Early detection of late-onset neonatal sepsis from noninvasive biosignals using deep learning: A multicenter prospective development and validation study

BACKGROUND

Neonatal sepsis is responsible for significant morbidity and mortality worldwide. Its accurate and timely diagnosis is hindered by vague symptoms and the urgent necessity for early antibiotic intervention. The gold standard for diagnosing the condition is the identification of a pathogenic organism from normally sterile sites via laboratory testing. However, this method is resource-intensive and cannot be conducted continuously.

OBJECTIVE

This study aimed to predict the onset of late-onset sepsis (LOS) with good diagnostic value as early as possible using non-invasive biosignal measurements from neonatal intensive care unit (NICU) monitors.

METHODS

In this prospective multicenter study, we developed a multimodal machine learning algorithm based on a convolutional neural network (CNN) structure that uses the power spectral density (PSD) of recorded biosignals to predict the onset of LOS. This approach aimed to discern LOS-related pathogenic spectral signatures without labor-intensive manual artifact removal.

RESULTS

The model achieved an area under the receiver operating characteristic score of 0.810 (95 % CI 0.698-0.922) on the validation dataset. With an optimal operating point, LOS detection had 83 % sensitivity and 73 % specificity. The median early detection was 44 h before clinical suspicion. The results highlighted the additive importance of electrocardiogram and respiratory impedance (RESP) signals in improving predictive accuracy. According to a more detailed analysis, the predictive power arose from the morphology of the electrocardiogram's R-wave and sudden changes in the RESP signal.

CONCLUSION

Raw biosignals from NICU monitors, in conjunction with PSD transformation, as input to the CNN, can provide state-of-the-art prediction performance for LOS without the need for artifact removal. To the knowledge of the authors, this is the first study to highlight the independent and additive predictive potential of electrocardiogram R-wave morphology and concurrent, sudden changes in the RESP waveform in predicting the onset of LOS using non-invasive biosignals.

Validity of Ultra-Short-Term HRV Analysis Using PPG-A Preliminary Study

Continuous measurement of heart rate variability (HRV) in the short and ultra-short-term using wearable devices allows monitoring of physiological status and prevention of diseases. This study aims to evaluate the agreement of HRV features between a commercial device (Bora Band, Biosency) measuring photoplethysmography (PPG) and reference electrocardiography (ECG) and to assess the validity of ultra-short-term HRV as a surrogate for short-term HRV features. PPG and ECG recordings were acquired from 5 healthy subjects over 18 nights in total. HRV features include time-domain, frequency-domain, nonlinear, and visibility graph features and are extracted from 5 min 30 s and 1 min 30 s duration PPG recordings. The extracted features are compared with reference features of 5 min 30 s duration ECG recordings using repeated-measures correlation, Bland-Altman plots with 95% limits of agreements, Cliff's delta, and an equivalence test. Results showed agreement between PPG recordings and ECG reference recordings for 37 out of 48 HRV features in short-term durations. Sixteen of the forty-eight HRV features were valid and retained very strong correlations, negligible to small bias, with statistical equivalence in the ultra-short recordings (1 min 30 s). The current study concludes that the Bora Band provides valid and reliable measurement of HRV features in short and ultra-short duration recordings.

Evolution of Heart Rate Complexity Indices in the Early Detection of Neonatal Sepsis

Despite advances in prenatal health care, neonatal sepsis remains a major cause of neonatal mortality. Early diagnosis and adequate treatment are essential to reduce morbidity and mortality related to this disease. In this paper, we propose a new method to detect neonatal sepsis based on heart rate (HR) complexity measures (entropy and compression indices) that takes into consideration neonatal gestational age. First, the percentile curves were computed for all the complexity indices using data from 118 control neonates. Eight indices were computed: the sample entropy (SampEn) and three indices to quantify the multiscale entropy (MSE) curve - the sum, the slope, and the product of the previous two - and the compression ratio (CR), using the bzip2 compressor, as well as the same three indices but related to the multiscale compression (MSC) curve. Then, the corresponding percentile was estimated for 23 sepsis neonates. Results show a significant decrease in the entropy indices SampEn and MSEsum and in the MSCslope a day before the detection of sepsis by the clinicians. The indices CR and MSCsum increased before the antibiotic take. These results imply that sepsis causes a random, uncorrelated pattern on the HR signal. Future studies should include a bigger data set to calculate a compound index comprising information of other physiological signals. Clinical Relevance - Prompt and accurate diagnosis of neona-tal sepsis is essential for the successful clinical management of neonates and significantly reduce morbidity and mortality. Complexity measures applied to the HR time series appear to detect sepsis in neonates starting one day before the clinical detection.

Extraction of Premature Newborns' Spontaneous Cries in the Real Context of Neonatal Intensive Care Units

Cry analysis is an important tool to evaluate the development of preterm infants. However, the context of Neonatal Intensive Care Units is challenging, since a wide variety of sounds can occur (e.g., alarms and adult voices). In this paper, a method to extract cries is proposed. It is based on an initial segmentation between silence and sound events, followed by feature extraction on the resulting audio segments and a cry and non-cry classification. A database of 198 cry events coming from 21 newborns and 439 non-cry events was created. Then, a set of features—including Mel-Frequency Cepstral Coefficients—issued from principal component analysis, was computed to describe each audio segment. For the first time in cry analysis, noise was handled using harmonic plus noise analysis. Several machine learning models have been compared. The K-Nearest Neighbours approach showed the best results with a precision of 92.9%. To test the approach in a monitoring application, 412 h of recordings were automatically processed. The cries automatically selected were replayed and a precision of 92.2% was obtained. The impact of errors on the fundamental frequency characterisation was also studied. Results show that despite a difficult context, automatic cry extraction for non-invasive monitoring of vocal development of preterm infants is achievable.

Functional age estimation through neonatal motion characterization using continuous video recordings

The follow-up of the development of the premature baby is a major component of its clinical care since it has been shown that it can reveal a pathology. However, no method allowing an automated and continuous monitoring of this development has been proposed. Within the framework of the Digi-NewB European project, our team wishes to offer new clinical indices qualifying the maturation of newborns. In this study, we propose a new method to characterize motor activity from video recordings. For this purpose, we have chosen to characterize the motion temporal organization by drawing inspiration from sleep organization. Thus, we propose a fully automatic process allowing to extract motion features and to combine them to estimate a functional age. By investigating two datasets, one of 28.5 hours (manually annotated) from 33 newborns and one of 4,920 hours from 46 newborns, we show that the proposed approach is relevant for monitoring in clinical routine and that the extracted features reflect the maturation of preterm newborns. Indeed, a compact and interpretable model using gestational age and three motion features (mean duration of intervals with motion, total percentage of time spent in motion and number of intervals without motion) was designed to predict post-menstrual age of newborns and showed an admissible mean absolute error of 1.3 weeks. While the temporal organization of motion was not studied clinically due to a lack of technological means, these results open the door to new developments, new investigations and new knowledge on the evolution of motion in newborns.

Proposal for a Home Sleep Monitoring Platform Employing a Smart Glove

The present paper proposes the design of a sleep monitoring platform. It consists of an entire sleep monitoring system based on a smart glove sensor called UpNEA worn during the night for signals acquisition, a mobile application, and a remote server called AeneA for cloud computing. UpNEA acquires a 3-axis accelerometer signal, a photoplethysmography (PPG), and a peripheral oxygen saturation (SpO2) signal from the index finger. Overnight recordings are sent from the hardware to a mobile application and then transferred to AeneA. After cloud computing, the results are shown in a web application, accessible for the user and the clinician. The AeneA sleep monitoring activity performs different tasks: sleep stages classification and oxygen desaturation assessment; heart rate and respiration rate estimation; tachycardia, bradycardia, atrial fibrillation, and premature ventricular contraction detection; and apnea and hypopnea identification and classification. The PPG breathing rate estimation algorithm showed an absolute median error of 0.5 breaths per minute for the 32 s window and 0.2 for the 64 s window. The apnea and hypopnea detection algorithm showed an accuracy (Acc) of 75.1%, by windowing the PPG in one-minute segments. The classification task revealed 92.6% Acc in separating central from obstructive apnea, 83.7% in separating central apnea from central hypopnea and 82.7% in separating obstructive apnea from obstructive hypopnea. The novelty of the integrated algorithms and the top-notch cloud computing products deployed, encourage the production of the proposed solution for home sleep monitoring.

Detection of changes in the behaviour of the elderly person

In this paper, we propose a solution for detecting changes in the behaviour of the elderly person based on the monitoring of activities of daily living (ADL). The elderly person's daily routine is characterized by the following five indexes: 1) percentage of time lying down, 2) percentage of time sitting, 3) percentage of time standing, 4) percentage of time absent from home, and 5) number of falls during the day. In our framework, these indexes are computed using characteristics extracted from depth and thermal data. We hypothesize that elderly persons have a well-defined, regular life routine, organized around their environment, habits, and social relations. Then, given the indexes values, a day is defined as routine or non-routine day. Thus, looking for changes of day type allows to detect changes in a person's routine. The method has been tested on a database of depth and thermal images recorded in a nursing home over an 85 days period. These tests proved the reliability of the proposed method.

Interest of the minimum edit distance to detect behaviour change of the elderly person

In this article, a solution to detect the change of behaviour of the elderly person based on the person's activities of daily living is proposed. This work is based on the hypothesis that the person attaches importance to a rhythmic sequence of days and activities per day. The day of the elderly person is described by a succession of activities, and each activity is associated to a posture (lying down, sitting, standing, absent). Postures are estimated from image analysis measured by thermal or depth cameras in order to preserve the anonymity of the person. The change in posture succession is calculated using the minimum edit distance with respect to the routine day. The number of permutations/inversions reflects the change in the person's behaviour. The method was tested on two elderly persons recorded by thermal and depth cameras during 85 days in a retirement home. It is shown that for a person with a life change behaviour, the average number of permutations and interquartile range, before and after changes, are 41 [28], [48] and 57 [55-62] respectively compared to the learned routine day. The Wilcoxon test confirmed the significant difference between these two periods.Clinical Relevance- Monitoring the daily routine provides indicators for detecting changes in the behaviour of an elderly person.

Criteria for assessing the quality of clinical practice guidelines in paediatrics and neonatology: a mixed-method study

Background

Evidenced-based practice is a key component of quality care. This study aims to explore users’ expectations concerning paediatric local clinical practice guidelines.

Methods

A mixed method approach was applied, including material from quantitative questionnaire and semi-structured interviews. Data were analysed using descriptive statistics and qualitative content analysis. Data were analysed with constant comparative method. Qualitative data were parsed and categorized to identify themes related to decision-making.

Results

A total of 83 physicians answered the survey (response rate 83%). 98% of the participants wanted protocols based on international guidelines, 80% expected a therapeutic content. 24 semi-structured interviews were conducted to understand implementation processes, barriers and facilitators. Qualitative analysis revealed 5 emerging themes: improvement of local clinical practice guidelines, patterns of usage, reasons for non-implementation, alternative sources and perspectives.

Conclusion

Some criteria should be considered for the redaction of local clinical practice guidelines: focus on therapeutic, ease of access, establish local clinical practice guidelines based on international guidelines adapted to the local setting, document references and include trainees such as residents in the redaction.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12911-021-01628-1.

Evaluation of maturation in preterm infants through an ensemble machine learning algorithm using physiological signals

This study was designed to test if heart rate variability (HRV) data from preterm and full-term infants could be used to estimate their functional maturational age (FMA), using a machine learning model. We propose that the FMA, and its deviation from the postmenstrual age (PMA) of the infants could inform physicians about the progress of the maturation of the infants. The HRV data was acquired from 50 healthy infants, born between 25 and 41 weeks of gestational age, who did not present any signs of abnormal maturation relative to their age group during the period of observation. The HRV features were used as input for a machine learning model that uses filtering and genetic algorithms for feature selection, and an ensemble machine learning (EML) algorithm, which combines linear and random forest regressions, to produce as output a FMA. Using HRV data, the FMA had a mean absolute error of 0.93 weeks, 95% CI [0.78, 1.08], compared to the PMA. These results demonstrate that HRV features of newborn infants can be used by an EML model to estimate their FMA. This method was also generalized using respiration rate variability (RRV) and bradycardia data, obtaining similar results. The FMA, predicted either by HRV, RRV or bradycardia, and its deviation from the true PMA of the infants, could be used as a surrogate measure of the maturational age of the infants, which could potentially be monitored non-invasively and in real-time in the setting of neonatal intensive care units.

Comparison of Activity Monitors Accuracy in Assessing Intermittent Outdoor Walking

PURPOSE

This study aimed to determine and compare the accuracy of different activity monitors in assessing intermittent outdoor walking in both healthy and clinical populations through the development and validation of processing methodologies.

METHODS

In study 1, an automated algorithm was implemented and tested for the detection of short (≤1 min) walking and stopping bouts during prescribed walking protocols performed by healthy subjects in environments with low and high levels of obstruction. The following parameters obtained from activity monitors were tested, with different recording epochs0.1s/0.033s/1s/3s/10s and wearing locationsscapula/hip/wrist/ankle: GlobalSat DG100 (GS) and Qstarz BT-Q1000XT/-Q1000eX (QS) speed; ActiGraph wGT3X+ (AG) vector magnitude (VM) raw data, VM counts, and steps; and StepWatch3 (SW) steps. Furthermore, linear mixed models were developed to estimate walking speeds and distances from the monitors parameters. Study 2 validated the performance of the activity monitors and processing methodologies in a clinical population showing profile of intermittent walking due to functional limitations during outdoor walking sessions.

RESULTS

In study 1, GS1s, scapula, QS1s, scapula/wrist speed, and AG0.033s, hip VM raw data provided the highest bout detection rates (>96.7%) and the lowest root mean square errors in speed (≤0.4 km·h-1) and distance (<18 m) estimation. Using SW3s, ankle steps, the root mean square error for walking/stopping duration estimation reached 13.6 min using proprietary software and 0.98 min using our algorithm (total recording duration, 282 min). In study 2, using AG0.033s, hip VM raw data, the bout detection rate (95% confidence interval) reached 100% (99%-100%), and the mean (SD) absolute percentage errors in speed and distance estimation were 9% (6.6%) and 12.5% (7.9%), respectively.

CONCLUSIONS

GPS receivers and AG demonstrated high performance in assessing intermittent outdoor walking in both healthy and clinical populations.

Early bradycardia detection and therapeutic interventions in preterm infant monitoring

In very preterm infants, cardio-respiratory events and associated hypoxemia occurring during early postnatal life have been associated with risks of retinopathy, growth alteration and neurodevelopment impairment. These events are commonly detected by continuous cardio-respiratory monitoring in neonatal intensive care units (NICU), through the associated bradycardia. NICU nurse interventions are mainly triggered by these alarms. In this work, we acquired data from 52 preterm infants during NICU monitoring, in order to propose an early bradycardia detector which is based on a decentralized fusion of three detectors. The main objective is to improve automatic detection under real-life conditions without altering performance with respect to that of a monitor commonly used in NICU. We used heart rate lower than 80 bpm during at least 10 sec to define bradycardia. With this definition we observed a high rate of false alarms (64%) in real-life and that 29% of the relevant alarms were not followed by manual interventions. Concerning the proposed detection method, when compared to current monitors, it provided a significant decrease of the detection delay of 2.9 seconds, without alteration of the sensitivity (97.6% vs 95.2%) and false alarm rate (63.7% vs 64.1%). We expect that such an early detection will improve the response of the newborn to the intervention and allow for the development of new automatic therapeutic strategies which could complement manual intervention and decrease the sepsis risk.

Preterm Newborn Presence Detection in Incubator and Open Bed Using Deep Transfer Learning

Video-based motion analysis recently appeared to be a promising approach in neonatal intensive care units for monitoring the state of preterm newborns since it is contact-less and noninvasive. However it is important to remove periods when the newborn is absent or an adult is present from the analysis. In this paper, we propose a method for automatic detection of preterm newborn presence in incubator and open bed. We learn a specific model for each bed type as the camera placement differs a lot and the encountered situations are different between both. We break the problem down into two binary classifications based on deep transfer learning that are fused afterwards: newborn presence detection on the one hand and adult presence detection on the other hand. Moreover, we adopt a strategy of decision intervals fusion in order to take advantage of temporal consistency. We test three deep neural network that were pre-trained on ImageNet: VGG16, MobileNetV2 and InceptionV3. Two classifiers are compared: support vector machine and a small neural network. Our experiments are conducted on a database of 120 newborns. The whole method is evaluated on a subset of 25 newborns including 66 days of video recordings. In incubator, we reach a balanced accuracy of 86%. In open bed, the performance is lower because of a much wider variety of situations whereas less data are available.

Early Detection of Late Onset Sepsis in Premature Infants Using Visibility Graph Analysis of Heart Rate Variability

OBJECTIVE

This study was designed to test the diagnostic value of visibility graph features derived from the heart rate time series to predict late onset sepsis (LOS) in preterm infants using machine learning.

METHODS

The heart rate variability (HRV) data was acquired from 49 premature newborns hospitalized in neonatal intensive care units (NICU). The LOS group consisted of patients who received more than five days of antibiotics, at least 72 hours after birth. The control group consisted of infants who did not receive antibiotics. HRV features in the days prior to the start of antibiotics (LOS group) or in a randomly selected period (control group) were compared against a baseline value calculated during a calibration period. After automatic feature selection, four machine learning algorithms were trained. All the tests were done using two variants of the feature set: one only included traditional HRV features, and the other additionally included visibility graph features. Performance was studied using area under the receiver operating characteristics curve (AUROC).

RESULTS

The best performance for detecting LOS was obtained with logistic regression, using the feature set including visibility graph features, with AUROC of 87.7% during the six hours preceding the start of antibiotics, and with predictive potential (AUROC above 70%) as early as 42 h before start of antibiotics.

CONCLUSION

These results demonstrate the usefulness of introducing visibility graph indexes in HRV analysis for sepsis prediction in newborns.

SIGNIFICANCE

The method proposed the possibility of non-invasive, real-time monitoring of risk of LOS in a NICU setting.

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.

Modeling Stress-Recovery Status Through Heart Rate Changes Along a Cycling Grand Tour

Background

Heart rate (HR) and HR variability (HRV) indices are established tools to detect abnormal recovery status in athletes. A low HR and vagally mediated HRV index change between supine and standing positions reflected a maladaptive training stress-recovery status.

Objectives

Our study was focused on a female multistage cycling event. Its overall aim was twofold: (1) quantify the correlation between (a) the change in HR and HRV indices during an active orthostatic test and (b) subjective/objective fatigue, physical load, and training level indicators; and (2) formulate a model predicting the stress-recovery status as indexed by Δ⁢R⁢R¯ and ΔLnRMSSD (defined as the difference between standing and supine mean RR intervals and LnRMSSD, respectively), based on subjective/objective fatigue indicators, physical load, and training levels.

Methods

Ten female cyclists traveled the route of the 2017 Tour de France, comprising 21 stages of 200 km on average. From 4 days before the beginning of the event itself, and until 1 day after its completion, every morning, each cyclist was subjected to HR and HRV measurements, first at rest in a supine position and then in a standing position. The correlation between HR and HRV indices and subjective/objective fatigue, physical load, and training level indicators was then computed. Finally, several multivariable linear models were tested to analyze the relationships between HR and HRV indices, fatigue, workload, and training level indicators.

Results

HR changes appeared as a reliable indicator of stress-recovery status. Fatigue, training level, and Δ⁢R⁢R¯ displayed a linear relationship. Among a large number of linear models tested, the best one to predict stress-recovery status was the following: Δ⁢R⁢R¯=1,249.37+12.32V̇O_2max + 0.36 km⋅week^–1−8.83 HR_max−5.8 RPE−28.41 perceived fatigue with an adjusted R² = 0.322.

Conclusion

The proposed model can help to directly assess the adaptation status of an athlete from RR measurements and thus to anticipate a decrease in performance due to fatigue, particularly during a multistage endurance event.

Smartphone-Based Clinical Pathways in Pediatrics: A Case Study

Evidence-based medicine is a major evolution in the way medical practice and reasoning are structured. This approach aims at guiding patient care through rigorous, explicit and judicious evidences. In this contribution, we present the case study of the deployment of a smartphone-based system to manage clinical pathways and its impact during two years in the pediatric department of the university hospital of Rennes, France. We also tackle smartphone acceptability and easiness of use by pediatricians.

Detecting falls and estimation of daily habits with depth images using machine learning algorithms

Different approaches have been proposed in the literature to detect the fall of an elderly person. In this paper, we propose a fall detection method based on the classification of parameters extracted from depth images. Three supervised learning methods are compared: decision tree, K-Nearest Neighbors (K-NN) and Random Forests (RF). The methods have been tested on a database of depth images recorded in a nursing home over a period of 43 days. The Random Forests based method yields the best results, achieving 93% sensitivity and 100% specificity when we restrict our study around the bed. Furthermore, this paper also proposes a 37 days follow-up of the person, to try and estimate his or her daily habits.

Detection and Classification of Sleep Apnea and Hypopnea Using PPG and SpO 2 Signals

In this work, a detection and classification method for sleep apnea and hypopnea, using photopletysmography (PPG) and peripheral oxygen saturation (SpO ₂) signals, is proposed. The detector consists of two parts: one that detects reductions in amplitude fluctuation of PPG (DAP)and one that detects oxygen desaturations. To further differentiate among sleep disordered breathing events (SDBE), the pulse rate variability (PRV) was extracted from the PPG signal, and then used to extract features that enhance the sympatho-vagal arousals during apneas and hypopneas. A classification was performed to discriminate between central and obstructive events, apneas and hypopneas. The algorithms were tested on 96 overnight signals recorded at the UZ Leuven hospital, annotated by clinical experts, and from patients without any kind of co-morbidity. An accuracy of 75.1% for the detection of apneas and hypopneas, in one-minute segments,was reached. The classification of the detected events showed 92.6% accuracy in separating central from obstructive apnea, 83.7% for central apnea and central hypopnea and 82.7% for obstructive apnea and obstructive hypopnea. The low implementation cost showed a potential for the proposed method of being used as screening device, in ambulatory scenarios.

Deep transfer learning for video-based detection of newborn presence in incubator

Preterm newborns are prone to late-onset sepsis, leading to a high risk of mortality. Video-based analysis of motion is a promising non-invasive approach because the behavior of the newborn is related to his physiological state. But it is needed to analyze only images where the newborn is solely present in incubator. In this context, we propose a method for video-based detection of newborn presence. We use deep transfer learning: bottleneck features are extracted from a pre-trained deep neural network and then a classifier is trained with these features on our database. Moreover, we propose a strategy that allows to take advantage of temporal consistency. On a database of 11 newborns with 56 days of video recordings, the results show a balanced accuracy of 80%.

Cardiorespiratory Alterations in a Newborn Ovine Model of Systemic Inflammation Induced by Lipopolysaccharide Injection

Although it is well known that neonatal sepsis can induce important alterations in cardiorespiratory control, their detailed early features and the mechanisms involved remain poorly understood. As a first step in resolving this issue, the main goal of this study was to characterize these alterations more extensively by setting up a full-term newborn lamb model of systemic inflammation using lipopolysaccharide (LPS) injection. Two 6-h polysomnographic recordings were performed on two consecutive days on eight full-term lambs: the first after an IV saline injection (control condition, CTRL); the second, after an IV injection of 2.5 μg/kg Escherichia coli LPS 0127:B8 (LPS condition). Rectal temperature, locomotor activity, state of alertness, arterial blood gases, respiratory frequency and heart rate, mean arterial blood pressure, apneas and cardiac decelerations, and heart-rate and respiratory-rate variability (HRV and RRV) were assessed. LPS injection decreased locomotor activity (p = 0.03) and active wakefulness (p = 0.01) compared to the CTRL. In addition, LPS injection led to a biphasic increase in rectal temperature (p = 0.01 at ∼30 and 180 min) and in respiratory frequency and heart rate (p = 0.0005 and 0.005, respectively), and to an increase in cardiac decelerations (p = 0.05). An overall decrease in HRV and RRV was also observed. Interestingly, the novel analysis of the representations of the horizontal and vertical visibility network yielded the most statistically significant alterations in HRV structure, suggesting its potential clinical importance for providing an earlier diagnosis of neonatal bacterial sepsis. A second goal was to assess whether the reflexivity of the autonomic nervous system was altered after LPS injection by studying the cardiorespiratory components of the laryngeal and pulmonary chemoreflexes. No difference was found. Lastly, preliminary results provide proof of principle that brainstem inflammation (increased IL-8 and TNF-α mRNA expression) can be shown 6 h after LPS injection. In conclusion, this full-term lamb model of systemic inflammation reproduces several important aspects of neonatal bacterial sepsis and paves the way for studies in preterm lambs aiming to assess both the effect of prematurity and the central neural mechanisms of cardiorespiratory control alterations observed during neonatal sepsis.

Robust, real-time generic detector based on a multi-feature probabilistic method

Robust, real-time event detection from physiological signals acquired during long-term ambulatory monitoring still represents a major challenge for highly-artifacted signals. In this paper, we propose an original and generic multi-feature probabilistic detector (MFPD) and apply it to real-time QRS complex detection under noisy conditions. The MFPD method calculates a binary Bayesian probability for each derived feature and makes a centralized fusion, using the Kullback-Leibler divergence. The method is evaluated on two ECG databases: 1) the MIT-BIH arrhythmia database from Physionet containing clean ECG signals, 2) a benchmark noisy database created by adding noise recordings of the MIT-BIH noise stress test database, also from Physionet, to the MIT-BIH arrhythmia database. Results are compared with a well-known wavelet-based detector, and two recently published detectors: one based on spatiotemporal characteristic of the QRS complex and the second, as the MFDP, based on feature calculations from the University of New South Wales detector (UNSW). For both benchmark Physionet databases, the proposed MFPD method achieves the lowest standard deviation in sensitivity and positive predictivity (+P) despite its online algorithm architecture. While the statistics are comparable for low-to mildly artifactual ECG signals, the MFPD outperforms reference methods for artifacted ECG with low SNR levels reaching 87.48 ± 14.21% in sensitivity and 89.39 ± 14.67% in +P as compared to 88.30 ± 17.66% and 86.06 ± 19.67% respectively from UNSW, the best performing reference method. With demonstrations on the extensively studied QRS detection problem, we consider that the proposed generic structure of the multi-feature probabilistic detector should offer promising perspectives for long-term monitoring applications for highly-artifacted signals.

Improving methodology in heart rate variability analysis for the premature infants: Impact of the time length

BACKGROUND

Heart rate variability (HRV) has been emerging in neonatal medicine. It may help for the early diagnosis of pathology and estimation of autonomous maturation. There is a lack of standardization and automation in the selection of the sequences to analyze and some features have not been explored in this specific population. The main objective of this study was to analyze the impact of the time length of the sequences on the estimation of linear and non-linear HRV features, including horizontal visibility graphs (HVG).

METHODS

HRV features were repeatedly measured with linear and non-linear methods on 2-, 5-, 10-minute sequences selected from the longest 15-min sequence and recorded on a weekly basis in 39 infants less than 31 weeks at birth. The associations between HRV measurements were analyzed through principal component analysis and k-means clustering. The effects of the time lengths on HRV measurements and post-menstrual age (PMA) were analyzed by linear mixed effect model for repeated measures.

RESULTS

The domains of analysis were concordant for their descriptive parameters of short (rMSSD, SD1 and HF) and long-term (SD, SD2 and LF) variability. α1 was correlated with the LF/HF and SD2/SD1. DC and AC were correlated with short-term variability estimates and significantly increased with GA and PMA. Shortening the windows of analysis increased the random measurement error for all the features and increased the bias for all but short term features and HVGs.

CONCLUSION

The linear and non-linear measurements of HRV are correlated each other. Shortening the windows of analysis increased the random error for all the features and increased the bias for all but short term features and HVGs. Short-term HRV can be an index for evaluating the maturation in whatever sequence length.

Influence of nasal CPAP on cardiorespiratory control in healthy neonate

The present study aimed to further unravel the effects of nasal continuous positive airway pressure (nCPAP) on the cardiovascular and respiratory systems in the neonatal period. Six-hour polysomnographic recordings were first performed in seven healthy newborn lambs, aged 2-3 days, without and with nCPAP application at 6 cmH₂O (nCPAP-6), in randomized order. The effects of nCPAP-6 on heart rate variability, respiratory rate variability, and cardiorespiratory interrelations were analyzed using a semiautomatic signal processing approach applied to ECG and respiration recordings. Thereafter, a cardiorespiratory mathematical model was adapted to the experimental conditions to gain further physiological interpretation and to simulate higher nCPAP levels (8 and 10 cmH₂O). Results from the signal processing approach suggest that nCPAP-6 applied in newborns with healthy lungs: 1) increases heart rate and decreases the time and frequency domain indices of heart rate variability, especially those representing parasympathetic activity, while increasing the complexity of the RR-interval time series; 2) prolongs the respiratory cycle and expiration duration and decreases respiratory rate variability; and 3) slightly impairs cardiorespiratory interrelations. Model-based analysis revealed that nCPAP-6 increases the heart rate and decreases respiratory sinus arrhythmia amplitude, in association with a reduced parasympathetic efferent activity. These results were accentuated when simulating an increased CPAP level. Overall, our results provide a further understanding of the effects of nCPAP in neonates, in the absence of lung disease.NEW & NOTEWORTHY Application of nasal continuous positive airway pressure (CPAP) at 6 cmH₂O, a level very frequently used in newborns, alters heart and respiratory rate variability, as well as cardiorespiratory interrelations in a full-term newborn model without lung disease. Moreover, whereas nasal CPAP at 6 cmH₂O decreases parasympathetic efferent activity, there is no change in sympathetic efferent activity.

Influence of Moderate Hyperbilirubinemia on Cardiorespiratory Control in Preterm Lambs

Hyperbilirubinemia (HB) is responsible for neonatal jaundice in 60% of term newborns and 90% of preterm infants. Neonatal HB can induce neurological damage (acute HB encephalopathy) and has been associated with persistent apneas. The objective of the present study was to investigate the immediate and delayed effects of moderate, clinically-relevant HB on cardiorespiratory control in preterm lambs. Two groups of five preterm lambs, namely control and HB, were studied. At day five of life, moderate HB (150-250 μmol/L) was induced and maintained during 17 h in the HB group while control lambs received a placebo solution. Six hours after HB onset, 7-h polysomnographic recordings with electrocardiogram (ECG) and respiratory (RESP) signals were performed to assess the immediate effects of HB on heart rate variability (HRV), respiratory rate variability (RRV), and cardiorespiratory interrelations. Identical recordings were repeated 72 h after HB induction to examine the delayed effects of HB on HRV, RRV and cardiorespiratory interrelations. Our results demonstrate a higher HRV and vagal activity immediately after induction of moderate HB. Meanwhile, a decrease in respiratory rate with an increase in both long- and short-term RRV was also noted, as well as a higher amplitude of the respiratory sinus arrhythmia and cardiorespiratory coupling. Seventy-two hours later, the alterations in HRV, RRV, and cardiorespiratory interrelations were attenuated, although a number of them were still present, suggesting a lasting influence of HB on the basal control of the cardiorespiratory system. Our results pave the way for studies in human preterms to assess the relevance of monitoring HRV, RRV, and cardiorespiratory interrelations to detect the acute neurological effects of HB and consequently adapt the treatment of neonatal jaundice.

Daily fatigue-recovery balance monitoring with heart rate variability in well-trained female cyclists on the Tour de France circuit

Objectives

This study aimed to analyze the daily heart rate variability (HRV) in well-trained female cyclists during the 2017 Tour de France circuit and to relate it to the load and perceived exertion response.

Methods

Ten female cyclists volunteered to participate in the study. HRV was recorded with a portable heart rate monitor each morning at rest in supine (7 min.) and upright (7 min.) positions, as well as throughout each day’s stage. Pre-Tour baseline HRV recordings were made, as well as during the four weeks following completion of the Tour. Exercise daily load was assessed using the training impulse score (TRIMPS). Post-exercise rate of perceived exertion (RPE) was assessed daily using the Borg CR-10 scale.

Results

The results show a HRV imbalance, increase of sympathetic and decrease of vagal activities respectively, along the event that correlated with rate of perceived exertion (r = 0.46), training impulse score (r = 0.60), and kilometers (r = 046). The greatest change in HRV balance was observed the days after the greatest relative physical load. Mean heart rate and heart rate variability values returned to their baseline values one week after completion of the event.

Conclusions

Despite incomplete recovery from day-to-day, fatigue is not summative or augmented with each successive stage and its physical load. Just one week is sufficient to restore baseline values. Heart rate and HRV can be used as a tool to strategically plan the effort of female cyclists that participate in multi-stage events.

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.

Heart Signal Recognition by Hidden Markov Models: The ECG Case

Wave recognition in ECG signals by Hidden Markov Models (HMMs) relies on the stationary assumption for the set of parameters used to describe ECG waves. This approach seems unnatural and consequently generates severe errors in practice. A new class of HMMs called Modified Continuous Variable Duration HMMs is proposed to account for the specific properties of the ECG signal. An application of the latter, coupled with a multiresolution front-end analysis of the ECG is presented. Results show these methods can increase the perfomance of ECG recognition compared to classical HMMs.

A novel controller based on state-transition models for closed-loop vagus nerve stimulation: Application to heart rate regulation

Vagus nerve stimulation (VNS) is an established adjunctive therapy for pharmacologically refractory epilepsy and depression and is currently in active clinical research for other applications. In current clinical studies, VNS is delivered in an open-loop approach, where VNS parameters are defined during a manual titration phase. However, the physiological response to a given VNS configuration shows significant inter and intra-patient variability and may significantly evolve through time. VNS closed-loop approaches, allowing for the optimization of the therapy in an adaptive manner, may be necessary to improve efficacy while reducing side effects. This paper proposes a generic, closed-loop control VNS system that is able to optimize a number of VNS parameters in an adaptive fashion, in order to keep a control variable within a specified range. Although the proposed control method is completely generic, an example application using the cardiac beat to beat interval (RR) as control variable will be developed in this paper. The proposed controller is based on a state transition model (STM) that can be configured using a partially or a fully-connected architecture, different model orders and different state-transition algorithms. The controller is applied to the adaptive regulation of heart rate and evaluated on 6 sheep, for 13 different targets, using partially-connected STM with 10 states. Also, partially and fully-connected STM defined by 30 states were applied to 7 other sheep for the same 10 targets. Results illustrate the interest of the proposed fully-connected STM and the feasibility of integrating this control system into an implantable neuromodulator.

Activity Recognition Using Complex Network Analysis

In this paper, we perform complex network analysis on a connectivity dataset retrieved from a monitoring system in order to classify simple daily activities. The monitoring system is composed of a set of wearable sensing modules positioned on the subject's body and the connectivity data consists of the correlation between each pair of modules. A number of network measures are then computed followed by the application of statistical significance and feature selection methods. These methods were implemented for the purpose of reducing the total number of modules in the monitoring system required to provide accurate activity classification. The obtained results show that an overall accuracy of 84.6% for activity classification is achieved, using a random forest classifier, and when considering a monitoring system composed of only two modules positioned at the neck and thigh of the subject's body.

Chorioamnionitis following preterm premature rupture of membranes and fetal heart rate variability

Introduction

The objective of this study was to identify prenatal markers of histological chorioamnionitis (HC) during pPROM using fetal computerized cardiotocography (cCTG).

Materials and methods

Retrospective review of medical records from pregnant women referred for pPROM between 26 and 34 weeks, in whom placental histology was available, in a tertiary level obstetric service over a 5-year period. Fetal heart rate variability was assessed using cCTG. Patients were included if they were monitored at least six times in the 72 hours preceding delivery. Clinical and biological cCTG parameters during the pPROM latency period were compared between cases with or without HC.

Results

In total, 222 pPROM cases were observed, but cCTG data was available in only 23 of these cases (10 with and 13 without HC) after exclusion of co-morbidities which may potentially perturb fetal heart rate variability measures. Groups were comparable for maternal age, parity, gestational age at pPROM, pPROM duration and neonatal characteristics (p>0.1). Baseline fetal heart rate was higher in the HC group [median 147.3 bpm IQR (144.2–149.2) vs. 141.3 bpm (137.1–145.4) in no HC group; p = 0.02]. The number of low variation episodes [6.4, (3.5–15.3) vs. 2.3 (1–5.2); p = 0.04] was also higher in the HC group, whereas short term variations were lower in the HC group [7.1 ms (6–7.4) vs. 8.1 ms (7.4–9); p = 0.01] within 72 hours before delivery. Differences were especially discriminant within 24 hours before delivery, with less short-term variation [5 ms (3.7–5.9) vs. 7.8 ms (5.4–8.7); p = 0.007] and high variation episodes [3.9 (4.9–3.2) vs. 0.8 (1.5–0.2); p < 0.001] in the HC group.

Conclusion

These results show differences in fetal heart rate variability, suggesting that cCTG could be used clinically to diagnoses chorioamnionitis during the pPROM latency period.

Sensitivity Analysis of Vagus Nerve Stimulation Parameters on Acute Cardiac Autonomic Responses: Chronotropic, Inotropic and Dromotropic Effects

Although the therapeutic effects of Vagus Nerve Stimulation (VNS) have been recognized in pre-clinical and pilot clinical studies, the effect of different stimulation configurations on the cardiovascular response is still an open question, especially in the case of VNS delivered synchronously with cardiac activity. In this paper, we propose a formal mathematical methodology to analyze the acute cardiac response to different VNS configurations, jointly considering the chronotropic, dromotropic and inotropic cardiac effects. A latin hypercube sampling method was chosen to design a uniform experimental plan, composed of 75 different VNS configurations, with different values for the main parameters (current amplitude, number of delivered pulses, pulse width, interpulse period and the delay between the detected cardiac event and VNS onset). These VNS configurations were applied to 6 healthy, anesthetized sheep, while acquiring the associated cardiovascular response. Unobserved VNS configurations were estimated using a Gaussian process regression (GPR) model. In order to quantitatively analyze the effect of each parameter and their combinations on the cardiac response, the Sobol sensitivity method was applied to the obtained GPR model and inter-individual sensitivity markers were estimated using a bootstrap approach. Results highlight the dominant effect of pulse current, pulse width and number of pulses, which explain respectively 49.4%, 19.7% and 6.0% of the mean global cardiovascular variability provoked by VNS. More interestingly, results also quantify the effect of the interactions between VNS parameters. In particular, the interactions between current and pulse width provoke higher cardiac effects than the changes on the number of pulses alone (between 6 and 25% of the variability). Although the sensitivity of individual VNS parameters seems similar for chronotropic, dromotropic and inotropic responses, the interacting effects of VNS parameters provoke significantly different cardiac responses, showing the feasibility of a parameter-based functional selectivity. These results are of primary importance for the optimal, subject-specific definition of VNS parameters for a given therapy and may lead to new closed-loop methods allowing for the optimal adaptation of VNS therapy through time.

Vagus nerve stimulation: state of the art of stimulation and recording strategies to address autonomic function neuromodulation

OBJECTIVE

Neural signals along the vagus nerve (VN) drive many somatic and autonomic functions. The clinical interest of VN stimulation (VNS) is thus potentially huge and has already been demonstrated in epilepsy. However, side effects are often elicited, in addition to the targeted neuromodulation.

APPROACH

This review examines the state of the art of VNS applied to two emerging modulations of autonomic function: heart failure and obesity, especially morbid obesity.

MAIN RESULTS

We report that VNS may benefit from improved stimulation delivery using very advanced technologies. However, most of the results from fundamental animal studies still need to be demonstrated in humans.

Influence of Vagus Nerve Stimulation parameters on chronotropism and inotropism in heart failure

Vagus Nerve Stimulation (VNS) has been shown to be useful in heart failure patients, including antiarrhythmic effects, improvement of cardiac function and reduction of the mortality. However, the optimal configuration of VNS can be a difficult task, since there are several adjustable parameters, such as current amplitude (mA), pulse width (ms), burst frequency (Hz), number of pulses and, in the case of cardiac-triggered VNS, the delay (ms) between the R-wave and the beginning of the stimulation. The objective of this paper is to analyse the effect of these parameters, and their interaction, on the chronotropic and inotropic responses to vagal stimulation. 306 VNS sequences were tested on 12 sheep with induced heart failure. Autonomic markers of the chronotropic (changes in RR interval) and inotropic (changes in dP/dtmax) effects were extracted from the observed data. In order to analyse the influence of stimulation parameters on these markers, a sensitivity analysis method was applied. Results illustrate the strong interaction between the delay and the others parameters. The number of pulses, the current and the frequency seem to be particularly influent on chronotropism and inotropism although the effect of the frequency is highly non-linear or it depends on other parameters.

Detection of Levodopa Induced Dyskinesia in Parkinson's Disease patients based on activity classification

In this paper, we present an activity classification-based algorithm for the automatic detection of Levodopa Induced Dyskinesia in Parkinson's Disease (PD) patients. Two PD patients experiencing motor fluctuations related to chronic Levodopa therapy performed a protocol of simple daily life activities on at least two different occasions. A Random Forest classifier was able to classify the performed activities by the patients with an overall accuracy of 86%. Based on the detected activity, a K Nearest Neighbor classifier detected the presence of dyskinesia with accuracy ranging from 75% to 88%.

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.