Vital Sign Monitoring for Cancer Patients Based on Dual-Path Sensor and Divided-Frequency-CNN Model

Monitoring vital signs is a key part of standard medical care for cancer patients. However, the traditional methods have instability especially when big fluctuations of signals happen, while the deep-learning-based methods lack pertinence to the sensors. A dual-path micro-bend optical fiber sensor and a targeted model based on the Divided-Frequency-CNN (DFC) are developed in this paper to measure the heart rate (HR) and respiratory rate (RR). For each path, features of frequency division based on the mechanism of signal periodicity cooperate with the operation of stable phase extraction to reduce the interference of body movements for monitoring. Then, the DFC model is designed to learn the inner information from the features robustly. Lastly, a weighted strategy is used to estimate the HR and RR via dual paths to increase the anti-interference for errors from one source. The experiments were carried out on the actual clinical data of cancer patients by a hospital. The results show that the proposed method has good performance in error (3.51 (4.51 %) and 2.53 (3.28 %) beats per minute (bpm) for cancer patients with pain and without pain respectively), relevance, and consistency with the values from hospital equipment. Besides, the proposed method significantly improved the ability in the report time interval (30 to 9 min), and mean / confidential interval (3.60/[-22.61,29.81] to -0.64 / [-9.21,7.92] for patients with pain and 1.87 / [-5.49,9.23] to -0.16 / [-6.21,5.89] for patients without pain) compared with our previous work.

Advancing spinal cord injury care through non-invasive autonomic dysreflexia detection with AI

This paper presents an AI-powered solution for detecting and monitoring Autonomic Dysreflexia (AD) in individuals with spinal cord injuries. Current AD detection methods are limited, lacking non-invasive monitoring systems. We propose a model that combines skin nerve activity (SKNA) signals with a deep neural network (DNN) architecture to overcome this limitation. The DNN is trained on a meticulously curated dataset obtained through controlled colorectal distension, inducing AD events in rats with spinal cord surgery above the T6 level. The proposed system achieves an impressive average classification accuracy of 93.9% ± 2.5%, ensuring accurate AD identification with high precision (95.2% ± 2.1%). It demonstrates a balanced performance with an average F1 score of 94.4% ± 1.8%, indicating a harmonious balance between precision and recall. Additionally, the system exhibits a low average false-negative rate of 4.8% ± 1.6%, minimizing the misclassification of non-AD cases. The robustness and generalizability of the system are validated on unseen data, maintaining high accuracy, F1 score, and a low false-negative rate. This AI-powered solution represents a significant advancement in non-invasive, real-time AD monitoring, with the potential to improve patient outcomes and enhance AD management in individuals with spinal cord injuries. This research contributes a promising solution to the critical healthcare challenge of AD detection and monitoring.

A RUSBoosted tree method for k-complex detection using tunable Q-factor wavelet transform and multi-domain feature extraction

Background

K-complex detection traditionally relied on expert clinicians, which is time-consuming and onerous. Various automatic k-complex detection-based machine learning methods are presented. However, these methods always suffered from imbalanced datasets, which impede the subsequent processing steps.

New method

In this study, an efficient method for k-complex detection using electroencephalogram (EEG)-based multi-domain features extraction and selection method coupled with a RUSBoosted tree model is presented. EEG signals are first decomposed using a tunable Q-factor wavelet transform (TQWT). Then, multi-domain features based on TQWT are pulled out from TQWT sub-bands, and a self-adaptive feature set is obtained from a feature selection based on the consistency-based filter for the detection of k-complexes. Finally, the RUSBoosted tree model is used to perform k-complex detection.

Results

Experimental outcomes manifest the efficacy of our proposed scheme in terms of the average performance of recall measure, AUC, and F10-score. The proposed method yields 92.41 ± 7.47%, 95.4 ± 4.32%, and 83.13 ± 8.59% for k-complex detection in Scenario 1 and also achieves similar results in Scenario 2.

Comparison to state-of-the-art methods

The RUSBoosted tree model was compared with three other machine learning classifiers [i.e., linear discriminant analysis (LDA), logistic regression, and linear support vector machine (SVM)]. The performance based on the kappa coefficient, recall measure, and F10-score provided evidence that the proposed model surpassed other algorithms in the detection of the k-complexes, especially for the recall measure.

Conclusion

In summary, the RUSBoosted tree model presents a promising performance in dealing with highly imbalanced data. It can be an effective tool for doctors and neurologists to diagnose and treat sleep disorders.

QRS complex detection using stationary wavelet transform and adaptive thresholding

Purpose- Electrocardiogram (ECG) signal is a record of the electrical activity of the heart and contains important clinical data about cardiovascular-related misfunctioning. The goal of the present work is to develop an improved QRS detection algorithm for the detection of heart abnormalities. Methods- In this present work stationary wavelet transforms (SWT) based method has been proposed for precise detection of QRS complex with 'sym2' mother wavelet. The stationary wavelet transform is a systematic mathematical tool to decompose the signal without downsampling using scale analysis and provides high detection of QRS complex and accurate localization of signal components. In the proposed method four level of decomposition is applied and the initial thresholding value is computed by the maximum amplitude of scale one at level four in SWT coefficients without the zero-crossing amplitude detection method. The multi-layered dynamic thresholding method has been applied to detect the true R-peak values and locate the QRS complex in the ECG signal. Results- For evaluation of results, the presented methodology is assessed on MIT-BIH, QTDB, and Noise stress test databases. In MIT-BIH, the sensitivity = 99.88%, positive predictivity = 99.93%, accuracy = 99.80% and detection error rate = 0.18% is achieved. In NSTD database, sensitivity = 97.46%, positive predictivity = 94.20%, accuracy = 91.95% and detection error rate = 8.47% and in QTDB, sensitivity = 99.95%, positive predictivity = 99.90%, accuracy = 99.71% and detection error rate = 0.16% is executed. Conclusion- In the presented proposed methodology, the computation complexity is low and exhibits a simple technique rather than an empirical approach. The proposed technique corroborates the performance for the detection of QRS complex with improved accuracy.

Evaluation and optimization of novel extraction algorithms for the automatic detection of atrial activations recorded within the pulmonary veins during atrial fibrillation

Background and objective

The automated detection of atrial activations (AAs) recorded from intracardiac electrograms (IEGMs) during atrial fibrillation (AF) is challenging considering their various amplitudes, morphologies and cycle length. Activation time estimation is further complicated by the constant changes in the IEGM active zones in complex and/or fractionated signals. We propose a new method which provides reliable automatic extraction of intracardiac AAs recorded within the pulmonary veins during AF and an accurate estimation of their local activation times.

Methods

First, two recently developed algorithms were evaluated and optimized on 118 recordings of pulmonary vein IEGM taken from 35 patients undergoing ablation of persistent AF. The adaptive mathematical morphology algorithm (AMM) uses an adaptive structuring element to extract AAs based on their morphological features. The relative-energy algorithm (Rel-En) uses short- and long-term energies to enhance and detect the AAs in the IEGM signals. Second, following the AA extraction, the signal amplitude was weighted using statistics of the AA sequences in order to reduce over- and undersensing of the algorithms. The detection capacity of our algorithms was compared with manually annotated activations and with two previously developed algorithms based on the Teager–Kaiser energy operator and the AF cycle length iteration, respectively. Finally, a method based on the barycenter was developed to reduce artificial variations in the activation annotations of complex IEGM signals.

Results

The best detection was achieved using Rel-En, yielding a false negative rate of 0.76% and a false positive rate of only 0.12% (total error rate 0.88%) against expert annotation. The post-processing further reduced the total error rate of the Rel-En algorithm by 70% (yielding to a final total error rate of 0.28%).

Conclusion

The proposed method shows reliable detection and robust temporal annotation of AAs recorded within pulmonary veins in AF. The method has low computational cost and high robustness for automatic detection of AAs, which makes it a suitable approach for online use in a procedural context.

Event Scene Method of Legal Domain Knowledge Map Based on Neural Network Hybrid Model

Event extraction technology is one of the important researches in the field of information extraction, which helps people accurately retrieve, find, classify, and summarize effective information from a large amount of information streams. This paper uses the neural network hybrid model to identify the trigger words and event categories of the legal domain knowledge graph events, extracts the events of interest from a large amount of free text, and displays them in a structured format. First, the original text is preprocessed, and then, the distributed semantic word vector is combined with the dependent syntactic structure and location attributes to create a semantic representation in the form of a vector. The combined deep learning model is used to extract activated words, the long-term memory loop neural network uses temporal semantics to extract deep features, and the convergent neural network completes the extraction of activated words and event categories. Finally, the experimental results show that the accuracy of event extraction of the neural network hybrid model designed in this paper has reached 77.1%, and the recall rate has reached 76.8%, which is greatly improved compared with the traditional model.

Machine Learning Models and Videos of Facial Regions for Estimating Heart Rate: A Review on Patents, Datasets, and Literature

Estimating heart rate is important for monitoring users in various situations. Estimates based on facial videos are increasingly being researched because they allow the monitoring of cardiac information in a non-invasive way and because the devices are simpler, as they require only cameras that capture the user’s face. From these videos of the user’s face, machine learning can estimate heart rate. This study investigates the benefits and challenges of using machine learning models to estimate heart rate from facial videos through patents, datasets, and article review. We have searched the Derwent Innovation, IEEE Xplore, Scopus, and Web of Science knowledge bases and identified seven patent filings, eleven datasets, and twenty articles on heart rate, photoplethysmography, or electrocardiogram data. In terms of patents, we note the advantages of inventions related to heart rate estimation, as described by the authors. In terms of datasets, we have discovered that most of them are for academic purposes and with different signs and annotations that allow coverage for subjects other than heartbeat estimation. In terms of articles, we have discovered techniques, such as extracting regions of interest for heart rate reading and using video magnification for small motion extraction, and models, such as EVM-CNN and VGG-16, that extract the observed individual’s heart rate, the best regions of interest for signal extraction, and ways to process them.

Concurrent Evolution of Biomechanical and Physiological Parameters With Running-Induced Acute Fatigue

Understanding the influence of running-induced acute fatigue on the homeostasis of the body is essential to mitigate the adverse effects and optimize positive adaptations to training. Fatigue is a multifactorial phenomenon, which influences biomechanical, physiological, and psychological facets. This work aimed to assess the evolution of these three facets with acute fatigue during a half-marathon. 13 recreational runners were equipped with one inertial measurement unit (IMU) on each foot, one combined global navigation satellite system-IMU-electrocardiogram sensor on the chest, and an Android smartphone equipped with an audio recording application. Spatio-temporal parameters for the running gait, along with the heart rate, its variability and complexity were computed using validated algorithms. Perceived fatigability was assessed using the rating-of-fatigue (ROF) scale at every 10 min of the race. The data was split into eight equal segments, corresponding to at least one ROF value per segment, and only level running parts were retained for analysis. During the race, contact time, duty factor, and trunk anteroposterior acceleration increased, and the foot strike angle and vertical stiffness decreased significantly. Heart rate showed a progressive increase, while the metrics for heart rate variability and complexity decreased during the race. The biomechanical parameters showed a significant alteration even with a small change in perceived fatigue, whereas the heart rate dynamics altered at higher changes. When divided into two groups, the slower runners presented a higher change in heart rate dynamics throughout the race than the faster runners; they both showed similar trends for the gait parameters. When tested for linear and non-linear correlations, heart rate had the highest association with biomechanical parameters, while the trunk anteroposterior acceleration had the lowest association with heart rate dynamics. These results indicate the ability of faster runners to better judge their physiological limits and hint toward a higher sensitivity of perceived fatigue to neuromuscular changes in the running gait. This study highlights measurable influences of acute fatigue, which can be studied only through concurrent measurement of biomechanical, physiological, and psychological facets of running in real-world conditions.

ReBeatICG: Real-time Low-Complexity Beat-to-beat Impedance Cardiogram Delineation Algorithm

This work presents ReBeatICG, a real-time, low-complexity beat-to-beat impedance cardiography (ICG) delineation algorithm that allows hemodynamic parameters monitoring. The proposed procedure relies only on the ICG signal compared to most algorithms found in the literature that rely on synchronous electrocardiogram signal (ECG) recordings. ReBeatICG was designed with implementation on an ultra-low-power microcontroller (MCU) in mind. The detection accuracy of the developed algorithm is tested against points manually labeled by cardiologists. It achieves a detection Gmean accuracy of 94.9%, 98.6%, 90.3%, and 84.3% for the B, C, X, and O characteristic points, respectively. Furthermore, several hemodynamic parameters were calculated based on annotated characteristic points and compared with values generated from the cardiologists' annotations. ReBeatICG achieved mean error rates of 0.11 ms, 9.72 ms, 8.32 ms, and 3.97% for HR, LVET, IVRT, and relative C-point amplitude, respectively.

Real-Time EEG-Based Cognitive Workload Monitoring on Wearable Devices

OBJECTIVE

Cognitive workload monitoring (CWM) can enhance human-machine interaction by supporting task execution assistance considering the operators cognitive state. Therefore, we propose a machine learning design methodology and a data processing strategy to enable CWM on resource-constrained wearable devices.

METHODS

Our CWM solution is built upon edge computing on a simple wearable system, with only four peripheral channels of electroencephalography (EEG). We assess our solution on experimental data from 24 volunteers. Moreover, to overcome the system's memory constraints, we adopt an optimization strategy for model size reduction and a multi-batch data processing scheme for optimizing RAM memory footprint. Finally, we implement our data processing strategy on a state-of-the-art wearable platform and assess its execution and system battery life.

RESULTS

We achieve an accuracy of 74.5% and a 74.0% geometric mean between sensitivity and specificity for CWM classification on unseen data. Besides, the proposed model optimization strategy generates a 27.5x smaller model compared to the one generated with default parameters, and the multi-batch data processing scheme reduces RAM memory footprint by 14x compared to a single batch data processing. Finally, our algorithm uses only 1.28% of the available processing time, thus, allowing our system to achieve 28.5 hours of battery life.

CONCLUSION

We provide a reliable and optimized CWM solution using wearable devices, enabling more than a day of operation on a single battery charge.

SIGNIFICANCE

The proposed methodology enables real-time data processing on resource-constrained devices and supports real-time wearable monitoring based on EEG for applications as CWM in human-machine interaction.

An innovative method based on Shannon energy envelope and summit navigation for detecting R peaks of noise stress test signals

In recent decades, there has been an increased demand for the processing of electrocardiogram (ECG) signals because of its significant role in diagnosing cardiac diseases. The QRS complex is the dominant feature of the ECG signal. The detection of QRS complexes is thus an essential part of almost any ECG signal processing systems. This paper presents a developed QRS complex detection method using dominant peak extraction and Shannon energy envelope for useful ECG signal analysis. The algorithm is divided into three main stages: pre-processing, searching for dominant peaks, and removing false R peaks. The proposed algorithm is validated in static ECG recordings from the MIT-BIH Arrhythmia Database (MITDB) and noise-contaminated ECG stress tests from the Glasgow University Database (GUDB), separately. The method compares favorably with conventional and recently published results of many QRS detection algorithms on the same MITDB. Subsequently, valuable performance coefficients are also found on the GUDB. The average detection accuracy of finding R peaks exceed 99% on both the databases, especially for cardiac stress test records with high interference levels. The method enables a highly effective ECG signal processing tool under various noises, artifacts, abnormalities, and morphologies.

Precision Medicine and Artificial Intelligence: A Pilot Study on Deep Learning for Hypoglycemic Events Detection based on ECG

Tracking the fluctuations in blood glucose levels is important for healthy subjects and crucial diabetic patients. Tight glucose monitoring reduces the risk of hypoglycemia, which can result in a series of complications, especially in diabetic patients, such as confusion, irritability, seizure and can even be fatal in specific conditions. Hypoglycemia affects the electrophysiology of the heart. However, due to strong inter-subject heterogeneity, previous studies based on a cohort of subjects failed to deploy electrocardiogram (ECG)-based hypoglycemic detection systems reliably. The current study used personalised medicine approach and Artificial Intelligence (AI) to automatically detect nocturnal hypoglycemia using a few heartbeats of raw ECG signal recorded with non-invasive, wearable devices, in healthy individuals, monitored 24 hours for 14 consecutive days. Additionally, we present a visualisation method enabling clinicians to visualise which part of the ECG signal (e.g., T-wave, ST-interval) is significantly associated with the hypoglycemic event in each subject, overcoming the intelligibility problem of deep-learning methods. These results advance the feasibility of a real-time, non-invasive hypoglycemia alarming system using short excerpts of ECG signal.

An Efficient and Robust Digital Fractional Order Differentiator Based ECG Pre-Processor Design for QRS Detection

This paper presents an efficient infinite impulse response type digital fractional order differentiator (DFOD) based electrocardiogram (ECG) pre-processor to detect QRS complexes. First, an efficient optimizer namely, Antlion optimization algorithm is employed to solve the proposed DFOD design problem. Then, the designed DFOD is deployed in the pre-processing stage of a threshold independent R-peak detection technique. Finally, the proposed QRS complex detector is thoroughly assessed on the standard ECG datasets of MIT/BIH Arrhythmia, MIT/BIH ST Change, MIT/BIH Supraventricular Arrhythmia, European ST-T, QT, and T-Wave Alternans Challenge databases to show the wide sense practicability of the proposed DFOD-based QRS detector. The root-means-square magnitude error (RMSME) and the average group delay (τ_DD) metrics of the proposed DFOD are as low as -38.17 dB and 0.04 samples, respectively. The percentage of improvement in terms of RMSME metric compared to the best-reported approach is 15%. The overall sensitivity of 99.89% and positive predictivity of 99.88% are incurred by considering all the six databases. To the best of the authors' knowledge, it is the first time when the evolutionary algorithm based IIR-type DFOD is employed for the QRS complex detection and establishing its performance superiority. The results so obtained are compared with the results of all the recently reported QRS detectors. The proposed DFOD based ECG pre-processor has a great potential to robustly generate the feature signal related to the ECG QRS complex irrespective of the ECG morphology. Thus, the proposed DFOD based QRS detector can be employed in clinical ECG monitoring devices to augment the QRS detection performance.

REWARD: Design, Optimization, and Evaluation of a Real-Time Relative-Energy Wearable R-Peak Detection Algorithm. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2019;2019:3341-3347. [PMID: 31946597 DOI: 10.1109/embc.2019.8857226]

Wearable devices are an unobtrusive, cost-effective means of continuous ambulatory monitoring of chronic cardiovascular diseases. However, on these resource-constrained systems, electrocardiogram (ECG) processing algorithms must consume minimal power and memory, yet robustly provide accurate physiological information. This work presents REWARD, the Relative-Energy-based WeArable R-Peak Detection algorithm, which is a novel ECG R-peak detection mechanism based on a nonlinear filtering method called Relative-Energy (Rel-En). REWARD is designed and optimized for real-time execution on wearable systems. Then, this novel algorithm is compared against three state-of-the-art real-time R-peak detection algorithms in terms of accuracy, memory footprint, and energy consumption. The Physionet QT and NST Databases were employed to evaluate the algorithms' accuracy and robustness to noise, respectively. Then, a 32-bit ARM Cortex-M3-based microcontroller was used to measure the energy usage, computational burden, and memory footprint of the four algorithms. REWARD consumed at least 63% less energy and 32% less RAM than the other algorithms while obtaining comparable accuracy results. Therefore, REWARD would be a suitable choice of R-peak detection mechanism for wearable devices that perform more complex ECG analysis, whose algorithms require additional energy and memory resources.