Voxyvi: A system for long-term audio and video acquisitions in neonatal intensive care units.

BACKGROUND: In the European Union, 300,000 newborn babies are born prematurely every year. Their care is ensured in Neonatal Intensive Care Units (NICU) where vital signs are constantly monitored. In addition, other descriptors such as motion, facial and vocal activities have been shown to be essential to assess neurobehavioral development. AIM: In the scope of the European project Digi-NewB, we aimed to develop and evaluate a new audio-video device designed to non-invasively acquire multi-modal data (audio, video and thermal images), while fitting the wide variety of bedding environment in NICU. METHODS: Firstly, a multimodal system and associated software and guidelines to collect data in neonatal intensive care unit were proposed. Secondly, methods for post-evaluation of the acquisition phase were developed, including the study of clinician feedback and a qualitative analysis of the data. RESULTS: The deployment of 19 acquisition devices in six French hospitals allowed to record more than 500 newborns of different gestational and postmenstrual ages. After the acquisition phase, clinical feedback was mostly positive. In addition, quality of more than 300 recordings was inspected and showed that 77% of the data is exploitable. In depth, the percentage of sole presence of the newborn was estimated at 62% within recordings. CONCLUSIONS: This study demonstrates that audio-video acquisitions are feasible on a large scale in real life in NICU. The experience also allowed us to make a clear observation of the requirements and challenges that will have to be overcome in order to set up audio-video monitoring methods.

Video and audio processing in paediatrics: a review.

OBJECTIVE: Video and sound acquisition and processing technologies have seen great improvements in recent decades, with many applications in the biomedical area. The aim of this paper is to review the overall state of the art of advances within these topics in paediatrics and to evaluate their potential application for monitoring in the neonatal intensive care unit (NICU). APPROACH: For this purpose, more than 150 papers dealing with video and audio processing were reviewed. For both topics, clinical applications are described according to the considered cohorts-full-term newborns, infants and toddlers or preterm newborns. Then, processing methods are presented, in terms of data acquisition, feature extraction and characterization. MAIN RESULTS: The paper first focuses on the exploitation of video recordings; these began to be automatically processed in the 2000s and we show that they have mainly been used to characterize infant motion. Other applications, including respiration and heart rate estimation and facial analysis, are also presented. Audio processing is then reviewed, with a focus on the analysis of crying. The first studies in this field focused on induced-pain cries and the newest ones deal with spontaneous cries; the analyses are mainly based on frequency features. Then, some papers dealing with non-cry signals are also discussed. SIGNIFICANCE: Finally, we show that even if recent improvements in digital video and signal processing allow for increased automation of processing, the context of the NICU makes a fully automated analysis of long recordings problematic. A few proposals for overcoming some of the limitations are given.

Multi-feature classifiers for burst detection in single EEG channels from preterm infants.

OBJECTIVE: The study of electroencephalographic (EEG) bursts in preterm infants provides valuable information about maturation or prognostication after perinatal asphyxia. Over the last two decades, a number of works proposed algorithms to automatically detect EEG bursts in preterm infants, but they were designed for populations under 35 weeks of post menstrual age (PMA). However, as the brain activity evolves rapidly during postnatal life, these solutions might be under-performing with increasing PMA. In this work we focused on preterm infants reaching term ages (PMA ⩾36 weeks) using multi-feature classification on a single EEG channel. APPROACH: Five EEG burst detectors relying on different machine learning approaches were compared: logistic regression (LR), linear discriminant analysis (LDA), k-nearest neighbors (kNN), support vector machines (SVM) and thresholding (Th). Classifiers were trained by visually labeled EEG recordings from 14 very preterm infants (born after 28 weeks of gestation) with 36-41 weeks PMA. MAIN RESULTS: The most performing classifiers reached about 95% accuracy (kNN, SVM and LR) whereas Th obtained 84%. Compared to human-automatic agreements, LR provided the highest scores (Cohen's kappa = 0.71) using only three EEG features. Applying this classifier in an unlabeled database of 21 infants ⩾36 weeks PMA, we found that long EEG bursts and short inter-burst periods are characteristic of infants with the highest PMA and weights. SIGNIFICANCE: In view of these results, LR-based burst detection could be a suitable tool to study maturation in monitoring or portable devices using a single EEG channel.

Does Deep Bradycardia Increase the Risk of Arrhythmias and Syncope in Endurance Athletes?

The aim of this study was to evaluate whether endurance athletes who exhibit deep bradycardia are more prone to arrhythmias and reflex syncope than their non-bradycardic peers. 46 healthy men (ages 19-35) were divided into 3 groups based on whether they were sedentary (SED,<2 h/week) or endurance trained (ET,>6 h/week), and non-bradycardic (NB, resting heart rate (HR)≥60 bpm) or bradycardic (B, resting HR<50 bpm). Resting HR was lower in ETB vs. ETNB and SED (43.8±3.1, 61.3±3.3, 66.1±5.9 bpm, respectively; p<0.001). Thus, 16 SED, 13 ETNB and 17 ETB underwent resting echocardiography, maximal exercise test, tilt test (TT) and 24 h-Holter ECG. Subjects were followed-up during 4.7±1.1 years for training, syncope and cardiac events. Our results showed that incidence of arrhythmias and hypotensive susceptibility did not differ between groups. During follow-up, no episode of syncope or near-syncope was reported. However, cardio-inhibitory syncope occurrence tended to be higher in ETB. Left ventricular end-diastolic diameter index was increased in ETB vs. ETNB and was correlated with resting HR (r=- 0.64; p<0.001). As a result, athletes with deep bradycardia do not present more arrhythmias and more hypotensive susceptibility than their non-bradycardic peers. Cardiac enlargement and autonomic alteration both seem to be involved in an athlete's bradycardia.

Denoising preterm EEG by signal decomposition and adaptive filtering: a comparative study.

Electroencephalography (EEG) from preterm infant monitoring systems is usually contaminated by several sources of noise that have to be removed in order to correctly interpret signals and perform automated analysis reliably. Band-pass and adaptive filters (AF) continue to be systematically applied, but their efficacy may be decreased facing preterm EEG patterns such as the tracé alternant and slow delta-waves. In this paper, we propose the combination of EEG decomposition with AF to improve the overall denoising process. Using artificially contaminated signals from real EEGs, we compared the quality of filtered signals applying different decomposition techniques: the discrete wavelet transform, the empirical mode decomposition (EMD) and a recent improved version, the complete ensemble EMD with adaptive noise. Simulations demonstrate that introducing EMD-based techniques prior to AF can reduce up to 30% the root mean squared errors in denoised EEGs.

Kernel based support vector machine for the early detection of syncope during head-up tilt test.

This study aims to analyze the autonomic nervous system response during head-up tilt test (HUTT), by exploring the changes in dynamic properties of heart rate variability in subjects with and without syncopes, to predict the outcome of HUTT. Baroreflex response, as well as linear and non-linear parameters of RR-interval time series, have been extracted from the ECG of 66 subjects: 35 with and 31 without syncope during HUTT. The results show that, when considering the first 15 min of tilting position, the total power spectrum, the standard deviation, the long-term fractal scale of RR-interval and ΔRR-interval of time series increase, while the sample entropy decreases in the positive group compared to the negative one. These indices may be good predictors of positive response in patients with reflex syncope. Additionally, an analysis of the first 15 min of tilting position using kernel support vector machines leads to a correct classification of 85% of patients, within negative and positive response groups (specificity = 80.6% and sensitivity = 88.5%). In medical applications, it is important to avoid false negative diagnosis of syncopes during HUTT. Taking this into account, an overall accuracy of 72.1% can be obtained in the same window allowing the reduction of the examination time in the clinical domain.

Are electronic cardiac devices still evolving?

OBJECTIVES: The goal of this paper is to review some important issues occurring during the past year in Implantable devices. METHODS: First cardiac implantable device was proposed to maintain an adequate heart rate, either because the heart's natural pacemaker is not fast enough, or there is a block in the heart's electrical conduction system. During the last forty years, pacemakers have evolved considerably and become programmable and allow to configure specific patient optimum pacing modes. Various technological aspects (electrodes, connectors, algorithms diagnosis, therapies, ...) have been progressed and cardiac implants address several clinical applications: management of arrhythmias, cardioversion / defibrillation and cardiac resynchronization therapy. RESULTS: Observed progress was the miniaturization of device, increased longevity, coupled with efficient pacing functions, multisite pacing modes, leadless pacing and also a better recognition of supraventricular or ventricular tachycardia's in order to deliver appropriate therapy. Subcutaneous implant, new modes of stimulation (leadless implant or ultrasound lead), quadripolar lead and new sensor or new algorithm for the hemodynamic management are introduced and briefly described. Each times, the main result occurring during the two past years are underlined and repositioned from the history, remaining limitations are also addressed. CONCLUSION: Some important technological improvements were described. Nevertheless, news trends for the future are also considered in a specific session such as the remote follow-up of the patient or the treatment of heart failure by neuromodulation.

Respiratory and spontaneous arousals in patients with Sleep Apnea Hypopnea Syndrome.

Sleep in patients with Sleep Apnea-Hypopnea Syndrome (SAHS) is frequently interrupted with arousals. Increased amounts of arousals result in shortening total sleep time and repeated sleep-arousal change can result in sleep fragmentation. According to the American Sleep Disorders Association (ASDA) an arousal is a marker of sleep disruption representing a detrimental and harmful feature for sleep. The nature of arousals and its role on the regulation of the sleep process raises controversy and has sparked the debate in the last years. In this work, we analyzed and compared the EEG spectral content of respiratory and spontaneous arousals on a database of 45 SAHS subjects. A total of 3980 arousals (1996 respiratory and 1984 spontaneous) were analyzed. The results showed no differences between the spectral content of the two kinds of arousals. Our findings raise doubt as to whether these two kinds of arousals are truly triggered by different organic mechanisms. Furthermore, they may also challenge the current beliefs regarding the underestimation of the importance of spontaneous arousals and their contribution to sleep fragmentation in patients suffering from SAHS.

Analysis of the QRS complex for apnea-bradycardia characterization in preterm infants.

This work presents an analysis of the information content of new features derived from the electrocardiogram (ECG) for the characterization of apnea-bradycardia events in preterm infants. Automatic beat detection and segmentation methods have been adapted to the ECG signals from preterm infants, through the application of two evolutionary algorithms. ECG data acquired from 32 preterm infants with persistent apnea-bradycardia have been used for quantitative evaluation. The adaptation procedure led to an improved sensitivity and positive predictive value, and a reduced jitter for the detection of the R-wave, QRS onset, QRS offset, and iso-electric level. Additionally, time series representing the RR interval, R-wave amplitude and QRS duration, were automatically extracted for periods at rest, before, during and after apnea-bradycardia episodes. Significant variations (p<0.05) were observed for all time-series when comparing the difference between values at rest versus values just before the bradycardia event, with the difference between values at rest versus values during the bradycardia event. These results reveal changes in the R-wave amplitude and QRS duration, appearing at the onset and termination of apnea-bradycardia episodes, which could be potentially useful for the early detection and characterization of these episodes.

Application of multiple correspondence analysis to asses the relation between time after transplantation and sympathetic activity in cardiac transplant recipient.

The aim of the present study was to evidence that the heart reinnervation can occur and it is related with the time after transplantation (evolution with time). Data were evaluated using Multiple Correspondence Analyses (MCA), which is the ideal method to study the relation, probably nonlinear, between the Time After Transplantation (TAT) and the probable restoration of normal heart rate responses of sinus node regulated by the autonomic nervous system. Twenty four nonrejecting transplant recipients (60 +/- 48 months after transplantation) and nine healthy subjects were studied by heart rate variability parameters. Results showed that sympathetic activity is restored some time after transplantation. Until 48 months, the recent HTR are in direct correlation to low values SD and LF and for the oldest transplant recipient, these parameters are similar to that observed in normal subjects.

Clustering follow-up time-series recorded by cardiac implantable devices.

Follow-up of patients treated by cardiac resynchronization therapy (CRT) is of great interest to prevent health deterioration in the postoperative period. In this purpose, data recorded in implantable devices (ID) can be informative. They are large, multivariate, evolutive with time, and then difficult to interpret. This study proposes a methodology, based on Multiple Correspondence Analysis (MCA) and fuzzy coding, to i) reduce the dimensionality of these data, ii) characterize the factorial axes by the variables recorded in the ID, iii) identify evolutions of patients in the factorial plane which are related to health deterioration. The results show that the first three axes of the MCA contain more than 90% of the total variance and are linked to the functional state of the patients. A simple clustering of the patients is performed and defines an area of the factorial plane, correlated with a degradation of patients' clinical state.

LF/(LF+HF) index in ventricular repolarization variability correlated and uncorrelated with heart rate variability.

The purpose of this study, was to asses whether LF/(LF+HF) obtained from ventricular repolarization variability (VRV) reflects the state of sympathovagal balance. The VRV time series and heart rate variability (HRV) time series from seventy two electrocardiogram (ECG) records in four different autonomic nervous system (ANS) profiles (athletes, cardiac transplant patient, heart failure patients and normal subjects) were extracted. A dynamic linear parametric model was applied to separate the VRV in two parts, VRV correlated with HRV (VRV(r)) and VRV uncorrelated with HRV (VRV(u)). Spectral indices were obtained from HRV, VRV, VRV(u) and VRV(u) time series. Changes of these indicators from rest to tilt position were analyzed. Results showed that: i) only LF/(LF+HF) from HRV time series increases significantly from rest to tilt in all ANS profiles, this information could not be retrieved in the other three series (VRV, VRV (u) and VRV(u)) ii) LF/(LF+HF) index in HRV series are significantly different between normal subjects and heart failure patients, while cardiac transplant patients show a low coherence between HRV and VRV power spectra and iii) HF rhythm in VRV series seem to be related to the mechanical effect of respiration.

Evaluation of real-time QRS detection algorithms in variable contexts.

A method is presented to evaluate the detection performance of real-time QRS detection algorithms to propose a strategy for the adaptive selection of ORS detectors, in variable signal contexts. Signal contexts are defined as different combinations of QRS morphologies and clinical noise. Four QRS detectors are compared in these contexts by means of a multivariate analysis. This evaluation strategy is general and can be easily extended to a larger number of detectors. A set of morphology contexts, corresponding to eight QRS morphologies (normal, PVC, premature atrial beat, paced beat, LBBB, fusion, RBBB, junctional premature beat), was extracted from 17 standard ECG records. For each morphology context, the set of extracted beats, ranging from 30 to 23000, was resampled to generate 50 realisations of 20 concatenated beats. These realisations were then used as input to the QRS detectors, without noise, and with three different types of additive clinical noise (electrode motion artifact, muscle artifact, baseline wander) at three signal-to-noise ratios (5 dB, -5 dB, -15 dB). Performance was assessed by the number of errors, which reflected both false alarms and missed beats. The results show that the evaluated detectors are indeed complementary. For example, the Pan-Tompkins detector is the best in most contexts but the Okada detector generates fewer errors in the presence of electrode motion artifact. These results will be particularly useful to the development of a real-time system that will be able to choose the best ORS detector according to the current context.

A bond graph model of the cardiovascular system.

The study of the autonomic nervous system (ANS) function has shown to provide useful indicators for risk stratification and early detection on a variety of cardiovascular pathologies. However, data gathered during different tests of the ANS are difficult to analyse, mainly due to the complex mechanisms involved in the autonomic regulation of the cardiovascular system (CVS). Although model-based analysis of ANS data has been already proposed as a way to cope with this complexity, only a few models coupling the main elements involved have been presented in the literature. In this paper, a new model of the CVS, representing the ventricles, the circulatory system and the regulation of the CVS activity by the ANS, is presented. The models of the vascular system and the ventricular activity have been developed using the Bond Graph formalism, as it proposes a unified representation for all energetic domains, facilitating the integration of mechanic and hydraulic phenomena. In order to take into account the electro-mechanical behaviour of both ventricles, an electrophysiologic model of the cardiac action potential, represented by a set of ordinary differential equations, has been integrated. The short-term ANS regulation of heart rate, cardiac contractility and peripheral vasoconstriction is represented by means of continuous transfer functions. These models, represented in different continuous formalisms, are coupled by using a multi-formalism simulation library. Results are presented for two different autonomic tests, namely the Tilt Test and the Valsalva Manoeuvre, by comparing real and simulated signals.

Multi-formalism modelling and simulation: application to cardiac modelling.

Cardiovascular modelling has been a major research subject for the last decade. Different cardiac models have been developed at a cellular level as well as at the whole organ level. Most of these models are defined by a comprehensive cellular modelling using continuous formalisms or by a tissue-level modelling often based on discrete formalisms. Nevertheless, both views still suffer from difficulties that reduce their clinical applications: the first approach requires heavy computational resources while the second one is not able to reproduce certain pathologies. This paper presents an original methodology trying to gather advantages from both approaches, by means of a hybrid model mixing discrete and continuous formalisms. This method has been applied to define a hybrid model of cardiac action potential propagation on a 2D grid of endocardial cells, combining cellular automata and a set of cells defined by the Beeler-Reuter model. For simulations under physiological and ischemic conditions, results show that the action potential propagation as well as electrogram reconstructions are consistent with clinical diagnosis. Finally, the advantage of the proposed approach is discussed within the frame of cardiac modelling and simulation.

Temporal abstraction and inductive logic programming for arrhythmia recognition from electrocardiograms.

This paper proposes a novel approach to cardiac arrhythmia recognition from electrocardiograms (ECGs). ECGs record the electrical activity of the heart and are used to diagnose many heart disorders. The numerical ECG is first temporally abstracted into series of time-stamped events. Temporal abstraction makes use of artificial neural networks to extract interesting waves and their features from the input signals. A temporal reasoner called a chronicle recogniser processes such series in order to discover temporal patterns called chronicles which can be related to cardiac arrhythmias. Generally, it is difficult to elicit an accurate set of chronicles from a doctor. Thus, we propose to learn automatically from symbolic ECG examples the chronicles discriminating the arrhythmias belonging to some specific subset. Since temporal relationships are of major importance, inductive logic programming (ILP) is the tool of choice as it enables first-order relational learning. The approach has been evaluated on real ECGs taken from the MIT-BIH database. The performance of the different modules as well as the efficiency of the whole system is presented. The results are rather good and demonstrate that integrating numerical techniques for low level perception and symbolic techniques for high level classification is very valuable.

[Evaluation of pendulum testing of spasticity]. / Evaluation de la mesure de la spasticité par le pendulum test.

OBJECTIVES: To identify valid measurements of spasticity derived from the pendulum test of the leg in a representative population of spastic patients. MATERIAL AND METHODS: Pendulum testing was performed in 15 spastic and 10 matched healthy subjects. The reflex-mediated torque evoked in quadriceps femoris, as well as muscle mechanical parameters (viscosity and elasticity), were calculated using mathematical modelling. Correlation with the two main measures derived from the pendulum test reported in the literature (the Relaxation Index and the area under the curve) was calculated in order to select the most valid. RESULTS, DISCUSSION: Among mechanical parameters, only viscosity was found to be significantly higher in the spastic group. As expected, the computed integral of the reflex-mediated torque was found to be larger in spastics than in healthy subjects. A significant non-linear (logarithmic) correlation was found between the clinically-assessed muscle spasticity (Ashworth grading) and the computed reflex-mediated torque, emphasising the non-linear behaviour of this scale. Among measurements derived from the pendulum test which are proposed in the literature for routine estimation of spasticity, the Relaxation Index exhibited an unsuitable U-shaped pattern of variation with increasing reflex-mediated torque. On the opposite, the area under the curve revealed a linear regression, which is more convenient for routine estimation of spasticity. CONCLUSION: The pendulum test of the leg is a simple technique for the assessment of spastic hypertonia. However, the measurement generally used in the literature (the Relaxation Index) exhibits serious limitations, and would benefit to be replaced by more valid measures, such as the area under the goniometric curve, especially for the assessment of therapeutics.

A comprehensive model of spastic hypertonia derived from the pendulum test of the leg.

We propose a comprehensive model of spastic hypertonia based on clinical neurophysiology and validated using experimental data obtained from the pendulum test of the leg in 8 healthy volunteers and 15 spastic patients. This nonlinear computational model includes mechanical parameters and a stretch reflex representation involving three neural parameters: a threshold coefficient, the gain of the stretch reflex, and a time lag accounting for the reflex loop latency and the electromechanical coupling delay. Variation of the threshold coefficient alone allowed an overall reproduction of experimental data obtained from spastic and healthy subjects. We propose that this parameter could represent the supraspinal drive, supposed to be preserved in control subjects and decreased in spastic patients. No subsequent variation of the reflex gain was required to simulate spastic traces. Adjustment of the time lag influenced the duration of the swinging phase and oscillatory phenomena possibly occurring during the pendulum test. It could be related to the involvement of either short- or long-latency stretch reflex loops. With respect to current neurophysiological concepts of motor control, this modeling approach may help in understanding mechanisms underlying spastic hypertonia, and in predicting the clinical effect of antispasticity agents.

Real-time ECG transmission via Internet for nonclinical applications.

Telemedicine is producing a great impact in the monitoring of patients located in remote nonclinical environments such as homes, elder communities, gymnasiums, schools, remote military bases, ships, and the like. A number of applications, ranging from data collection, to chronic patient surveillance, and even to the control of therapeutic procedures, are being implemented in many parts of the world. As part of this growing trend, this paper discusses the problems in electrocardiogram (ECG) real-time data acquisition, transmission, and visualization over the Internet. ECG signals are transmitted in real time from a patient in a remote nonclinical environment to the specialist in a hospital or clinic using the current capabilities and availability of the Internet. A prototype system is composed of a portable data acquisition and preprocessing module connected to the computer in the remote site via its RS-232 port, a Java-based client-server platform, and software modules to handle communication protocols between data acquisition module and the patient's personal computer, and to handle client-server communication. The purpose of the system is the provision of extended monitoring for patients under drug therapy after infarction, data collection in some particular cases, remote consultation, and low-cost ECG monitoring for the elderly.