An Advanced Two-Step DNN-Based Framework for Arrhythmia Detection
Ontology highlight
ABSTRACT:
SUBMITTER: Lauw H
PROVIDER: S-EPMC7206245 | biostudies-literature | 2020 Apr
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Tactile sensory input of surgeons is severely limited in minimally invasive surgery, therefore manual palpation cannot be performed for intraoperative tumor detection. Computer-aided palpation, in which tactile information is acquired by devices and relayed to the surgeon, is one solution for overcoming this limitation. An important design factor is the method by which the acquired information is fed back to the surgeon. However, currently there is no systematic method for achieving this aim, and it is possible that a badly implemented feedback mechanism could adversely affect the performance of the surgeon. In this study, we propose an assistance algorithm for intraoperative tumor detection in laparoscopic surgery. Our scenario is that the surgeon manipulates a sensor probe, makes a decision based on the feedback provided from the sensor, while simultaneously, the algorithm analyzes the time series of the sensor output. Thus, the algorithm assists the surgeon in making decisions by providing independent detection results. A deep neural network model with three hidden layers was used to analyze the sensor output. We propose methods to input the time series of the sensor output to the model for real-time analysis, and to determine the criterion for detection by the model. This study aims to validate the feasibility of the algorithm by using data acquired in our previous psychophysical experiment. There, novice participants were asked to detect a phantom of an early-stage gastric tumor through visual feedback from the tactile sensor. In addition to the analysis of the accuracy, signal detection theory was employed to assess the potential detection performance of the model. The detection performance was compared with that of human participants. We conducted two types of validation, and found that the detection performance of the model was not significantly different from that of the human participants if the data from a known user was included in the model construction. The result supports the feasibility of the proposed algorithm for detection assistance in computer-aided palpation.
| S-EPMC7806085 | biostudies-literature
Project description:Automatic detection of arrhythmia is of great significance for early prevention and diagnosis of cardiovascular disease. Traditional feature engineering methods based on expert knowledge lack multidimensional and multi-view information abstraction and data representation ability, so the traditional research on pattern recognition of arrhythmia detection cannot achieve satisfactory results. Recently, with the increase of deep learning technology, automatic feature extraction of ECG data based on deep neural networks has been widely discussed. In order to utilize the complementary strength between different schemes, in this paper, we propose an arrhythmia detection method based on the multi-resolution representation (MRR) of ECG signals. This method utilizes four different up to date deep neural networks as four channel models for ECG vector representations learning. The deep learning based representations, together with hand-crafted features of ECG, forms the MRR, which is the input of the downstream classification strategy. The experimental results of big ECG dataset multi-label classification confirm that the F1 score of the proposed method is 0.9238, which is 1.31%, 0.62%, 1.18% and 0.6% higher than that of each channel model. From the perspective of architecture, this proposed method is highly scalable and can be employed as an example for arrhythmia recognition.
| S-EPMC7175329 | biostudies-literature
Project description:BACKGROUND:In microbiome studies, it is important to detect taxa which are associated with pathological outcomes at the lowest definable taxonomic rank, such as genus or species. Traditionally, taxa at the target rank are tested for individual association, followed by the Benjamini-Hochberg (BH) procedure to control for false discovery rate (FDR). However, this approach neglects the dependence structure among taxa and may lead to conservative results. The taxonomic tree of microbiome data represents alignment from phylum to species rank and characterizes evolutionary relationships across microbial taxa. Taxa that are closer on the tree usually have similar responses to the exposure (environment). The statistical power in microbial association tests can be enhanced by efficiently employing the prior evolutionary information via the taxonomic tree. METHODS:We propose a two-stage microbial association mapping framework (massMap) which uses grouping information from the taxonomic tree to strengthen statistical power in association tests at the target rank. massMap first screens the association of taxonomic groups at a pre-selected higher taxonomic rank using a powerful microbial group test OMiAT. The method then proceeds to test the association for each candidate taxon at the target rank within the significant taxonomic groups identified in the first stage. Hierarchical BH (HBH) and selected subset testing (SST) procedures are evaluated to control the FDR for the two-stage structured tests. RESULTS:Our simulations show that massMap incorporating OMiAT and the advanced FDR controlling methodologies largely alleviates the multiplicity issue. It is statistically more powerful than the traditional association mapping directly at the target rank while controlling the FDR at desired levels under most scenarios. In our real data analyses, massMap detects more or the same amount of associated species with smaller adjusted p values compared to the traditional method, which further illustrates the efficiency of the proposed framework. The R package of massMap is publicly available at https://sites.google.com/site/huilinli09/software and https://github.com/JiyuanHu/ . CONCLUSIONS:massMap is a novel microbial association mapping framework and achieves additional efficiency by utilizing the intrinsic taxonomic structure of microbiome data.
| S-EPMC6060480 | biostudies-literature
Project description:Sensitive detection of nucleic acids and identification of single nucleotide polymorphism (SNP) is crucial in diagnosis of genetic diseases. Many strategies have been developed for detection and analysis of DNA, including fluorescence, electrical, optical, and mechanical methods. Recent advances in fluorescence resonance energy transfer (FRET)-based sensing have provided a new avenue for sensitive and quantitative detection of various types of biomolecules in simple, rapid, and recyclable platforms. Here, we report single-step FRET-based DNA sensors designed to work via a toehold-mediated strand displacement (TMSD) process, leading to a distinct change in the FRET efficiency upon target binding. Using single-molecule FRET (smFRET), we show that these sensors can be regenerated in situ, and they allow detection of femtomoles DNA without the need for target amplification while still using a dramatically small sample size (fewer than three orders of magnitude compared to the typical sample size of bulk fluorescence). In addition, these single-molecule sensors exhibit a dynamic range of approximately two orders of magnitude. Using one of the sensors, we demonstrate that the single-base mismatch sequence can be discriminated from a fully matched DNA target, showing a high specificity of the method. These sensors with simple and recyclable design, sensitive detection of DNA, and the ability to discriminate single-base mismatch sequences may find applications in quantitative analysis of nucleic acid biomarkers.
| S-EPMC6719117 | biostudies-literature
Project description:We present EP-DNN, a protocol for predicting enhancers based on chromatin features, in different cell types. Specifically, we use a deep neural network (DNN)-based architecture to extract enhancer signatures in a representative human embryonic stem cell type (H1) and a differentiated lung cell type (IMR90). We train EP-DNN using p300 binding sites, as enhancers, and TSS and random non-DHS sites, as non-enhancers. We perform same-cell and cross-cell predictions to quantify the validation rate and compare against two state-of-the-art methods, DEEP-ENCODE and RFECS. We find that EP-DNN has superior accuracy with a validation rate of 91.6%, relative to 85.3% for DEEP-ENCODE and 85.5% for RFECS, for a given number of enhancer predictions and also scales better for a larger number of enhancer predictions. Moreover, our H1 ? IMR90 predictions turn out to be more accurate than IMR90 ? IMR90, potentially because H1 exhibits a richer signature set and our EP-DNN model is expressive enough to extract these subtleties. Our work shows how to leverage the full expressivity of deep learning models, using multiple hidden layers, while avoiding overfitting on the training data. We also lay the foundation for exploration of cross-cell enhancer predictions, potentially reducing the need for expensive experimentation.
| S-EPMC5144062 | biostudies-literature
Project description:BackgroundAt present, the bioinformatics research on the relationship between aging-related diseases and genes is mainly through the establishment of a machine learning multi-label model to classify each gene. Most of the existing methods for predicting pathogenic genes mainly rely on specific types of gene features, or directly encode multiple features with different dimensions, use the same encoder to concatenate and predict the final results, which will be subject to many limitations in the applicability of the algorithm. Possible shortcomings of the above include: incomplete coverage of gene features by a single type of biomics data, overfitting of small dimensional datasets by a single encoder, or underfitting of larger dimensional datasets.MethodsWe use the known gene disease association data and gene descriptors, such as gene ontology terms (GO), protein interaction data (PPI), PathDIP, Kyoto Encyclopedia of genes and genomes Genes (KEGG), etc, as input for deep learning to predict the association between genes and diseases. Our innovation is to use Mashup algorithm to reduce the dimensionality of PPI, GO and other large biological networks, and add new pathway data in KEGG database, and then combine a variety of biological information sources through modular Deep Neural Network (DNN) to predict the genes related to aging diseases.Result and conclusionThe results show that our algorithm is more effective than the standard neural network algorithm (the Area Under the ROC curve from 0.8795 to 0.9153), gradient enhanced tree classifier and logistic regression classifier. In this paper, we firstly use DNN to learn the similar genes associated with the known diseases from the complex multi-dimensional feature space, and then provide the evidence that the assumed genes are associated with a certain disease.
| S-EPMC8680025 | biostudies-literature
Project description:In this work, a memristive spike-based computing in memory (CIM) system with adaptive neuron (MSPAN) is proposed to realize energy-efficient remote arrhythmia detection with high accuracy in edge devices by software and hardware co-design. A multi-layer deep integrative spiking neural network (DiSNN) is first designed with an accuracy of 93.6% in 4-class ECG classification tasks. Then a memristor-based CIM architecture and the corresponding mapping method are proposed to deploy the DiSNN. By evaluation, the overall system achieves an accuracy of over 92.25% on the MIT-BIH dataset while the area is 3.438 mm2 and the power consumption is 0.178 μJ per heartbeat at a clock frequency of 500 MHz. These results reveal that the proposed MSPAN system is promising for arrhythmia detection in edge devices.
| S-EPMC8715923 | biostudies-literature
Project description:Target image detection based on a rapid serial visual presentation (RSVP) paradigm is a typical brain-computer interface system with various applications, such as image retrieval. In an RSVP paradigm, a P300 component is detected to determine target images. This strategy requires high-precision single-trial P300 detection methods. However, the performance of single-trial detection methods is relatively lower than that of multitrial P300 detection methods. Image retrieval based on multitrial P300 is a new research direction. In this paper, we propose a triple-RSVP paradigm with three images being presented simultaneously and a target image appearing three times. Thus, multitrial P300 classification methods can be used to improve detection accuracy. In this study, these mechanisms were extended and validated, and the characteristics of the multi-RSVP framework were further explored. Two different P300 detection algorithms were also utilized in multi-RSVP to demonstrate that the scheme is universally applicable. Results revealed that the detection accuracy of the multi-RSVP paradigm was higher than that of the standard RSVP paradigm. The results validate the effectiveness of the proposed method, and this method can provide a whole new idea in the field of EEG-based target detection.
| S-EPMC5541818 | biostudies-other
Project description:During cardiac arrhythmias, dynamical patterns of electrical activation form and evolve, which are of interest to understand and cure heart rhythm disorders. The analysis of these patterns is commonly performed by calculating the local activation phase and searching for phase singularities (PSs), i.e., points around which all phases are present. Here we propose an alternative framework, which focuses on phase defect lines (PDLs) and surfaces (PDSs) as more general mechanisms, which include PSs as a specific case. The proposed framework enables two conceptual unifications: between the local activation time and phase description, and between conduction block lines and the central regions of linear-core rotors. A simple PDL detection method is proposed and applied to data from simulations and optical mapping experiments. Our analysis of ventricular tachycardia in rabbit hearts (n = 6) shows that nearly all detected PSs were found on PDLs, but the PDLs had a significantly longer lifespan than the detected PSs. Since the proposed framework revisits basic building blocks of cardiac activation patterns, it can become a useful tool for further theory development and experimental analysis.
| S-EPMC8494009 | biostudies-literature
Project description:Introduction: Pharmacists need to be adaptable, flexible, and capable of advancing their practice to adapt to rapidly changing population health needs. We describe an educational approach to pharmacy workforce transformation in Indonesia through an advanced practice competency framework development using an "adopt and adapt" methodology. Methods: The competency framework development process comprised a translation phase, an adopt and adapt phase, validation through a nationwide mapping survey, and a completion phase through leadership consensus panels. We conducted a forward-backwards translation of a previously validated Advanced to Consultancy Level Framework (ACLF) to yield the Indonesian Advanced Development Framework (IADF) draft. The subsequent adoption and adaptation process was conducted through a series of consensus panels. We validated the IADF through a nationwide workforce survey. The final phase included leadership consensus panels with the professional leadership body in Indonesia. We analyzed the qualitative data thematically and the quantitative data using a Multiple Correspondence Analysis (MCA) technique. Results: We identified conceptual challenges in adopting and adapting the existing ACLF, which were addressed by providing a national glossary and concrete examples. A total of 6,212 pharmacists participated in the national workforce survey, of which 43% had <2 years of post-license (post-registration) experience. The MCA results showed that practitioner self-assessment to the IADF could discriminate their career development stages. The results also indicated a four-stage career model (including early years career training). Embedding this model in a structured national training program will enhance the professional workforce development through a more structured career journey. Conclusions: We describe the first validation of an advanced competency development framework for the pharmacy workforce in a non-Anglophone country, showing the possibility of transnational applicability of this framework. We argue that this methodology can be used in Low and Middle-income countries (LMICs) for the more rapid advancement of pharmaceutical care practice.
| S-EPMC8655862 | biostudies-literature