Showing papers in "Physiological Measurement in 2018"

An open source benchmarked toolbox for cardiovascular waveform and interval analysis.

Abstract: OBJECTIVE This work aims to validate a set of data processing methods for variability metrics, which hold promise as potential indicators for autonomic function, prediction of adverse cardiovascular outcomes, psychophysiological status, and general wellness Although the investigation of heart rate variability (HRV) has been prevalent for several decades, the methods used for preprocessing, windowing, and choosing appropriate parameters lacks consensus among academic and clinical investigators Moreover, many of the important steps are omitted from publications, preventing reproducibility APPROACH To address this, we have compiled a comprehensive and open-source modular toolbox for calculating HRV metrics and other related variability indices, on both raw cardiovascular time series and RR intervals The software, known as the PhysioNet Cardiovascular Signal Toolbox, is implemented in the MATLAB programming language, with standard (open) input and output formats, and requires no external libraries The functioning of our software is compared with other widely used and referenced HRV toolboxes to identify important differences MAIN RESULTS Our findings demonstrate how modest differences in the approach to HRV analysis can lead to divergent results, a factor that might have contributed to the lack of repeatability of studies and clinical applicability of HRV metrics SIGNIFICANCE Existing HRV toolboxes do not include standardized preprocessing, signal quality indices (for noisy segment removal), and abnormal rhythm detection and are therefore likely to lead to significant errors in the presence of moderate to high noise or arrhythmias We therefore describe the inclusion of validated tools to address these issues We also make recommendations for default values and testing/reporting

Passive BCI beyond the lab: current trends and future directions

Abstract: Over the last decade, passive brain-computer interface (BCI) algorithms and biosignal acquisition technologies have experienced a significant growth that has allowed the real-time analysis of biosignals, with the aim to quantify relevant insights, such as mental and emotional states, of the users Several passive BCI-based applications have been tested in laboratory settings, and just a few of them in real or, at least, simulated but highly realistic settings Nevertheless, works performed in laboratory settings are not able to take into account all those factors (artefacts, non-brain influences, other mental states) that could impair the usability of passive BCIs during real applications, naturally characterized by higher complexity The present review takes into account the most recent trends in using advanced passive BCI technologies in real settings, especially for real-time mental state evaluations in operational environments, evaluation of team resources, training and expertise assessment, gaming and neuromarketing applications The objective of the work is to draw a mark on where we are to date and the future challenges, in order to make passive BCIs closer to being integrated into daily life applications

ECG signal classification for the detection of cardiac arrhythmias using a convolutional recurrent neural network.

Abstract: Objective The electrocardiogram (ECG) provides an effective, non-invasive approach for clinical diagnosis in patients with cardiac diseases such as atrial fibrillation (AF). AF is the most common cardiac rhythm disturbance and affects ~2% of the general population in industrialized countries. Automatic AF detection in clinics remains a challenging task due to the high inter-patient variability of ECGs, and unsatisfactory existing approaches for AF diagnosis (e.g. atrial or ventricular activity-based analyses). Approach We have developed RhythmNet, a 21-layer 1D convolutional recurrent neural network, trained using 8528 single-lead ECG recordings from the 2017 PhysioNet/Computing in Cardiology (CinC) Challenge, to classify ECGs of different rhythms including AF automatically. Our RhythmNet architecture contained 16 convolutions to extract features directly from raw ECG waveforms, followed by three recurrent layers to process ECGs of varying lengths and to detect arrhythmia events in long recordings. Large 15 × 1 convolutional filters were used to effectively learn the detailed variations of the signal within small time-frames such as the P-waves and QRS complexes. We employed residual connections throughout RhythmNet, along with batch-normalization and rectified linear activation units to improve convergence during training. Main results We evaluated our algorithm on 3658 testing data and obtained an F 1 accuracy of 82% for classifying sinus rhythm, AF, and other arrhythmias. RhythmNet was also ranked 5th in the 2017 CinC Challenge. Significance Potentially, our approach could aid AF diagnosis in clinics and be used for patient self-monitoring to improve the early detection and effective treatment of AF.

Assessing mental stress from the photoplethysmogram: a numerical study.

Abstract: Objective Mental stress is detrimental to cardiovascular health, being a risk factor for coronary heart disease and a trigger for cardiac events. However, it is not currently routinely assessed. The aim of this study was to identify features of the photoplethysmogram (PPG) pulse wave which are indicative of mental stress. Approach A numerical model of pulse wave propagation was used to simulate blood pressure signals, from which simulated PPG pulse waves were estimated using a transfer function. Pulse waves were simulated at six levels of stress by changing the model input parameters both simultaneously and individually, in accordance with haemodynamic changes associated with stress. Thirty-two feature measurements were extracted from pulse waves at three measurement sites: the brachial, radial and temporal arteries. Features which changed significantly with stress were identified using the Mann-Kendall monotonic trend test. Main results Seventeen features exhibited significant trends with stress in measurements from at least one site. Three features showed significant trends at all three sites: the time from pulse onset to peak, the time from the dicrotic notch to pulse end, and the pulse rate. More features showed significant trends at the radial artery (15) than the brachial (8) or temporal (7) arteries. Most features were influenced by multiple input parameters. Significance The features identified in this study could be used to monitor stress in healthcare and consumer devices. Measurements at the radial artery may provide superior performance than the brachial or temporal arteries. In vivo studies are required to confirm these observations.

A review of physiological and behavioral monitoring with digital sensors for neuropsychiatric illnesses

Abstract: Physiological, behavioral, and psychological changes associated with neuropsychiatric illness are reflected in several related signals, including actigraphy, location, word sentiment, voice tone, social activity, heart rate, and responses to standardized questionnaires. These signals can be passively monitored using sensors in smartphones, wearable accelerometers, Holter monitors, and multimodal sensing approaches that fuse multiple data types. Connection of these devices to the internet has made large scale studies feasible and is enabling a revolution in neuropsychiatric monitoring. Currently, evaluation and diagnosis of neuropsychiatric disorders relies on clinical visits, which are infrequent and out of the context of a patient's home environment. Moreover, the demand for clinical care far exceeds the supply of providers. The growing prevalence of context-aware and physiologically relevant digital sensors in consumer technology could help address these challenges, enable objective indexing of patient severity, and inform rapid adjustment of treatment in real-time. Here we review recent studies utilizing such sensors in the context of neuropsychiatric illnesses including stress and depression, bipolar disorder, schizophrenia, post traumatic stress disorder, Alzheimer's disease, and Parkinson's disease.

Detection of tuberculosis by automatic cough sound analysis.

Abstract: Objective: Globally, tuberculosis (TB) remains one of the most deadly diseases. Although several effective diagnosis methods exist, in lower income countries clinics may not be in a position to afford expensive equipment and employ the trained experts needed to interpret results. In these situations, symptoms including cough are commonly used to identify patients for testing. However, self-reported cough has suboptimal sensitivity and specificity, which may be improved by digital detection. Approach: This study investigates a simple and easily applied method for TB screening based on the automatic analysis of coughing sounds. A database of cough audio recordings was collected and used to develop statistical classifiers. Main results: These classifiers use short-term spectral information to automatically distinguish between the coughs of TB positive patients and healthy controls with an accuracy of 78% and an AUC of 0.95. When a set of five clinical measurements is available in addition to the audio, this accuracy improves to 82%. By choosing an appropriate decision threshold, the system can achieve a sensitivity of 95% at a specificity of approximately 72%. The experiments suggest that the classifiers are using some spectral information that is not perceivable by the human auditory system, and that certain frequencies are more useful for classification than others. Significance: We conclude that automatic classification of coughing sounds may represent a viable low-cost and low-complexity screening method for TB.

A robust deep convolutional neural network for the classification of abnormal cardiac rhythm using single lead electrocardiograms of variable length

Abstract: Objective: Atrial fibrillation (AF) is a major cause of hospitalization and death in the United States. Moreover, as the average age of individuals increases around the world, early detection and diagnosis of AF become even more pressing. In this paper, we introduce a novel deep learning architecture for the detection of normal sinus rhythm, AF, other abnormal rhythms, and noise. Approach: We have demonstrated through a systematic approach many hyperparameters, input sets, and optimization methods that yielded influence in both training time and performance accuracy. We have focused on these properties to identify an optimal 13-layer convolutional neural network (CNN) model which was trained on 8528 short single-lead ECG recordings and evaluated on a test dataset of 3658 recordings. Main results: The proposed CNN architecture achieved a state-of-the-art performance in identifying normal, AF and other rhythms with an average F 1-score of 0.83. Significance: We have presented a robust deep learning-based architecture that can identify abnormal cardiac rhythms using short single-lead ECG recordings. The proposed architecture is computationally fast and can also be used in real-time cardiac arrhythmia detection applications.

A convolutional neural network for ECG annotation as the basis for classification of cardiac rhythms.

Abstract: Objective Electrocardiography is the most common tool to diagnose cardiovascular diseases. Annotation, segmentation and rhythm classification of ECGs are challenging tasks, especially in the presence of atrial fibrillation and other arrhythmias. Our aim is to increase the accuracy of heart rhythm estimation by the use of extreme gradient boosting trees and the development of a deep convolutional neural network for ECG segmentation. Approach We trained a convolutional neural network with waveforms from PhysioNet databases to annotate QRS complexes, P waves, T waves, noise and interbeat ECG segments that characterize the essences of normal and irregular heart beats. We evaluated true positive rates, positive predictive values and mean absolute differences of our annotation based on reference annotations of the QT and MIT-BIH P-wave database. Moreover, we compared the results with standard QRS detectors and Ecgpuwave. Extreme gradient boosting trees were used to determine the heart rhythm based on hand-crafted features. More precisely, a noise estimation function was used in combination with heart rate and interval data. Furthermore we defined particular features based on ECG morphology, appearance of P waves and detection of irregular beats. We examined the feature importance and identified key features for normal sinus rhythm, atrial fibrillation, alternative rhythm and noisy recordings. The classification performance was evaluated externally using F 1 scores by applying the algorithm to the hidden test set provided by the PhysioNet/CinC Challenge 2017. Main results The true positive rate of the convolutional neural network in detection of manually revised R peaks in the QT database was [Formula: see text] and the positive predictive value was [Formula: see text]. The detection of P and T waves reached a true positive rate of [Formula: see text] and [Formula: see text] respectively, given a 50 ms tolerance when comparing the reference to the test annotation set. The rhythm classification performance reached an overall F 1 score of 0.82 when applying the algorithm to the hidden test set. Significance We achieved a shared rank #9 in the post-challenge phase of the PhysioNet/CinC Challenge 2017.

New photoplethysmogram indicators for improving cuffless and continuous blood pressure estimation accuracy.

Abstract: OBJECTIVE The accuracy of cuffless and continuous blood pressure (BP) estimation has been improved, but it is still unsatisfactory for clinical uses. This study was designed to further increase BP estimation accuracy. APPROACH In this study, a number of new indicators were extracted from photoplethysmogram (PPG) recordings and a linear regression method was used to construct BP estimation models based on the PPG indicators and pulse transit time (PTT). The performance of the BP estimation models was evaluated by the PPG recordings from 22 subjects when they performed mental arithmetic stress and Valsalva's manoeuvre tasks that could induce BP fluctuations. MAIN RESULTS Our results showed that the best PPG-based BP estimation model could achieve a decrease of 0.31 ± 0.08 mmHg in systolic BP (SBP) and 0.33 ± 0.01 mmHg in diastolic BP (DBP) on estimation errors of grand absolute mean (GAM) and standard deviation (GSD) in comparison to the previously reported PPG-based methods. The best estimation model based on the combination of PPG and PPT could achieve a decrease (GAM & GSD) of 0.81 ± 0.95 mmHg in SBP and 0.75 ± 0.54 mmHg in DBP in comparison to the PPT-based methods. SIGNIFICANCE The findings suggest that the newly proposed PPG indicators would be promising for improving the accuracy of continuous and cuffless BP estimation.

A deep learning approach for fetal QRS complex detection

Abstract: Objective: Non-invasive foetal electrocardiography (NI-FECG) has the potential to provide more additional clinical information for detecting and diagnosing fetal diseases. We propose and demonstrate a deep learning approach for fetal QRS complex detection from raw NI-FECG signals by using a convolutional neural network (CNN) model. The main objective is to investigate whether reliable fetal QRS complex detection performance can still be obtained from features of single-channel NI-FECG signals, without canceling maternal ECG (MECG) signals. Approach: A deep learning method is proposed for recognizing fetal QRS complexes. Firstly, we collect data from set-a of the PhysioNet/computing in Cardiology Challenge database. The sample entropy method is used for signal quality assessment. Part of the bad quality signals is excluded in the further analysis. Secondly, in the proposed method, the features of raw NI-FECG signals are normalized before they are fed to a CNN classifier to perform fetal QRS complex detection. We use precision, recall, F-measure and accuracy as the evaluation metrics to assess the performance of fetal QRS complex detection. Main results: The proposed deep learning method can achieve relatively high precision (75.33%), recall (80.54%), and F-measure scores (77.85%) compared with three other well-known pattern classification methods, namely KNN, naive Bayes and SVM. Significance: the proposed deep learning method can attain reliable fetal QRS complex detection performance from the raw NI-FECG signals without canceling MECG signals. In addition, the influence of different activation functions and signal quality assessment on classification performance are evaluated, and results show that Relu outperforms the Sigmoid and Tanh on this particular task, and better classification performance is obtained with the signal quality assessment step in this study.

Classification methods to detect sleep apnea in adults based on respiratory and oximetry signals: a systematic review.

Abstract: Objective Sleep apnea (SA), a common sleep disorder, can significantly decrease the quality of life, and is closely associated with major health risks such as cardiovascular disease, sudden death, depression, and hypertension. The normal diagnostic process of SA using polysomnography is costly and time consuming. In addition, the accuracy of different classification methods to detect SA varies with the use of different physiological signals. If an effective, reliable, and accurate classification method is developed, then the diagnosis of SA and its associated treatment will be time-efficient and economical. This study aims to systematically review the literature and present an overview of classification methods to detect SA using respiratory and oximetry signals and address the automated detection approach. Approach Sixty-two included studies revealed the application of single and multiple signals (respiratory and oximetry) for the diagnosis of SA. Main results Both airflow and oxygen saturation signals alone were effective in detecting SA in the case of binary decision-making, whereas multiple signals were good for multi-class detection. In addition, some machine learning methods were superior to the other classification methods for SA detection using respiratory and oximetry signals. Significance To deal with the respiratory and oximetry signals, a good choice of classification method as well as the consideration of associated factors would result in high accuracy in the detection of SA. An accurate classification method should provide a high detection rate with an automated (independent of human action) analysis of respiratory and oximetry signals. Future high-quality automated studies using large samples of data from multiple patient groups or record batches are recommended.

Electrohysterography in the diagnosis of preterm birth: a review.

Abstract: Preterm birth (PTB) is one of the most common and serious complications in pregnancy. About 15 million preterm neonates are born every year, with ratios of 10-15% of total births. In industrialized countries, preterm delivery is responsible for 70% of mortality and 75% of morbidity in the neonatal period. Diagnostic means for its timely risk assessment are lacking and the underlying physiological mechanisms are unclear. Surface recording of the uterine myoelectrical activity (electrohysterogram, EHG) has emerged as a better uterine dynamics monitoring technique than traditional surface pressure recordings and provides information on the condition of uterine muscle in different obstetrical scenarios with emphasis on predicting preterm deliveries. Objective A comprehensive review of the literature was performed on studies related to the use of the electrohysterogram in the PTB context. Approach This review presents and discusses the results according to the different types of parameter (temporal and spectral, non-linear and bivariate) used for EHG characterization. Main results Electrohysterogram analysis reveals that the uterine electrophysiological changes that precede spontaneous preterm labor are associated with contractions of more intensity, higher frequency content, faster and more organized propagated activity and stronger coupling of different uterine areas. Temporal, spectral, non-linear and bivariate EHG analyses therefore provide useful and complementary information. Classificatory techniques of different types and varying complexity have been developed to diagnose PTB. The information derived from these different types of EHG parameters, either individually or in combination, is able to provide more accurate predictions of PTB than current clinical methods. However, in order to extend EHG to clinical applications, the recording set-up should be simplified, be less intrusive and more robust-and signal analysis should be automated without requiring much supervision and yield physiologically interpretable results. Significance This review provides a general background to PTB and describes how EHG can be used to better understand its underlying physiological mechanisms and improve its prediction. The findings will help future research workers to decide the most appropriate EHG features to be used in their analyses and facilitate future clinical EHG applications in order to improve PTB prediction.

Shoulder physiotherapy exercise recognition: machine learning the inertial signals from a smartwatch.

Abstract: Objective Participation in a physical therapy program is considered one of the greatest predictors of successful conservative management of common shoulder disorders. However, adherence to these protocols is often poor and typically worse for unsupervised home exercise programs. Currently, there are limited tools available for objective measurement of adherence in the home setting. The goal of this study was to develop and evaluate the potential for performing home shoulder physiotherapy monitoring using a commercial smartwatch. Approach Twenty healthy adult subjects with no prior shoulder disorders performed seven exercises from an evidence-based rotator cuff physiotherapy protocol, while 6-axis inertial sensor data was collected from the active extremity. Within an activity recognition chain (ARC) framework, four supervised learning algorithms were trained and optimized to classify the exercises: k-nearest neighbor (k-NN), random forest (RF), support vector machine classifier (SVC), and a convolutional recurrent neural network (CRNN). Algorithm performance was evaluated using 5-fold cross-validation stratified first temporally and then by subject. Main results Categorical classification accuracy was above 94% for all algorithms on the temporally stratified cross validation, with the best performance achieved by the CRNN algorithm (99.4%). The subject stratified cross validation, which evaluated classifier performance on unseen subjects, yielded lower accuracies scores again with CRNN performing best (88.9%). Significance This proof of concept study demonstrates the technical feasibility of a smartwatch device and supervised machine learning approach to more easily monitor and assess the at-home adherence of shoulder physiotherapy exercise protocols.

A physicist's view of homeostasis: how time series of continuous monitoring reflect the function of physiological variables in regulatory mechanisms.

Abstract: OBJECTIVE Homeostasis is one of the key concepts of physiology and the basis to understand chronic-degenerative disease and human ageing, but is difficult to quantify in clinical practice. The variability of time series resulting from continuous and non-invasive physiological monitoring is conjectured to reflect the underlying homeostatic regulatory processes, but it is not clear why the variability of some variables such as heart rate gives a favourable health prognosis whereas the variability of other variables such as blood pressure implies an increased risk factor. The purpose of the present contribution is to quantify homeostasis using time-series analysis and to offer an explanation for the phenomenology of physiological time series. APPROACH Within the context of network physiology, which focusses on the interactions between various variables at multiple scales of time and space, it may be understood that different physiological variables may play distinct roles in their respective regulatory mechanisms. In the present contribution, we distinguish between regulated variables, such as blood pressure or core temperature, and physiological responses, such as heart rate and skin temperature. MAIN RESULTS We give evidence that in optimal conditions of youth and health the former are characterized by Gaussian statistics, low variability and represent the stability of the internal environment, whereas the latter are characterized by non-Gaussian distributions, large variability and reflect the adaptive capacity of the human body; in the adverse conditions of ageing and/or disease, adaptive capacity is lost and the variability of physiological responses is diminished, and as a consequence the stability of the internal environment is compromised and its variability increases. SIGNIFICANCE Time-series analysis allows one to quantify homeostasis in the optimal conditions of youth and health and the degradation of homeostasis or homeostenosis in the adverse conditions of ageing and/or disease, and may offer an alternative approach to diagnosis in clinical practice.

Deep learning in the cross-time frequency domain for sleep staging from a single-lead electrocardiogram

Abstract: OBJECTIVE This study classifies sleep stages from a single lead electrocardiogram (ECG) using beat detection, cardiorespiratory coupling in the time-frequency domain and a deep convolutional neural network (CNN). APPROACH An ECG-derived respiration (EDR) signal and synchronous beat-to-beat heart rate variability (HRV) time series were derived from the ECG using previously described robust algorithms. A measure of cardiorespiratory coupling (CRC) was extracted by calculating the coherence and cross-spectrogram of the EDR and HRV signal in 5 min windows. A CNN was then trained to classify the sleep stages (wake, rapid-eye-movement (REM) sleep, non-REM (NREM) light sleep and NREM deep sleep) from the corresponding CRC spectrograms. A support vector machine was then used to combine the output of CNN with the other features derived from the ECG, including phase-rectified signal averaging (PRSA), sample entropy, as well as standard spectral and temporal HRV measures. The MIT-BIH Polysomnographic Database (SLPDB), the PhysioNet/Computing in Cardiology Challenge 2018 database (CinC2018) and the Sleep Heart Health Study (SHHS) database, all expert-annotated for sleep stages, were used to train and validate the algorithm. MAIN RESULTS Ten-fold cross validation results showed that the proposed algorithm achieved an accuracy (Acc) of 75.4% and a Cohen's kappa coefficient of [Formula: see text] = 0.54 on the out of sample validation data in the classification of Wake, REM, NREM light and deep sleep in SLPDB. This rose to Acc = 81.6% and [Formula: see text] = 0.63 for the classification of Wake, REM sleep and NREM sleep and Acc = 85.1% and [Formula: see text] = 0.68 for the classification of NREM sleep versus REM/wakefulness in SLPDB. SIGNIFICANCE The proposed ECG-based sleep stage classification approach that represents the highest reported results on non-electroencephalographic data and uses datasets over ten times larger than those in previous studies. By using a state-of-the-art QRS detector and deep learning model, the system does not require human annotation and can therefore be scaled for mass analysis.

Comparison of photoplethysmogram measured from wrist and finger and the effect of measurement location on pulse arrival time.

Abstract: Objective The aim of this paper was to compare photoplethysmogram (PPG) signals measured from the wrist and finger and to evaluate if wrist PPG signal could be used to calculate pulse arrival time (PAT), the time delay between electrocardiogram (ECG) R peak and a feature (e.g. peak, foot, first derivative peak) in the PPG signal. Further, the correlation between pulse wave velocity (rePWV, defined as PWV from ECG R peak to extremity) and systolic blood pressure was studied. Approach Thirty subjects were measured at rest by a trained research nurse. For reference measurement, chest ECG and finger PPG were measured using commercial sensors. Wrist PPG and arm ECG were measured with a custom-made setup, where the PPG sensor was located at the back surface of the forearm. Main results Reference finger and wrist PPG signals were found to differ in shape and also in amplitude. The PPG foot or first derivative peak detection methods seemed to be the most suitable methods for wrist PAT calculation. The Pearson correlation coefficient between blood pressure and rePWV was found to be 0.44 for the reference finger measurement and 0.37 for the wrist measurement. Significance Wrist PPG signal is widely used in optical heart rate monitors. Based on the results obtained in this study, wrist PPG signal may be used also for PAT calculation. The use of PAT for blood pressure estimation still has challenges, but PAT as such could be used as an interesting indicator of vascular health.

Sleep-wake classification via quantifying heart rate variability by convolutional neural network.

Abstract: Objective Fluctuations in heart rate are intimately related to changes in the physiological state of the organism. We exploit this relationship by classifying a human participant's wake/sleep status using his instantaneous heart rate (IHR) series. Approach We use a convolutional neural network (CNN) to build features from the IHR series extracted from a whole-night electrocardiogram (ECG) and predict every 30 s whether the participant is awake or asleep. Our training database consists of 56 normal participants, and we consider three different databases for validation; one is private, and two are public with different races and apnea severities. Main results On our private database of 27 participants, our accuracy, sensitivity, specificity, and [Formula: see text] values for predicting the wake stage are [Formula: see text], 52.4%, 89.4%, and 0.83, respectively. Validation performance is similar on our two public databases. When we use the photoplethysmography instead of the ECG to obtain the IHR series, the performance is also comparable. A robustness check is carried out to confirm the obtained performance statistics. Significance This result advocates for an effective and scalable method for recognizing changes in physiological state using non-invasive heart rate monitoring. The CNN model adaptively quantifies IHR fluctuation as well as its location in time and is suitable for differentiating between the wake and sleep stages.

Comparison between electrocardiogram- and photoplethysmogram-derived features for atrial fibrillation detection in free-living conditions.

Abstract: OBJECTIVE: Atrial fibrillation (AF) is the most commonly experienced arrhythmia and it increases the risk of stroke and heart failure. The challenge in detecting the presence of AF is the occasional and asymptomatic manifestation of the condition. Long-term monitoring can increase the sensitivity of detecting intermittent AF episodes, however it is either cumbersome or invasive and costly with electrocardiography (ECG). Photoplethysmography (PPG) is an unobtrusive measuring modality enabling heart rate monitoring, and promising results have been presented in detecting AF. However, there is still limited knowledge about the applicability of the PPG solutions in free-living conditions. The aim of this study was to compare the inter-beat interval derived features for AF detection between ECG and wrist-worn PPG in daily life. APPROACH: The data consisted of 24 h ECG, PPG, and accelerometer measurements from 27 patients (eight AF, 19 non-AF). In total, seven features (Shannon entropy, root mean square of successive differences (RMSSD), normalized RMSSD, pNN40, pNN70, sample entropy, and coefficient of sample entropy (CosEn)) were compared. Body movement was measured with the accelerometer and used with three different thresholds to exclude PPG segments affected by movement. MAIN RESULTS: CosEn resulted as the best performing feature from ECG with Cohens kappa 0.95. When the strictest movement threshold was applied, the same performance was obtained with PPG (kappa = 0.96). In addition, pNN40 and pNN70 reached similar results with the same threshold (kappa = 0.95 and 0.94), but were more robust with respect to movement artefacts. The coverage of PPG was 24.0%-57.6% depending on the movement threshold compared to 92.1% of ECG. SIGNIFICANCE: The inter-beat interval features derived from PPG are equivalent to the ones from ECG for AF detection. Movement artefacts substantially worsen PPG-based AF monitoring in free-living conditions, therefore monitoring coverage needs to be carefully selected. Wrist-worn PPG still provides a promising technology for long-term AF monitoring.

Big data in Parkinson's disease: using smartphones to remotely detect longitudinal disease phenotypes.

Abstract: Objective To better understand the longitudinal characteristics of Parkinson's disease (PD) through the analysis of finger tapping and memory tests collected remotely using smartphones. Approach Using a large cohort (312 PD subjects and 236 controls) of participants in the mPower study, we extract clinically validated features from a finger tapping and memory test to monitor the longitudinal behaviour of study participants. We investigate any discrepancy in learning rates associated with motor and non-motor tasks between PD subjects and healthy controls. The ability of these features to predict self-assigned severity measures is assessed whilst simultaneously inspecting the severity scoring system for floor-ceiling effects. Finally, we study the relationship between motor and non-motor longitudinal behaviour to determine if separate aspects of the disease are dependent on one another. Main results We find that the test performances of the most severe subjects show significant correlations with self-assigned severity measures. Interestingly, less severe subjects do not show significant correlations, which is shown to be a consequence of floor-ceiling effects within the mPower self-reporting severity system. We find that motor performance after practise is a better predictor of severity than baseline performance suggesting that starting performance at a new motor task is less representative of disease severity than the performance after the test has been learnt. We find PD subjects show significant impairments in motor ability as assessed through the alternating finger tapping (AFT) test in both the short- and long-term analyses. In the AFT and memory tests we demonstrate that PD subjects show a larger degree of longitudinal performance variability in addition to requiring more instances of a test to reach a steady state performance than healthy subjects. Significance Our findings pave the way forward for objective assessment and quantification of longitudinal learning rates in PD. This can be particularly useful for symptom monitoring and assessing medication response. This study tries to tackle some of the major challenges associated with self-assessed severity labels by designing and validating features extracted from big datasets in PD, which could help identify digital biomarkers capable of providing measures of disease severity outside of a clinical environment.

Eye tracking technology in sports-related concussion: a systematic review and meta-analysis.

Abstract: Objective The development of objective quantitative tools for the assessment and monitoring of sports-related concussion is critical. Eye tracking is a novel tool that may provide suitable metrics. The aim of this review was to appraise current evidence for the use of eye tracking technology in sports-related concussion assessment and monitoring. Approach A systematic literature review was conducted following the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. A search was run using Google Scholar, Microsoft Academic and PubMed for literature published between January 1980 and May 2018. Included were empirical research studies in English where at least 50% of the research participants were athletes, the participants were individuals with a diagnosis of concussion, and eye movements were measured using an eye tracking device. Main results This systematic review integrates 21 publications on sports-related concussion and eye tracking technology, nine of which also qualified for the meta-analysis. Overall, the literature reported significant findings for variables in each of the four classes of eye tracking measurements (movement, position, count, and latency). Meta-comparison was made for seven variables for the acute concussions (the difference between the concussed and the control groups was significant for all of them) and one variable for the latent concussions (the difference was not significant). Significance Most saccadic and pursuit deficits may be missed during clinical examination, and therefore eye tracking technology may be a useful and sensitive screening and monitoring tool for sports-related concussions. The inconsistencies between the eye movement metrics and methodology still make inferences challenging; however, using tasks that are closely related to brain areas involved in executive functions (such as memory-based saccade or antisaccade tasks) in the acute injury phase holds promise in differentiating between athletes who have a concussion compared to those who do not.

Monitoring of heart rate and inter-beat intervals with wrist plethysmography in patients with atrial fibrillation.

Abstract: OBJECTIVE Atrial fibrillation (AF) causes marked risk for patients, while silent fibrillation may remain unnoticed if not suspected and screened. Development of comfortable yet accurate beat-to-beat heart rate (HR) monitoring with good AF detection sensitivity would facilitate screening and improve treatment. The purpose of this study was to evaluate whether a wrist-worn photoplethysmography (PPG) device can be used to monitor beat-to-beat HR accurately during post-operative treatment in patients suffering from AF and whether wrist-PPG can be used to distinguish AF from sinus rhythm (SR). APPROACH Twenty-nine patients (14 with AF, 15 with SR, mean age 71.5 years) with multiple comorbidities were monitored during routine post-operative treatment. The monitoring included standard ECG, finger PPG monitoring and a wrist-worn PPG monitor with green and infrared light sources. The HR from PPG sensors was compared against ECG-derived HR. MAIN RESULTS The wrist PPG technology had very good HR and beat detection accuracy when using green light. For the SR group, the mean absolute error (MAE) for HR was 1.50 bpm, and for the inter-beat intervals (IBI), the MAE was 7.64 ms. For the AF group, the MAE for HR was 4.28 bpm and for IBI, the MAE was 14.67 ms. Accuracy for the infrared (IR) channel was worse. Finger PPG provided similar accuracy for HR and better accuracy for the IBI. AF detection sensitivity using green light was 99.0% and the specificity was 93.0%. Performance can be improved by discarding unreliable IBI periods. SIGNIFICANCE Results suggest that wrist PPG measurement allows accurate HR and beat-to-beat HR monitoring also in AF patients, and could be used for differentiating between SR and AF with very good sensitivity.

Multiclass classification of obstructive sleep apnea/hypopnea based on a convolutional neural network from a single-lead electrocardiogram

Abstract: Objective In this paper, we propose a convolutional neural network (CNN)-based deep learning architecture for multiclass classification of obstructive sleep apnea and hypopnea (OSAH) using single-lead electrocardiogram (ECG) recordings. OSAH is the most common sleep-related breathing disorder. Many subjects who suffer from OSAH remain undiagnosed; thus, early detection of OSAH is important. Approach In this study, automatic classification of three classes-normal, hypopnea, and apnea-based on a CNN is performed. An optimal six-layer CNN model is trained on a training dataset (45 096 events) and evaluated on a test dataset (11 274 events). The training set (69 subjects) and test set (17 subjects) were collected from 86 subjects with length of approximately 6 h and segmented into 10 s durations. Main results The proposed CNN model reaches a mean [Formula: see text]-score of 93.0 for the training dataset and 87.0 for the test dataset. Significance Thus, proposed deep learning architecture achieved a high performance for multiclass classification of OSAH using single-lead ECG recordings. The proposed method can be employed in screening of patients suspected of having OSAH.

Ensembling convolutional and long short-term memory networks for electrocardiogram arrhythmia detection.

Abstract: Objective: Atrial fibrillation is a common type of heart rhythm abnormality caused by a problem with the heart's electrical system. Early detection of this disease has important implications for stroke prevention and management. Our objective is to construct an intelligent tool that assists cardiologists in identifying automatically cardiac arrhythmias and noise in electrocardiogram (ECG) recordings. Approach: Our base deep classifier combined a convolutional neural network (CNNs) and a sequence of long short-term memory units, with pooling, dropout and normalization techniques to improve their accuracy. The network predicted a classification at every 18th input sample and the final prediction was selected for classification. Ten standalone models that used our base classifier architecture were first cross-validated separately on 90% of the PhysioNet/CinC Challenge 2017 dataset and then tested on 10%. An ensemble classifier selected the label of the best average probability from the ten sub-models to improve prediction quality. Main results: Our original result submitted to the challenge gave a mean F1-measure of 80%. The new proposed method improved the test score to 82%, which was tied for the third-highest score in the follow-up phase of the challenge. Significance: Without employing a time-consuming feature engineering step, the ensemble classifier trained with this architecture provided a robust solution to the problem of detecting cardiac arrhythmia from noisy ECG signals. In addition, interpretation of the classifier by inspection of its network parameters and predictions revealed what aspects of the ECG signal the classifier considered most discriminating.

Analyzing single-lead short ECG recordings using dense convolutional neural networks and feature-based post-processing to detect atrial fibrillation

Abstract: Objective The prevalence of atrial fibrillation (AF) in the general population is 0.5%-1%. As AF is the most common sustained cardiac arrhythmia that is associated with an increased morbidity and mortality, its timely diagnosis is clinically desirable. The main aim of this study as our contribution to the PhysioNet/CinC Challenge 2017 was to develop an automatic algorithm for classification of normal sinus rhythm (NSR), AF, other rhythm (O), and noise using a short single-channel ECG. Furthermore, the impact of changing labels/annotations on performance of the proposed algorithm was studied in this article. Approach The challenge training dataset (8528 ECG recordings) and a complementary dataset (6312 ECG recordings) from other sources were used for algorithm development. Version 3 (v3), which is an updated version of the annotations at the official phase of the challenge (v2), was used in this study. In the proposed algorithm, densely connected convolutional networks were combined with feature-based post-processing after initial signal quality analysis for the classification of ECG recordings. Main results The F1 scores for classification of NSR, AF, and O were 0.91, 0.83, and 0.72, respectively, which led to a F1 of 0.82. There was a small or no performance difference between the top entries in the official phase of the challenge and our proposed method. An increase of 2.5% in F1 score was observed when the same annotations for training and test was used (using v3 annotations) compared to using different annotations (v2 annotations for training and v3 annotations for the test). Significance Our promising results suggest that the availability of more data with improved labeling along with improvement in signal quality analysis make our algorithm suitable for practical clinical applications.

Clinical performance of a novel textile interface for neonatal chest electrical impedance tomography.

Abstract: Objective: Critically ill neonates and infants might particularly benefit from continuous chest electrical impedance tomography (EIT) monitoring at the bedside. In this study a textile 32-electrode interface for neonatal EIT examination has been developed and tested to validate its clinical performance. The objectives were to assess ease of use in a clinical setting, stability of contact impedance at the electrode–skin interface and possible adverse effects. Approach: Thirty preterm infants (gestational age: 30.3 ± 3.9 week (mean ± SD), postnatal age: 13.8 ± 28.2 d, body weight at inclusion: 1727 ± 869 g) were included in this multicentre study. The electrode–skin contact impedances were measured continuously for up to 3 d and analysed during the initial 20-min phase after fastening the belt and during a 10 h measurement interval without any clinical interventions. The skin condition was assessed by attending clinicians. Main results: Our findings imply that the textile electrode interface is suitable for long-term neonatal chest EIT imaging. It does not cause any distress for the preterm infants or discomfort. Stable contact impedance of about 300 Ohm was observed immediately after fastening the electrode belt and during the subsequent 20 min period. A slight increase in contact impedance was observed over time. Tidal variation of contact impedance was less than 5 Ohm. Significance: The availability of a textile 32-electrode belt for neonatal EIT imaging with simple, fast, accurate and reproducible placement on the chest strengthens the potential of EIT to be used for regional lung monitoring in critically ill neonates and infants.

Parallel use of a convolutional neural network and bagged tree ensemble for the classification of Holter ECG

Abstract: The automated detection of arrhythmia in a Holter ECG signal is a challenging task due to its complex clinical content and data quantity. It is also challenging due to the fact that Holter ECG is usually affected by noise. Such noise may be the result of the regular activity of patients using the Holter ECG—partially unplugged electrodes, short-time disconnections due to movement, or disturbances caused by electric devices or infrastructure. Furthermore, regular patient activities such as movement also affect the ECG signals and, in connection with artificial noise, may render the ECG non-readable or may lead to misinterpretation of the ECG. Objective: In accordance with the PhysioNet/CinC Challenge 2017, we propose a method for automated classification of 1-lead Holter ECG recordings. Approach: The proposed method classifies a tested record into one of four classes—'normal', 'atrial fibrillation', 'other arrhythmia' or 'too noisy to classify'. It uses two machine learning methods in parallel. The first—a bagged tree ensemble (BTE)—processes a set of 43 features based on QRS detection and PQRS morphology. The second—a convolutional neural network connected to a shallow neural network (CNN/NN)—uses ECG filtered by nine different filters (8× envelograms, 1× band-pass). If the output of CNN/NN reaches a specific level of certainty, its output is used. Otherwise, the BTE output is preferred. Main results: The proposed method was trained using a reduced version of the public PhysioNet/CinC Challenge 2017 dataset (8183 records) and remotely tested on the hidden dataset on PhysioNet servers (3658 records). The method achieved F1 test scores of 0.92, 0.82 and 0.74 for normal recordings, atrial fibrillation and recordings containing other arrhythmias, respectively. The overall F1 score measured on the hidden test-set was 0.83. Significance: This F1 score led to shared rank #2 in the follow-up PhysioNet/CinC Challenge 2017 ranking.

A comparison of entropy approaches for AF discrimination.

Abstract: Objective This study focuses on the comparison of single entropy measures for ventricular response analysis-based AF detection. Approach To enhance the performance of entropy-based AF detectors, we developed a normalized fuzzy entropy, [Formula: see text], a novel metric that (1) uses a fuzzy function to determine vector similarity, (2) replaces probability estimation with density estimation for entropy approximation, (3) utilizes a flexible distance threshold parameter, and (4) adjusts for heart rate by subtracting the natural log value of the mean RR interval. An AF detector based on [Formula: see text] was trained using the MIT-BIH atrial fibrillation (AF) database, and tested on the MIT-BIH normal sinus rhythm (NSR) and MIT-BIH arrhythmia databases. The [Formula: see text]-based AF detector was compared to AF detectors based on three other entropy measures: sample entropy ([Formula: see text]), fuzzy measure entropy ([Formula: see text]) and coefficient of sample entropy ([Formula: see text]), over three standard window sizes. Main results To classify AF and non-AF rhythms, [Formula: see text] achieved the highest area under receiver operating characteristic curve (AUC) values of 92.72%, 95.27% and 96.76% for 12-, 30- and 60-beat window lengths respectively. This was higher than the performance of the next best technique, [Formula: see text], over all windows sizes, which provided respective AUCs of 91.12%, 91.86% and 90.55%. [Formula: see text] and [Formula: see text] resulted in lower AUCs (below 90%) over all window sizes. [Formula: see text] also provided superior performance for all other tested statistics, including the Youden index, sensitivity, specificity, accuracy, positive predictivity and negative predictivity. In conclusion, we show that [Formula: see text] can be used to accurately identify AF from RR interval time series. Furthermore, longer window lengths (up to one minute) increase the performance of all entropy-based AF detectors under evaluation except the [Formula: see text]-based method. Significance Our results demonstrate that the new developed normalized fuzzy entropy is an accurate measure for detecting AF.

Prediction and early detection of delirium in the intensive care unit by using heart rate variability and machine learning

Abstract: Objective: Delirium is an important syndrome found in patients in the intensive care unit (ICU), however, it is usually under-recognized during treatment. This study was performed to investigate whether delirious patients can be successfully distinguished from non-delirious patients by using heart rate variability (HRV) and machine learning. Approach: Electrocardiography data of 140 patients was acquired during daily ICU care, and HRV data were analyzed. Delirium, including its type, severity, and etiologies, was evaluated daily by trained psychiatrists. HRV data and various machine learning algorithms including linear support vector machine (SVM), SVM with radial basis function (RBF) kernels, linear extreme learning machine (ELM), ELM with RBF kernels, linear discriminant analysis, and quadratic discriminant analysis were utilized to distinguish delirium patients from non-delirium patients. Main results: HRV data of 4797 ECGs were included, and 39 patients had delirium at least once during their ICU stay. The maximum classification accuracy was acquired using SVM with RBF kernels. Our prediction method based on HRV with machine learning was comparable to previous delirium prediction models using massive amounts of clinical information. Significance: Our results show that autonomic alterations could be a significant feature of patients with delirium in the ICU, suggesting the potential for the automatic prediction and early detection of delirium based on HRV with machine learning.

Comment on 'Invasive and noninvasive means of measuring intracranial pressure: a review'.

Abstract: We read with great interest the review by Zhang et al (2017) concerning the invasive and noninvasive means of measuring intracranial pressure. We would like to congratulate the authors for their interesting paper, but we were a bit disappointed reading the section dedicated to the role of ultrasound in the measurement of the optic nerve sheath diameter (ONSD).

Wearable technology reveals gait compensations, unstable walking patterns and fatigue in people with multiple sclerosis.

Abstract: Objective: People with multiple sclerosis (PwMS) often experience a decline in gait performance, which can compromise their independence and increase falls. Ankle joint contractures in PwMS are common and often result in compensatory gait patterns to accommodate reduced ankle range of motion (ROM). Approach: Using advances in wearable technology, the aim of this study was to quantify head and pelvis movement patterns that occur in PwMS with disability and determine how these secondary gait compensations impact on gait stability. Twelve healthy participants and 12 PwMS participated in the study. Head and pelvis movements were measured using two tri-axial accelerometers. Measures of gait compensation, mobility, variability, asymmetry, stability and fatigue were assessed during a 6 min walking test. Main results: Compared to healthy controls, PwMS had greater vertical asymmetry in their head and pelvic movements (Cohen's d = 1.85 and 1.60). Lower harmonic ratios indicated that PwMS were more unstable than controls (Cohen's d = −1.61 to −3.06), even after adjusting for their slower walking speeds. In the PwMS, increased compensatory movements were correlated with reduced ankle active ROM (r = −0.71), higher disability (EDSS) scores (r = 0.58), unstable gait (r = −0.76), reduced mobility (r = −0.76) and increased variability (r = 0.83). Significance: Wearable device technology provides an efficient and reliable way to screen for excessive compensatory movements often present in PwMS and provides clinically important information that impacts on mobility, stride time variability and gait stability. This information may help clinicians identify PwMS at high risk of falling and develop better rehabilitation interventions that, in addition to improving mobility, may help target the underlying causes of unstable gait.