Sleep Recording is important for clinical diagnosis and treatment of sleep disorders. Sleep staging is one of the most important steps in sleep analysis and is typically performed based on the characteristics of electrophysiological signals including EEG, EOG, and EMG. Normal healthy sleep consists of sequences of stages. According to the traditional Rechtschaffen & Kales (R&K) rules, these stages include: Awake, Light Sleep, Deep Sleep, and Rapid-Eye Movement (REM) Sleep. Our study develops a simple four-stage process to classify sleep into wakefulness, stage 1, stage 2, slow wave sleep (SWS) and rapid eye movement (REM) sleep based on single LEOG channel. To achieve this goal, this study first generates feature variables from LEOG signal. The proposed feature selection method is applied to select a subset of features to improve the accuracy of the classifier. By applying the proposed approach to 48727 distinct LEOG epochs that are obtained from 62 subjects, the accuracy rate is about 72.6%. The largest amount of errors occurs in the identification of Stage 1, 56.3% of which was misclassified into stages 2 or wake. The second largest error is associated with REM sleep, 23.7% of which was misclassified into stages 2.

One primary topic of sleep studies is the depth of sleep. According to definitions of R&K rules, human sleep can be roughly divided into three different stages: Awake, Non-rapid-eye-movement (NREM) Sleep, and Rapid-eye-movement (REM) Sleep. Moreover, sleep stages are scored mainly by EEG signals and complementally by EOG and EMG signals. Many researchers have indicated that diseases or disorders occur during sleep will affect life quality of patients. For example, REM sleep-related dyssomnia is highly correlated with neurodegenerative or mental disorders such as major depression. Furthermore, sleep apnea is one of the most common sleep disorders at present. Untreated sleep apnea can increase the risk of mental and cardiovascular diseases. This research proposes a detection method of REM sleep. Take into account the environment of homecare, we just extract and analyze EOG signals for the sake of convenience in comparison with EEG channels. By analyzing elementary waveforms of EOG signals based on VQ method, the proposed method performs a classification accuracy of 67.71% in a group application. The corresponding sensitivity and specificity are 73.38% and 68.95% respectively. In contrast, the average classification accuracy is 82.02% in personalized applications. And the corresponding average sensitivity and specificity are 83.05% and 81.62% respectively. Experimental results demonstrate the feasibility of detecting REM sleep via the proposed method, especially in personalized applications. This will be propitious to a long term tracing and research of personal sleep status.

Since studies have revealed sleeping quality is highly related to our health conditions, sleep-medicine has attracted more and more attention in recent years. Sleep staging is one of the most important elements of sleep-medicine. Traditionally, it¡¦s done by observing the information form of EEG, EOG and EMG signals. But this is almost not possible to achieve at home. Automatic detection of REM sleep is the main goal of this study. Via comparing the classification performances of different combinations of EEG, EOG and EMG signals, this study also tries to simplify the number of signal channels. By using features extracted from EEG, EOG and EMG signals, the back-propagation neural networks are used to distinguish REM and NREM sleep. By refining the outputs of the neural networks, this study extensively test the efficacy of the proposed approach by using databases from two different sleep centers. This work also investigates the influences of the number of signal channels, REM sleep ratio, AHI, and age on classification results. Data acquired from the sleep centers of China Medical University Hospital (CMUH) and Sheng-Mei Hospital are arranged in ten different groups. For our largest datasets, which consists of 1318 subjects from CMUH, the results show that the proposed method achieves 95.5% epoch-to-epoch agreement with Cohen's Kappa 0.833, sensitivity 85.9% and specificity 97.3%. The generalization accuracy is 94.1% with Cohen's Kappa 0.782, sensitivity 78.5% and specificity 97.3%.

Sleep staging can be used to assess whether sleep structure is abnormal. According to the R&K rule, human sleep can be divided into four different stages: Awake, Light Sleep, Deep Sleep and Rapid-Eye-Movement (REM) Sleep. Conventionally, sleep staging are scored mainly by EEG signals and complementally by EOG and EMG signals. The goal of this study is to detect slow wave sleep (SWS) automatically by using different combinations of EEG, EOG, and EMG signals. In particular, a total of 16 combinations of channels have been studied. Based on high amplitude slow wave characteristics of SWS, this study develops many of feature variables to characterize SWS. A subset of these features are employed to design neural network classifier to detect SWS. This study has noted interpersonal-differences in physiological signals between people and proposes solutions to this problem to improve the performance of SWS detection. The number of tested subjects from two different sleep centers is 1318 and 947 subjects, respectively. These subjects were divided into five groups for training and testing data in order to test performance of our proposed approach. By applying the proposed approach to 1318 subjects, the experimental results show that the proposed method achieves kappa of 0.63 by using a single EEG channel, kappa of 0.6 by using two channels EOG and kappa of 0.66 by using the best combination of multi-channel singals. The size of dataset used in this work is significantly large than those of previous studies and thus provide more reliable experimental results. The experimental results show that the proposed approach can provide satisfactory performance in dealing with dataset with more than 1000 subjects.

The rapid-eye-movement (REM) sleep is one of the most important parts in overnight sleep. In this study, an automatic REM sleep staging rule is introduced. Compared with the traditional REM detection method, a distinct feature of this method is that it only requires two EOG signals and thus reduces the number of input signal channels significantly. We calculate the correlation coefficient series between two EOG signals. By representing such a series with a VQ coding method, several techniques are proposed to improve the classification rate. Experimental results are given to demonstrate the effectiveness of the proposed approach.

One of the important measures for sleep quailty is sleep structure. Normal sleep consists of awake, rapid eye movement (REM) sleep and nonrapid eye movement (NREM) sleep states. NREM sleep can be further classified into stage 1, stage 2 and slow wave sleep (SWS). These stages can be analyzed quantitatively from various electrical signals such as the electroencephalogram (EEG), electro-oculogram (EOG), and electromyogram (EMG). The goal of this research is to develop a simple four-stage process to classify sleep into wake, REM, stage 1, stage 2 and SWS by using a single EEG channel. By applying the proposed approach to 48727 distinct epochs which are acquired from 62 persons, the experimental results show that the proposed method is achieves 76.98% of accuracy. The sensitivity and PPV for wake are 85.96% and 68.35%. Furthermore, the sensitivity and PPV for REM are 82.13% and 74.11%, respectively. The sensitivity and PPV for the stage 1 are 9.02% and 39.00%. The sensitivity and PPV for the stage 2 are 84.19% and 79.36%. The sensitivity and PPV for SWS are 81.53% and 85.40%.

Intracranial electroencephalography (iEEG) provides superior diagnostic and research benefits over non-invasive EEG in terms of spatial resolution and the level of electrophysiological detail. Post-operative Computed Tomography (CT) scans provide the precision in electrode localization required for clinical purposes; however, to use this data for basic sleep research the challenge lies in identifying the precise locations of the implanted electrodes’ recording sites in terms of neuroanatomical regions as well as reliable scoring of their sleep data without the aid of facial electrodes. While existing methods can be combined to determine their exact locations in three-dimensional space, they fail to identify the functionally relevant gray matter areas that lie closest to them, especially if the points lie in the white matter. We introduce an iterative sphere inflation algorithm in conjunction with a unified pipeline to detect the exact as well as nearest regions of interest for these recording sites. Next, for sleep scoring purposes, we establish differences observed in alpha band activity between wakefulness and rapid eye movement (REM) sleep in frontal and temporal regions of iEEG patients. Lastly, we implement an automated sleep scoring method relying on the variations in alpha and delta bands power during sleep which can be applied to large sets of iEEG data recorded without accompanying electrooculogram (EOG) and electromyogram (EMG) electrodes available across labs for use in studies pertaining to neural dynamics during sleep. / M.S. / Patients with epilepsy (a neurological disorder characterized by seizures) who do not respond to medication often undergo invasive monitoring of their brains’ electrical activity using intracranial electroencephalography (iEEG). iEEG requires a surgery in which electrodes are inserted directly into the patient’s brain for better measurements. While they are monitored, these patients offer a unique opportunity for research studies that investigate the role of sleep in various learning, memory mechanisms and other health-related areas. This is because the direct contact of the electrodes with the brain tissue provides far superior quality and resolution of brain activity data in comparison to non-invasive cap-based EEG that healthy subjects wear over their scalp. However, in order to derive meaningful conclusions from these invasive recordings, we must first know the exact areas of the brain from which each site records the electrical data. We must then be able to identify which stage of sleep the patient is in at any given point in time, to be able to successfully correlate specific sleep stage-related activity with our research objectives; these patients often lack the facial electrodes used for standard sleep scoring procedures. To solve the first problem, we present an electrode localization method along with an algorithm to determine which neighboring regions contribute most to a given site’s recorded data. For the second problem, we first establish a difference in the behavior of alpha waves in the brain between wakefulness and rapid eye movement (REM) sleep. Lastly, we present an automated method to classify sleep data into different stages based on the variation in alpha waves and delta waves found during sleep.

Made available in DSpace on 2014-12-17T15:36:56Z (GMT). No. of bitstreams: 1 AndreLHP_Capa_a_Pag78.pdf: 4907555 bytes, checksum: 6e6da787d601f791eab35a63d8938972 (MD5) Previous issue date: 2009-05-22 / Several lines of evidence indicate that sleep is beneficial for learning, but there is no experimental evidence yet that the content of dreams is adaptive, i.e., that dreams help the dreamer to cope with challenges of the following day. Our aim here is to investigate the role of dreams in the acquisition of a complex cognitive task. We investigated electroencephalographic recordings and dream reports of adult subjects exposed to a computer game comprising perceptual, motor, spatial, emotional and higher-level cognitive aspects (Doom). Subjects slept two nights in the sleep laboratory, a completely dark room with a comfortable bed and controlled temperature. Electroencephalographic recordings with 28 channels were continuously performed throughout the experiment to identify episodes of rapid-eye-movement (REM) sleep. Behaviors were continuously recorded in audio and video with an infrared camera. Dream reports were collected upon forced awakening from late REM sleep, and again in the morning after spontaneous awakening. On day 1, subjects were habituated to the sleep laboratory, no computer game was played, and negative controls for gamerelated dream reports were collected. On day 2, subjects played the computer game before and after sleep. Each game session lasted for an hour, and sleep for 7-9 hours. 9 different measures of performance indicated significant improve overnight. 81% of the subjects experienced intrusion of elements of the game into their dreams, including potentially adaptative strategies (insights). There was a linear correlation between performance and dream intrusion as well as for game improval and quantity of reported dreaming. In the electrophysiological analysis we mapped the subjects brain activities in different stages (SWS 1, REM 1, SWS 2, REM 2, Game 1 and Game 2), and found a modest reverberation in motor areas related to the joystick control during the sleep. When separated by gender, we found a significant difference on female subjects in the channels that indicate motor learning. Analysis of dream reports showed that the amount of gamerelated elements in dreams correlated with performance gains according to an inverted-U function analogous to the Yerkes-Dodson law that governs the relationship between arousal and learning. The results indicate that dreaming is an adaptive behavior / V?rias linhas de pesquisa indicam que o sono ? ben?fico para a aprendizagem, mas ainda n?o h? evid?ncias experimentais de que o conte?do dos sonhos ? adaptativo, isto ?, que os sonhos ajudam o sonhador a lidar com os desafios do dia seguinte. O objetivo deste trabalho ? investigar a rela??o dos sonhos como fator de aprendizagem e adapta??o a uma tarefa cognitiva complexa. Investigamos registros eletroencefalogr?ficos e relatos de sonhos de volunt?rios adultos expostos a um jogo de computador ("Doom") que envolve o aprendizado perceptual, motor, espacial e emocional, al?m de aspectos cognitivos de mais alta ordem ("insights"). Os volunt?rios dormiram duas noites no laborat?rio do sono, uma sala completamente escura com uma cama confort?vel e temperatura controlada. Registros polissonogr?ficos com 28 canais foram continuamente realizados durante todo o experimento para identificar epis?dios de sono REM, durante o qual prevalece a atividade on?rica. O comportamento dos volunt?rios foi continuamente registrado em ?udio e v?deo com uma c?mera infravermelha. Os relatos de sonho foram coletados mediante despertar for?ado de sono REM. No dia 1, os indiv?duos foram habituados ao laborat?rio de sono. No dia 2, os volunt?rios jogaram o jogo no computador antes e depois do sono. Cada sess?o do jogo durou 1 hora, e o sono durou 7-9 horas. Medidas de 9 tipos de desempenho mostraram melhora significativa ap?s uma noite de sono. 81% dos indiv?duos relataram intrus?o de elementos do jogo nos seus sonhos, incluindo estrat?gias potencialmente adaptativas (insight). Vimos uma correla??o linear entre os desempenhos e as intrus?es de elementos do jogo. Na an?lise eletrofisiol?gica do mapeamento cerebral de todos os sujeitos em estados espec?ficos (SWS 1, REM 1, SWS 2, REM 2, Jogo 1 e Jogo 2), verificamos uma modesta reverbera??o em ?reas motoras relacionadas com o controle do joystick durante o sono. Quando separamos os jogadores por sexo, verificamos no grupo de mulheres uma diferen?a significativa em canais que indicam este aprendizado motor. Os resultados indicam que sonhar pode ser um comportamento adaptativo

Dans cette thèse, un système personnalisé de stadification automatique du sommeil est proposé, combinant fusion symbolique et système de contrôle rétroactif. La fusion symbolique est inspirée par le processus décisionnel mis en œuvre par les cliniciens experts du sommeil lors la reconnaissance visuelle des stades de sommeil. Il commence par l'extraction de paramètres numériques à partir des signaux polysomnographiques bruts. L'interprétation symbolique de haut niveau se fait par l'intermédiaire de l'extraction de caractéristiques à partir des paramètres numériques. Enfin, la décision est générée en utilisant des règles inspirées par les recommandations internationales en médecine du sommeil. Les symboles et les valeurs des caractéristiques dépendent d'un ensemble de seuils, dont la détermination est une question clé. Dans cette thèse, deux algorithmes de recherche différents, Differential Evolution et Cross Entropy ont été étudiés pour calculer la valeur de ces seuils automatiquement. La variabilité individuelle a souvent été ignorée dans les systèmes automatiques de stadification du sommeil existants. Cependant, elle a été démontrée dans plusieurs travaux de recherche vis à vis de nombreux aspects du sommeil (comme les enregistrements polysomnographiques, les habitudes de sommeil, l'architecture du sommeil, la durée du sommeil, les événements liés au sommeil, etc.). Afin d'améliorer l'efficacité des classificateurs des stades de sommeil, un système automatisé de sommeil automatique adapté aux différentes personnes et tenant compte de la variabilité individuelle a été exploré et évalué. / In this thesis, a personalized automatic sleep staging system is proposed by combining symbolic fusion and feedback system control technique. Symbolic fusion is inspired by the decision-making process of clinical sleep staging. It starts from the extraction of digital parameters from raw polysomnography signals and it goes up to a high-level symbolic interpretation through a features extraction process. At last, the decision is generated using rules inspired by international guidelines in sleep medicine. Meanwhile, the symbols and the features computations depend on a set of thresholds, whose determination is a key issue. In this thesis, two different search algorithms, Differential Evolution and Cross Entropy, were studied to compute these thresholds automatically.Individual variability was often ignored in existing automatic sleep staging systems. However, an individual variability was observed in many aspects of sleep research (such as polysomnography recordings, sleep patterns, sleep architecture, sleep duration, sleep events, etc.). In order to improve the effectiveness of the sleep stages classifiers, a personalized automatic sleep staging system that can be adapted the different persons and take individual variability into consideration was explored and evaluated.The perspectives of this work are based on evaluating the complexity and the performances of these algorithms in terms of latencies and hardware resource requirements, in order to target a personalized automated embedded sleep staging system.