Dataset Information

A New Multichannel Parallel Network Framework for the Special Structure of Multilead ECG.

ABSTRACT: Electrocardiogram (ECG) contains the rhythmic features of continuous heartbeat and morphological features of ECG waveforms and varies among different diseases. Based on ECG signal features, we propose a combination of multiple neural networks, the multichannel parallel neural network (MLCNN-BiLSTM), to explore feature information contained in ECG. The MLCNN channel is used in extracting the morphological features of ECG waveforms. Compared with traditional convolutional neural network (CNN), the MLCNN can accurately extract strong relevant information on multilead ECG while ignoring irrelevant information. It is suitable for the special structures of multilead ECG. The Bidirectional Long Short-Term Memory (BiLSTM) channel is used in extracting the rhythmic features of ECG continuous heartbeat. Finally, by initializing the core threshold parameters and using the backpropagation algorithm to update automatically, the weighted fusion of the temporal-spatial features extracted from multiple channels in parallel is used in exploring the sensitivity of different cardiovascular diseases to morphological and rhythmic features. Experimental results show that the accuracy rate of multiple cardiovascular diseases is 87.81%, sensitivity is 86.00%, and specificity is 87.76%. We proposed the MLCNN-BiLSTM neural network that can be used as the first-round screening tool for clinical diagnosis of ECG.

SUBMITTER: Lu P

PROVIDER: S-EPMC7728482 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A New Multichannel Parallel Network Framework for the Special Structure of Multilead ECG.

Lu Peng P Xi Hao H Zhou Bing B Zhang Hongpo H Lin Yusong Y Chen Liwei L Gao Yang Y Zhang Yabin Y Hu Yanhua Y

Journal of healthcare engineering 20201203

Electrocardiogram (ECG) contains the rhythmic features of continuous heartbeat and morphological features of ECG waveforms and varies among different diseases. Based on ECG signal features, we propose a combination of multiple neural networks, the multichannel parallel neural network (MLCNN-BiLSTM), to explore feature information contained in ECG. The MLCNN channel is used in extracting the morphological features of ECG waveforms. Compared with traditional convolutional neural network (CNN), the ...[more]

PMID: 33343853

Similar Datasets

Project description:BackgroundThe development of wearable health monitoring systems is garnering tremendous interest in research, technology and commercial applications. Their ability of providing unique capabilities in continuous, real-time, and non-invasive tracking of the physiological markers of users can provide insights into the performance and health of individuals. Electrocardiogram (ECG) signals are of particular interest, as cardiovascular disease is the leading cause of death globally. Monitoring heart health and its conditions such as ventricular disturbances and arrhythmias can be achieved through evaluating various features of ECG such as R-peaks, QRS complex, T-wave, and P-wave. Despite recent advances in biosensors for wearable applications, most of the currently available solutions rely solely on a single system attached to the body, limiting the ability to obtain reliable and multi-location biosignals. However, in engineering systems, sensor fusion, which is the optimal integration and processing of data from multiple sensors, has been a common theme and should be considered for wearables. In recent years, due to an increase in the availability and variety of different types of sensors, the possibility of achieving sensor fusion in wearable systems has become more attainable. Sensor fusion in multi-sensing systems results in significant enhancements of information inferences compared to those from systems with a sole sensor. One step towards the development of sensor fusion for wearable health monitoring systems is the accessibility to multiple reliable electrophysiological signals, which can be recorded continuously.ResultsIn this paper, we develop a textile-based multichannel ECG band that has the ability to measure ECG from multiple locations on the waist. As a proof of concept, we demonstrate that ECG signals can be reliably obtained from different locations on the waist where the shape of the QRS complex is nearly comparable with recordings from the chest using traditional gel electrodes. In addition, we develop a probabilistic approach-based on prediction and update strategies-to detect R-peaks from noisy textile data in different statuses, including sitting, standing, and jogging. In this approach, an optimal search method is utilized to detect R-peaks based on the history of the intervals between previously detected R-peaks. We show that the performance of our probabilistic approach in R-peak detection is significantly better than that based on Pan-Tompkins and optimal-threshold methods.ConclusionA textile-based multichannel ECG band was developed to track the heart rate changes from multiple locations on the waist. We demonstrated that (i) the ECG signal can be detected from different locations on the waist, and (ii) the accuracy of the detected R-peaks from textile sensors was improved by using our proposed probabilistic approach. Despite the limitations of the textile sensors that might compromise the quality of ECG signals, we anticipate that the textile-based multichannel ECG band can be considered as an effective wearable system to facilitate the development of sensor fusion methodology for pervasive and non-invasive health monitoring through continuous tracking of heart rate variability (HRV) from the waist. In addition, from the commercialization point of view, we anticipate that the developed band has the potential to be integrated into garments such as underwear, bras or pants so that individuals can use it on a daily basis.

Project description:OBJECTIVE:Although substantial progress has been made in mapping the connections of the brain, less is known about how this organization translates into brain function. In particular, the massive interconnectivity of the brain has made it difficult to specifically examine data transmission between two nodes of the connectome, a central component of the 'neural code.' Here, we investigated the propagation of multiple streams of asynchronous neuronal activity across an isolated in vitro 'connectome unit.' APPROACH:We used the novel technique of axon stretch growth to create a model of a long-range cortico-cortical network, a modular system consisting of paired nodes of cortical neurons connected by axon tracts. Using optical stimulation and multi-electrode array recording techniques, we explored how input patterns are represented by cortical networks, how these representations shift as they are transmitted between cortical nodes and perturbed by external conditions, and how well the downstream node distinguishes different patterns. MAIN RESULTS:Stimulus representations included direct, synaptic, and multiplexed responses that grew in complexity as the distance between the stimulation source and recorded neuron increased. These representations collapsed into patterns with lower information content at higher stimulation frequencies. With internodal activity propagation, a hierarchy of network pathways, including latent circuits, was revealed using glutamatergic blockade. As stimulus channels were added, divergent, non-linear effects were observed in local versus distant network layers. Pairwise difference analysis of neuronal responses suggested that neuronal ensembles generally outperformed individual cells in discriminating input patterns. SIGNIFICANCE:Our data illuminate the complexity of spiking activity propagation in cortical networks in vitro, which is characterized by the transformation of an input into myriad outputs over several network layers. These results provide insight into how the brain potentially processes information and generates the neural code and could guide the development of clinical therapies based on multichannel brain stimulation.

Project description:Nuclear co-localization labels are critical to ocular research. Among these, the TUNEL assay has been established as a gold standard of cell death and apoptosis. While several validated computer-based methods exist to quantitate these markers, including ImageJ Retina Analysis (RA) Toolkit and ImagePro, none verify the count with the nuclear counter stain to confirm nuclear co-localization. We established a new ImageJ-based automated multichannel thresholding (MCT) method to quantitate nuclear co-localized labeling. The MCT method was validated by comparing it with the two published TUNEL analysis in TUNEL-positive photoreceptors in an experimental retinal detachment (RD) model. RDs were induced in murine eyes and cross-sectional images of TUNEL and DAPI counter stain were obtained. Images were classified as "typical" or high density "hotspot" TUNEL regions (n?=?10/group). Images were analyzed and compared between the MCT method and the published techniques including both "standard" and "high" settings of the RA Toolkit for detecting lower or higher TUNEL densities, respectively. Additional testing of the MCT method with built-in ImageJ thresholding algorithms was performed to produce fully automated measurements. All images were compared with Bland-Altman mean difference plots to assess the difference in counts and linear regression plots to assess correlation. Comparison between the MCT method and the ImagePro method were found to be well correlated (typical: R2?=?0.8972, hotspot: R2?=?0.9000) with minor to non-significant differences. The RA Toolkit settings were found to be mostly well correlated as well (standard/typical: R2?=?0.8036, standard/hotspot: R2?=?0.4309, high/typical: R2?=?0.7895, high/hotspot: R2?=?0.8738) but were often found to have significantly higher counts than the MCT. In conclusion, the MCT method compared favorably with validated computer-based methods of nuclear marker immunofluorescence quantitation and avoids staining artifacts through the incorporation of the nuclear counter stain to confirm positive cells.

Dataset Information

A New Multichannel Parallel Network Framework for the Special Structure of Multilead ECG.

Publications

A New Multichannel Parallel Network Framework for the Special Structure of Multilead ECG.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure