: A Useful Similarity Measure between EEG and EOG Signals in Computer-aided Classification of Sleep Stages E Estrada 1 , H Nazeran 1, 2 , P Nava 1 , K Behbehani * , J Burk ** , and E Lucas ** 1 Department of Electrical and Computer Engineering, The University of Texas at El Paso, El Paso, Texas, USA 2 School of Informatics and Engineering, Flinders University of South Australia, Adelaide, Australia * Biomedical Engineering Program, The University of Texas at Arlington/UTSWMC at Dallas, Arlington, Texas, USA ** Sleep Consultants Inc., Ft Worth, Texas, USA Proceedings of the 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference Shanghai, China, September 1-4, 2005 0-7803-8740-6/05/$20.00 ©2005 IEEE. 1189