Adrien Ugon

K-Complex Detection Using a Hybrid-Synergic Machine Learning Method

Sleep stage identification is the first step in modern sleep disorder diagnostics process. K-complex is an indicator for the sleep stage 2. However, due to the ambiguity of the translation of the medical standards into a computer-based procedure, reliability of automated K-complex detection from the EEG wave is still far from expectation. More specifically, there are some significant barriers to the research of automatic K-complex detection. First, there is no adequate description of K-complex that makes it difficult to develop automatic detection algorithm. Second, human experts only provided the label for whether a whole EEG segment contains K-complex or not, rather than individual labels for each subsegment. These barriers render most pattern recognition algorithms inapplicable in detecting K-complex. In this paper, we attempt to address these two challenges, by designing a new feature extraction method that can transform visual features of the EEG wave with any length into mathematical representation and proposing a hybrid-synergic machine learning method to build a K-complex classifier. The tenfold cross-validation results indicate that both the accuracy and the precision of this proposed model are at least as good as a human expert in K-complex detection.

Convex optimisation-based methods for K-complex detection

We develop three convex optimisation-based models for automatic detection of K-complexes.They extract key features of an EEG signal (a biological application).They significantly reduce the dimension of the problem and the computational time.They enhance the classification accuracy of an EEG signal in presence of K-complex.K-complexes are successfully detected in an EEG background. K-complex is a special type of electroencephalogram (EEG, brain activity) waveform that is used in sleep stage scoring. An automated detection of K-complexes is a desirable component of sleep stage monitoring. This automation is difficult due to the ambiguity of the scoring rules, complexity and extreme size of data. We develop three convex optimisation models that extract key features of EEG signals. These features are essential for detecting K-complexes. Our models are based on approximation of the original signals by sine functions with piecewise polynomial amplitudes. Then, the parameters of the corresponding approximations (rather than raw data) are used to detect the presence of K-complexes. The proposed approach significantly reduces the dimension of the classification problem (by extracting essential features) and the computational time while the classification accuracy is improved in most cases. Numerical results show that these models are efficient for detecting K-complexes.

Using visual analytics for presenting comparative information on new drugs

OBJECTIVE When a new drug is marketed, physicians must decide whether they will consider it for their future practice. However, information about new drugs can be biased or hard to find. In this work, our objective was to study whether visual analytics could be used for comparing drug properties such as contraindications and adverse effects, and whether this visual comparison can help physicians to forge their own well-founded opinions about a new drug. MATERIALS AND METHODS First, an ontology for comparative drug information was designed, based on the expectations expressed during focus groups comprised of physicians. Second, a prototype of a visual drug comparator website was developed. It implements several visualization methods: rainbow boxes (a new technique for overlapping set visualization), dynamic tables, bar charts and icons. Third, the website was evaluated by 22 GPs for four new drugs. We recorded the general satisfaction, the physicians decision whether to consider the new drug for future prescription, both before and after consulting the website, and their arguments to justify their choice. RESULTS The prototype website permits the visual comparison of up to 10 drugs, including efficacy, contraindications, interactions, adverse effects, prices, dosage regimens,…All physicians found that the website allowed them to forge a well-founded opinion on the four new drugs. The physicians changed their decision about using a new drug in their future practice in 29 cases (out of 88) after consulting the website. DISCUSSION AND CONCLUSION Visual analytics is a promising approach for presenting drug information and for comparing drugs. The visual comparison of drug properties allows physicians to forge their opinions on drugs. Since drug properties are available in reference texts, reviewed by public health agencies, it could contribute to the independent of drug information.

On Personalization in IoT

In recent years, a large number of connected objects for health and well-being have emerged. The attractiveness of these objects for several categories of people and their decreasing cost make it an interesting opportunity especially in well-being and health applications. Despite the popularity of these objects, many people still have some concerns about their use due to various considerations such as lack of security of the collected data, %the feeling of being watched, lack of personalization, particularly considering preferences, constraints and needs of the user, etc. In this paper we first briefly survey IoT oriented towards health and well-being. Then we highlight in particular some hot-issues regarding personalization. We propose a model for IoT-based intelligent and personalized system for improving sleep quality.

How to use symbolic fusion to support the sleep apnea syndrome diagnosis

The Sleep Apnea Syndrome is a sleep disorder characterized by frequently repeated respiratory disorders during sleep. It needs the simultaneous recording of many physiological parameters to be diagnosed. The analysis of these curves is a time consuming task made by sleep Physicians. First, they detect some physiological events on each curve and then, they point out links between respiratory events and their consequences. To support the diagnosis, we used symbolic fusion on elementary events, which connects events to their sleep context - sleep-stage and body position - and to the respiratory event responsible of their occurrence. The reference indicator is the Apnea-Hypopnea Index (AHI), defined as the average hourly frequency of arisen of Apneas or Hypopneas while the patient is sleeping. We worked on the polysomnography of 59 patients, that were first completely analyzed by a sleep Physician and then analyzed by our method. We compared the ratio of the AHI got by the automatic analysis and the AHI got by the sleep Physician. δ = AHI(automaticanalysis)/AHI(SleepPhysicianAnalysis) Globally, we overvalued the count of apneas and hypopneas for the group of patients with AHI ≤ 5, that are considered as healthy patients. In average, for these patients, δ = 2, 71. For patients with mild or moderate Sleep Apnea Syndrome we globally found a similar AHI. In average, for these patients, δ = 1, 04. For patients with severe Sleep Apnea Syndrome, we undervalued a little the count of respiratory events. In average, for these patients, δ = 0, 83. This leads to the same severity class for most of the patients.

Decision System Integrating Preferences to Support Sleep Staging.

Scoring sleep stages can be considered as a classification problem. Once the whole recording segmented into 30-seconds epochs, features, extracted from raw signals, are typically injected into machine learning algorithms in order to build a model able to assign a sleep stage, trying to mimic what experts have done on the training set. Such approaches ignore the advances in sleep medicine, in which guidelines have been published by the AASM, providing definitions and rules that should be followed to score sleep stages. In addition, these approaches are not able to solve conflict situations, in which criteria of different sleep stages are met. This work proposes a novel approach based on AASM guidelines. Rules are formalized integrating, for some of them, preferences allowing to support decision in conflict situations. Applied to a doubtful epoch, our approach has taken the appropriate decision.

Cross entropy-based automatic thresholds setting-up method for sleep staging system

Sleep staging is a fundamental step in diagnosis and treatment of sleep disorders. In current sleep staging systems, normally a set of thresholds should be set up to determine the boundaries in differentiating different linguistic or symbolic features. However, as far as we know, there are no fully satisfying automatic method to do this task. Thresholds are mostly set up manually. In this paper, an automatic thresholds setting-up method based on Cross Entropy is proposed. Person-dependent thresholds can be provided automatically by using Cross Entropy and used in personalized sleep staging analysis while considering individual variability. The feasibility of Cross Entropy has also been evaluated, computational results exhibit that the Cross Entropy-based method is an efficient, convenient and applicable stochastic method for automatically setting-up thresholds in sleep staging system. Compared with manual method, average F-Measures are improved more than 10% for all the stages and up-to 74% for stage N3 by using proposed method.

Evaluating optisas, a visual method to analyse sleep apnea syndromes

The Sleep apnea syndrome is a real public health problem. Improving its diagnosis using the polysomnography is of huge importance. Optisas was a visual method allowing translating the polysomnographic data into a meaningful image. In a previous paper, it was shown to bring extrainformation in 62% of cases. Here its capacity for displaying information of the same relevance as the one got using the classical report of the polysomnography is studied. The main result is that this capacity is weak and seems to be present only to identify the obstructive sleep apnea syndrome. Moreover this study suggests to improve the standardization of the classical report in the framework of a quality insurance process.

On the Role of Similarity in Analogical Transfer

Analogical transfer consists in making the assumption that if two situations are alike in some respect, they may be alike in others. This article explores the links that exist between analogical transfer and the qualitative measurement of differences. The main idea is to formulate the similarity principle as a dependency between two measurements of difference. Analogical transfer is formulated as a similarity-based reasoning: it is plausible that equally different pairs in a certain dimension are also equally different in another dimension, at least for pairs that are not too (analogically) dissimilar.

Thermal-Signature-Based Sleep Analysis Sensor

This paper addresses the development of a new technique in the sleep analysis domain. Sleep is defined as a periodic physiological state during which vigilance is suspended and reactivity to external stimulations diminished. We sleep on average between six and nine hours per night and our sleep is composed of four to six cycles of about 90 min each. Each of these cycles is composed of a succession of several stages of sleep that vary in depth. Analysis of sleep is usually done via polysomnography. This examination consists of recording, among other things, electrical cerebral activity by electroencephalography (EEG), ocular movements by electrooculography (EOG), and chin muscle tone by electromyography (EMG). Recordings are made mostly in a hospital, more specifically in a service for monitoring the pathologies related to sleep. The readings are then interpreted manually by an expert to generate a hypnogram, a curve showing the succession of sleep stages during the night in 30s epochs. The proposed method is based on the follow-up of the thermal signature that makes it possible to classify the activity into three classes: “awakening,” “calm sleep,” and “restless sleep”. The contribution of this non-invasive method is part of the screening of sleep disorders, to be validated by a more complete analysis of the sleep. The measure provided by this new system, based on temperature monitoring (patient and ambient), aims to be integrated into the tele-medicine platform developed within the framework of the Smart-EEG project by the SYEL–SYstemes ELectroniques team. Analysis of the data collected during the first surveys carried out with this method showed a correlation between thermal signature and activity during sleep. The advantage of this method lies in its simplicity and the possibility of carrying out measurements of activity during sleep and without direct contact with the patient at home or hospitals.