Microb Biotechnol. 2024 Jun;17(6):e14478. doi: 10.1111/1751-7915.14478.

ABSTRACT

Clostridioides difficile (CD) infections are defined by toxins A (TcdA) and B (TcdB) along with the binary toxin (CDT). The emergence of the 'hypervirulent' (Hv) strain PR 027, along with PR 176 and 181, two decades ago, reshaped CD infection epidemiology in Europe. This study assessed MALDI-TOF mass spectrometry (MALDI-TOF MS) combined with machine learning (ML) and Deep Learning (DL) to identify toxigenic strains (producing TcdA, TcdB with or without CDT) and Hv strains. In total, 201 CD strains were analysed, comprising 151 toxigenic (24 ToxA+B+CDT+, 22 ToxA+B+CDT+ Hv+ and 105 ToxA+B+CDT-) and 50 non-toxigenic (ToxA-B-) strains. The DL-based classifier exhibited a 0.95 negative predictive value for excluding ToxA-B- strains, showcasing accuracy in identifying this strain category. Sensitivity in correctly identifying ToxA+B+CDT- strains ranged from 0.68 to 0.91. Additionally, all classifiers consistently demonstrated high specificity (>0.96) in detecting ToxA+B+CDT+ strains. The classifiers' performances for Hv strain detection were linked to high specificity (≥0.96). This study highlights MALDI-TOF MS enhanced by ML techniques as a rapid and cost-effective tool for identifying CD strain virulence factors. Our results brought a proof-of-concept concerning the ability of MALDI-TOF MS coupled with ML techniques to detect virulence factor and potentially improve the outbreak's management.

PMID:38850267 | DOI:10.1111/1751-7915.14478

Biochem Mol Biol Educ. 2024 Jun 8. doi: 10.1002/bmb.21843. Online ahead of print.

ABSTRACT

Analogies are used to make abstract topics meaningful and more easily comprehensible to learners. Incorporating simple analogies into STEM classrooms is a fairly common practice, but the analogies are typically generated and explained by the instructor for the learners. We hypothesize that challenging learners to create complex, extended analogies themselves can promote integration of content knowledge and development of critical thinking skills, which are essential for deep learning, but are challenging to teach. In this qualitative study, college biology students (n = 30) were asked to construct a complex analogy about the flow of genetic information using a familiar item. One week later, participants constructed a second analogy about the same topic, but this time using a more challenging item. Twenty participants worked on the challenging analogy in pairs, while the other 10 worked alone. Analysis of the 50 interviews resulted in a novel-scoring scheme, which measured both content knowledge (understanding of biology terms) and critical thinking (alignment of relationships between elements of the analogy). Most participants improved slightly due to practice, but they improved dramatically when working with a partner. The biggest gains were seen in critical thinking, not content knowledge. Having students construct complex, sophisticated analogies in pairs is a high-impact practice that can help students develop their critical thinking skills, which are crucial in academic and professional settings. The discussion between partners likely requires students to justify their explanations and critique their partner's explanations, which are characteristics of critical thinking.

PMID:38850246 | DOI:10.1002/bmb.21843

J Magn Reson Imaging. 2024 Jun 8. doi: 10.1002/jmri.29405. Online ahead of print.

ABSTRACT

BACKGROUND: Pathological complete response (pCR) is an essential criterion for adjusting follow-up treatment plans for patients with breast cancer (BC). The value of the visual geometry group and long short-term memory (VGG-LSTM) network using time-series dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) for pCR identification in BC is unclear.

PURPOSE: To identify pCR to neoadjuvant chemotherapy (NAC) using deep learning (DL) models based on the VGG-LSTM network.

STUDY TYPE: Retrospective.

POPULATION: Center A: 235 patients (47.7 ± 10.0 years) were divided 7:3 into training (n = 164) and validation set (n = 71). Center B: 150 patients (48.5 ± 10.4 years) were used as test set.

FIELD STRENGTH/SEQUENCE: 3-T, T2-weighted spin-echo sequence imaging, and gradient echo DCE sequence imaging.

ASSESSMENT: Patients underwent MRI examinations at three sequential time points: pretreatment, after three cycles of treatment, and prior to surgery, with tumor regions of interest manually delineated. Histopathology was the gold standard. We used VGG-LSTM network to establish seven DL models using time-series DCE-MR images: pre-NAC images (t0 model), early NAC images (t1 model), post-NAC images (t2 model), pre-NAC and early NAC images (t0 + t1 model), pre-NAC and post-NAC images (t0 + t2 model), pre-NAC, early NAC and post-NAC images (t0 + t1 + t2 model), and the optimal model combined with the clinical features and imaging features (combined model). The models were trained and optimized on the training and validation set, and tested on the test set.

STATISTICAL TESTS: The DeLong, Student's t-test, Mann-Whitney U, Chi-squared, Fisher's exact, Hosmer-Lemeshow tests, decision curve analysis, and receiver operating characteristics analysis were performed. P < 0.05 was considered significant.

RESULTS: Compared with the other six models, the combined model achieved the best performance in the test set yielding an AUC of 0.927.

DATA CONCLUSION: The combined model that used time-series DCE-MR images, clinical features and imaging features shows promise for identifying pCR in BC.

TECHNICAL EFFICACY: Stage 4.

PMID:38850180 | DOI:10.1002/jmri.29405

Mol Inform. 2024 Jun 8:e202400021. doi: 10.1002/minf.202400021. Online ahead of print.

ABSTRACT

Drug development is a long and costly process, often limited by the toxicity and adverse drug reactions (ADRs) caused by drug candidates. Even on the market, some drugs can cause strong ADRs that can vary depending on an individual polymorphism. The development of Genome-wide association studies (GWAS) allowed the discovery of genetic variants of interest that may cause these effects. In this study, the objective was to investigate a deep learning approach to predict genetic variations potentially related to ADRs. We used single nucleotide polymorphisms (SNPs) information from dbSNP to create a network based on ADR-drug-target-mutations and extracted matrixes of interaction to build deep Neural Networks (DNN) models. Considering only information about mutations known to impact drug efficacy and drug safety from PharmGKB and drug adverse reactions based on the MedDRA System Organ Classes (SOCs), these DNN models reached a balanced accuracy of 0.61 in average. Including molecular fingerprints representing structural features of the drugs did not improve the performance of the models. To our knowledge, this is the first model that exploits DNN to predict ADR-drug-target-mutations. Although some improvements are suggested, these models can be of interest to analyze multiple compounds over all of the genes and polymorphisms information accessible and thus pave the way in precision medicine.

PMID:38850150 | DOI:10.1002/minf.202400021

Yeast. 2024 Jun 8. doi: 10.1002/yea.3965. Online ahead of print.

ABSTRACT

Meiotic crossovers play a vital role in proper chromosome segregation and evolution of most sexually reproducing organisms. Meiotic recombination can be visually observed in Saccharomyces cerevisiae tetrads using linked spore-autonomous fluorescent markers placed at defined intervals within the genome, which allows for analysis of meiotic segregation without the need for tetrad dissection. To automate the analysis, we developed a deep learning-based image recognition and classification pipeline for high-throughput tetrad detection and meiotic crossover classification. As a proof of concept, we analyzed a large image data set from wild-type and selected gene knock-out mutants to quantify crossover frequency, interference, chromosome missegregation, and gene conversion events. The deep learning-based method has the potential to accelerate the discovery of new genes involved in meiotic recombination in S. cerevisiae such as the underlying factors controlling crossover frequency and interference.

PMID:38850080 | DOI:10.1002/yea.3965

J Cheminform. 2024 Jun 7;16(1):66. doi: 10.1186/s13321-024-00865-6.

ABSTRACT

Accurate ligand binding site prediction (LBSP) within proteins is essential for drug discovery. We developed ProteinUNetResNetV2.0 (PUResNetV2.0), leveraging sparse representation of protein structures to improve LBSP accuracy. Our training dataset included protein complexes from 4729 protein families. Evaluations on benchmark datasets showed that PUResNetV2.0 achieved an 85.4% Distance Center Atom (DCA) success rate and a 74.7% F1 Score on the Holo801 dataset, outperforming existing methods. However, its performance in specific cases, such as RNA, DNA, peptide-like ligand, and ion binding site prediction, was limited due to constraints in our training data. Our findings underscore the potential of sparse representation in LBSP, especially for oligomeric structures, suggesting PUResNetV2.0 as a promising tool for computational drug discovery.

PMID:38849917 | DOI:10.1186/s13321-024-00865-6

J Cheminform. 2024 Jun 7;16(1):67. doi: 10.1186/s13321-024-00862-9.

ABSTRACT

Identification of interactions between chemical compounds and proteins is crucial for various applications, including drug discovery, target identification, network pharmacology, and elucidation of protein functions. Deep neural network-based approaches are becoming increasingly popular in efficiently identifying compound-protein interactions with high-throughput capabilities, narrowing down the scope of candidates for traditional labor-intensive, time-consuming and expensive experimental techniques. In this study, we proposed an end-to-end approach termed SPVec-SGCN-CPI, which utilized simplified graph convolutional network (SGCN) model with low-dimensional and continuous features generated from our previously developed model SPVec and graph topology information to predict compound-protein interactions. The SGCN technique, dividing the local neighborhood aggregation and nonlinearity layer-wise propagation steps, effectively aggregates K-order neighbor information while avoiding neighbor explosion and expediting training. The performance of the SPVec-SGCN-CPI method was assessed across three datasets and compared against four machine learning- and deep learning-based methods, as well as six state-of-the-art methods. Experimental results revealed that SPVec-SGCN-CPI outperformed all these competing methods, particularly excelling in unbalanced data scenarios. By propagating node features and topological information to the feature space, SPVec-SGCN-CPI effectively incorporates interactions between compounds and proteins, enabling the fusion of heterogeneity. Furthermore, our method scored all unlabeled data in ChEMBL, confirming the top five ranked compound-protein interactions through molecular docking and existing evidence. These findings suggest that our model can reliably uncover compound-protein interactions within unlabeled compound-protein pairs, carrying substantial implications for drug re-profiling and discovery. In summary, SPVec-SGCN demonstrates its efficacy in accurately predicting compound-protein interactions, showcasing potential to enhance target identification and streamline drug discovery processes.Scientific contributionsThe methodology presented in this work not only enables the comparatively accurate prediction of compound-protein interactions but also, for the first time, take sample imbalance which is very common in real world and computation efficiency into consideration simultaneously, accelerating the target identification and drug discovery process.

PMID:38849874 | DOI:10.1186/s13321-024-00862-9

Clin Oral Investig. 2024 Jun 8;28(7):364. doi: 10.1007/s00784-024-05762-8.

ABSTRACT

OBJECTIVES: Diagnosing oral potentially malignant disorders (OPMD) is critical to prevent oral cancer. This study aims to automatically detect and classify the most common pre-malignant oral lesions, such as leukoplakia and oral lichen planus (OLP), and distinguish them from oral squamous cell carcinomas (OSCC) and healthy oral mucosa on clinical photographs using vision transformers.

METHODS: 4,161 photographs of healthy mucosa, leukoplakia, OLP, and OSCC were included. Findings were annotated pixel-wise and reviewed by three clinicians. The photographs were divided into 3,337 for training and validation and 824 for testing. The training and validation images were further divided into five folds with stratification. A Mask R-CNN with a Swin Transformer was trained five times with cross-validation, and the held-out test split was used to evaluate the model performance. The precision, F1-score, sensitivity, specificity, and accuracy were calculated. The area under the receiver operating characteristics curve (AUC) and the confusion matrix of the most effective model were presented.

RESULTS: The detection of OSCC with the employed model yielded an F1 of 0.852 and AUC of 0.974. The detection of OLP had an F1 of 0.825 and AUC of 0.948. For leukoplakia the F1 was 0.796 and the AUC was 0.938.

CONCLUSIONS: OSCC were effectively detected with the employed model, whereas the detection of OLP and leukoplakia was moderately effective.

CLINICAL RELEVANCE: Oral cancer is often detected in advanced stages. The demonstrated technology may support the detection and observation of OPMD to lower the disease burden and identify malignant oral cavity lesions earlier.

PMID:38849649 | DOI:10.1007/s00784-024-05762-8

Int J Comput Assist Radiol Surg. 2024 Jun 7. doi: 10.1007/s11548-024-03205-z. Online ahead of print.

ABSTRACT

OBJECTIVES: In patients having naïve glioblastoma multiforme (GBM), this study aims to assess the efficacy of Deep Learning algorithms in automating the segmentation of brain magnetic resonance (MR) images to accurately determine 3D masks for 4 distinct regions: enhanced tumor, peritumoral edema, non-enhanced/necrotic tumor, and total tumor.

MATERIAL AND METHODS: A 3D U-Net neural network algorithm was developed for semantic segmentation of GBM. The training dataset was manually delineated by a group of expert neuroradiologists on MR images from the Brain Tumor Segmentation Challenge 2021 (BraTS2021) image repository, as ground truth labels for diverse glioma (GBM and low-grade glioma) subregions across four MR sequences (T1w, T1w-contrast enhanced, T2w, and FLAIR) in 1251 patients. The in-house test was performed on 50 GBM patients from our cohort (PerProGlio project). By exploring various hyperparameters, the network's performance was optimized, and the most optimal parameter configuration was identified. The assessment of the optimized network's performance utilized Dice scores, precision, and sensitivity metrics.

RESULTS: Our adaptation of the 3D U-net with additional residual blocks demonstrated reliable performance on both the BraTS2021 dataset and the in-house PerProGlio cohort, employing only T1w-ce sequences for enhancement and non-enhanced/necrotic tumor models and T1w-ce + T2w + FLAIR for peritumoral edema and total tumor. The mean Dice scores (training and test) were 0.89 and 0.75; 0.75 and 0.64; 0.79 and 0.71; and 0.60 and 0.55, for total tumor, edema, enhanced tumor, and non-enhanced/necrotic tumor, respectively.

CONCLUSIONS: The results underscore the high precision with which our network can effectively segment GBM tumors and their distinct subregions. The level of accuracy achieved agrees with the coefficients recorded in previous GBM studies. In particular, our approach allows model specialization for each of the different tumor subregions employing only those MR sequences that provide value for segmentation.

PMID:38849632 | DOI:10.1007/s11548-024-03205-z

Nat Rev Clin Oncol. 2024 Jun 7. doi: 10.1038/s41571-024-00909-8. Online ahead of print.

ABSTRACT

Artificial intelligence (AI) stands at the threshold of revolutionizing clinical oncology, with considerable potential to improve early cancer detection and risk assessment, and to enable more accurate personalized treatment recommendations. However, a notable imbalance exists in the distribution of the benefits of AI, which disproportionately favour those living in specific geographical locations and in specific populations. In this Perspective, we discuss the need to foster the development of equitable AI tools that are both accurate in and accessible to a diverse range of patient populations, including those in low-income to middle-income countries. We also discuss some of the challenges and potential solutions in attaining equitable AI, including addressing the historically limited representation of diverse populations in existing clinical datasets and the use of inadequate clinical validation methods. Additionally, we focus on extant sources of inequity including the type of model approach (such as deep learning, and feature engineering-based methods), the implications of dataset curation strategies, the need for rigorous validation across a variety of populations and settings, and the risk of introducing contextual bias that comes with developing tools predominantly in high-income countries.

PMID:38849530 | DOI:10.1038/s41571-024-00909-8

Commun Biol. 2024 Jun 7;7(1):702. doi: 10.1038/s42003-024-06371-7.

ABSTRACT

The Drosophila model is pivotal in deciphering the pathophysiological underpinnings of various human ailments, notably aging and cardiovascular diseases. Cutting-edge imaging techniques and physiology yield vast high-resolution videos, demanding advanced analysis methods. Our platform leverages deep learning to segment optical microscopy images of Drosophila hearts, enabling the quantification of cardiac parameters in aging and dilated cardiomyopathy (DCM). Validation using experimental datasets confirms the efficacy of our aging model. We employ two innovative approaches deep-learning video classification and machine-learning based on cardiac parameters to predict fly aging, achieving accuracies of 83.3% (AUC 0.90) and 79.1%, (AUC 0.87) respectively. Moreover, we extend our deep-learning methodology to assess cardiac dysfunction associated with the knock-down of oxoglutarate dehydrogenase (OGDH), revealing its potential in studying DCM. This versatile approach promises accelerated cardiac assays for modeling various human diseases in Drosophila and holds promise for application in animal and human cardiac physiology under diverse conditions.

PMID:38849449 | DOI:10.1038/s42003-024-06371-7

Sci Rep. 2024 Jun 7;14(1):13163. doi: 10.1038/s41598-024-63611-w.

ABSTRACT

Pear pollination is performed by artificial pollination because the pollination rate through insect pollination is not stable. Pollen must be collected to secure sufficient pollen for artificial pollination. However, recently, collecting sufficient amounts of pollen in Japan has become difficult, resulting in increased imports from overseas. To solve this problem, improving the efficiency of pollen collection and strengthening the domestic supply and demand system is necessary. In this study, we proposed an Artificial Intelligence (AI)-based method to estimate the amount of pear pollen. The proposed method used a deep learning-based object detection algorithm, You Only Look Once (YOLO), to classify and detect flower shapes in five stages, from bud to flowering, and to estimate the pollen amount. In this study, the performance of the proposed method was discussed by analyzing the accuracy and error of classification for multiple flower varieties. Although this study only discussed the performance of estimating the amount of pollen collected, in the future, we aim to establish a technique for estimating the time of maximum pollen collection using the method proposed in this study.

PMID:38849427 | DOI:10.1038/s41598-024-63611-w

Sci Rep. 2024 Jun 7;14(1):13126. doi: 10.1038/s41598-024-63776-4.

ABSTRACT

In human-computer interaction systems, speech emotion recognition (SER) plays a crucial role because it enables computers to understand and react to users' emotions. In the past, SER has significantly emphasised acoustic properties extracted from speech signals. The use of visual signals for enhancing SER performance, however, has been made possible by recent developments in deep learning and computer vision. This work utilizes a lightweight Vision Transformer (ViT) model to propose a novel method for improving speech emotion recognition. We leverage the ViT model's capabilities to capture spatial dependencies and high-level features in images which are adequate indicators of emotional states from mel spectrogram input fed into the model. To determine the efficiency of our proposed approach, we conduct a comprehensive experiment on two benchmark speech emotion datasets, the Toronto English Speech Set (TESS) and the Berlin Emotional Database (EMODB). The results of our extensive experiment demonstrate a considerable improvement in speech emotion recognition accuracy attesting to its generalizability as it achieved 98%, 91%, and 93% (TESS-EMODB) accuracy respectively on the datasets. The outcomes of the comparative experiment show that the non-overlapping patch-based feature extraction method substantially improves the discipline of speech emotion recognition. Our research indicates the potential for integrating vision transformer models into SER systems, opening up fresh opportunities for real-world applications requiring accurate emotion recognition from speech compared with other state-of-the-art techniques.

PMID:38849422 | DOI:10.1038/s41598-024-63776-4

Sci Rep. 2024 Jun 7;14(1):13122. doi: 10.1038/s41598-024-63656-x.

ABSTRACT

Deep learning-based methods have demonstrated high classification performance in the detection of cardiovascular diseases from electrocardiograms (ECGs). However, their blackbox character and the associated lack of interpretability limit their clinical applicability. To overcome existing limitations, we present a novel deep learning architecture for interpretable ECG analysis (xECGArch). For the first time, short- and long-term features are analyzed by two independent convolutional neural networks (CNNs) and combined into an ensemble, which is extended by methods of explainable artificial intelligence (xAI) to whiten the blackbox. To demonstrate the trustworthiness of xECGArch, perturbation analysis was used to compare 13 different xAI methods. We parameterized xECGArch for atrial fibrillation (AF) detection using four public ECG databases ( n = 9854 ECGs) and achieved an F1 score of 95.43% in AF versus non-AF classification on an unseen ECG test dataset. A systematic comparison of xAI methods showed that deep Taylor decomposition provided the most trustworthy explanations ( + 24 % compared to the second-best approach). xECGArch can account for short- and long-term features corresponding to clinical features of morphology and rhythm, respectively. Further research will focus on the relationship between xECGArch features and clinical features, which may help in medical applications for diagnosis and therapy.

PMID:38849417 | DOI:10.1038/s41598-024-63656-x

EBioMedicine. 2024 Jun 6;104:105183. doi: 10.1016/j.ebiom.2024.105183. Online ahead of print.

ABSTRACT

BACKGROUND: Contrast-enhanced CT scans provide a means to detect unsuspected colorectal cancer. However, colorectal cancers in contrast-enhanced CT without bowel preparation may elude detection by radiologists. We aimed to develop a deep learning (DL) model for accurate detection of colorectal cancer, and evaluate whether it could improve the detection performance of radiologists.

METHODS: We developed a DL model using a manually annotated dataset (1196 cancer vs 1034 normal). The DL model was tested using an internal test set (98 vs 115), two external test sets (202 vs 265 in 1, and 252 vs 481 in 2), and a real-world test set (53 vs 1524). We compared the detection performance of the DL model with radiologists, and evaluated its capacity to enhance radiologists' detection performance.

FINDINGS: In the four test sets, the DL model had the area under the receiver operating characteristic curves (AUCs) ranging between 0.957 and 0.994. In both the internal test set and external test set 1, the DL model yielded higher accuracy than that of radiologists (97.2% vs 86.0%, p < 0.0001; 94.9% vs 85.3%, p < 0.0001), and significantly improved the accuracy of radiologists (93.4% vs 86.0%, p < 0.0001; 93.6% vs 85.3%, p < 0.0001). In the real-world test set, the DL model delivered sensitivity comparable to that of radiologists who had been informed about clinical indications for most cancer cases (94.3% vs 96.2%, p > 0.99), and it detected 2 cases that had been missed by radiologists.

INTERPRETATION: The developed DL model can accurately detect colorectal cancer and improve radiologists' detection performance, showing its potential as an effective computer-aided detection tool.

FUNDING: This study was supported by National Science Fund for Distinguished Young Scholars of China (No. 81925023); Regional Innovation and Development Joint Fund of National Natural Science Foundation of China (No. U22A20345); National Natural Science Foundation of China (No. 82072090 and No. 82371954); Guangdong Provincial Key Laboratory of Artificial Intelligence in Medical Image Analysis and Application (No. 2022B1212010011); High-level Hospital Construction Project (No. DFJHBF202105).

PMID:38848616 | DOI:10.1016/j.ebiom.2024.105183

IEEE/ACM Trans Comput Biol Bioinform. 2024 Jun 7;PP. doi: 10.1109/TCBB.2024.3411024. Online ahead of print.

ABSTRACT

Controlling the gene expression is the most important development in a living organism, which makes it easier to find different kinds of diseases and their causes. It's very difficult to know what factors control the gene expression. Transcription Factor (TF) is a protein that plays an important role in gene expression. Discovering the transcription factor has immense biological significance, however, it is challenging to develop novel techniques and evaluation for regulatory developments in biological structures. In this research, we mainly focus on 'sequence specificities' that can be ascertained from experimental data with 'deep learning' techniques, which offer a scalable, flexible and unified computational approach for predicting transcription factor binding. Specifically, Multiple Expression motifs for Motif Elicitation (MEME) technique with Convolution Neural Network (CNN) named as CnNet, has been used for discovering the 'sequence specificities' of DNA gene sequences dataset. This process involves two steps: a) discovering the motifs that are capable of identifying useful TF binding site by using MEME technique, and b) computing a score indicating the likelihood of a given sequence being a useful binding site by using CNN technique. The proposed CnNet approach predicts the TF binding score with much better accuracy compared to existing approaches. The source code and datasets used in this work are available at https://github.com/masoodbai/CnNet-Approach-for-TFBS.git.

PMID:38848239 | DOI:10.1109/TCBB.2024.3411024

IEEE Trans Pattern Anal Mach Intell. 2024 Jun 7;PP. doi: 10.1109/TPAMI.2024.3411292. Online ahead of print.

ABSTRACT

We propose a framework that combines traditional, hand-crafted algorithms and recent advances in deep learning to obtain high-quality, high-resolution disparity maps from stereo images. By casting the refinement process as a continuous feature sampling strategy, our neural disparity refinement network can estimate an enhanced disparity map at any output resolution. Our solution can process any disparity map produced by classical stereo algorithms, as well as those predicted by modern stereo networks or even different depth-from-images approaches, such as the COLMAP structure-from-motion pipeline. Nonetheless, when deployed in the former configuration, our framework performs at its best in terms of zero-shot generalization from synthetic to real images. Moreover, its continuous formulation allows for easily handling the unbalanced stereo setup very diffused in mobile phones.

PMID:38848234 | DOI:10.1109/TPAMI.2024.3411292

IEEE J Biomed Health Inform. 2024 Jun 7;PP. doi: 10.1109/JBHI.2024.3411056. Online ahead of print.

ABSTRACT

Due to their convenience, adhesive patch-type electrocardiographs are commonly used for arrhythmia screening. This study aimed to develop a reliable method that can improve the classification performance of atrial fibrillation (AF) using these devices based on the 2020 European Society of Cardiology (ESC) guidelines for AF diagnosis in clinical practice. We developed a deep learning model that utilizes RR interval frames for precise, beat-wise classification of electrocardiogram (ECG) signals. This model is specifically designed to sequentially classify each R peak on the ECG, considering the rhythms surrounding each beat. It features a two-stage bidirectional Recurrent Neural Network (RNN) with a many-to-many architecture, which is particularly optimized for processing sequential and time-series data. The structure aims to extract local features and capture long-term dependencies associated with AF. After inference, outputs which indicating either AF or non-AF, derived from various temporal sequences are combined through an ensembling technique to enhance prediction accuracy. We collected AF data from a clinical trial that utilized the MEMO Patch, an adhesive patch-type electrocardiograph. When trained on public databases, the model demonstrated high accuracy on the patch dataset (accuracy: 0.986, precision: 0.981, sensitivity: 0.979, specificity: 0.992, and F1 score: 0.98), maintaining consistent performance across public datasets. SeqAFNet was robust for AF classification, making it a potential tool in real-world applications.

PMID:38848232 | DOI:10.1109/JBHI.2024.3411056

IEEE Trans Neural Netw Learn Syst. 2024 Jun 7;PP. doi: 10.1109/TNNLS.2024.3404276. Online ahead of print.

ABSTRACT

Accurate teeth delineation on 3-D dental models is essential for individualized orthodontic treatment planning. Pioneering works like PointNet suggest a promising direction to conduct efficient and accurate 3-D dental model analyses in end-to-end learnable fashions. Recent studies further imply that multistream architectures to concurrently learn geometric representations from different inputs/views (e.g., coordinates and normals) are beneficial for segmenting teeth with varying conditions. However, such multistream networks typically adopt simple late-fusion strategies to combine features captured from raw inputs that encode complementary but fundamentally different geometric information, potentially hampering their accuracy in end-to-end semantic segmentation. This article presents a hierarchical cross-stream aggregation (HiCA) network to learn more discriminative point/cell-wise representations from multiview inputs for fine-grained 3-D semantic segmentation. Specifically, based upon our multistream backbone with input-tailored feature extractors, we first design a contextual cross-steam aggregation (CA) module conditioned on interstream consistency to boost each view's contextual representation learning jointly. Then, before the late fusion of different streams' outputs for segmentation, we further deploy a discriminative cross-stream aggregation (DA) module to concurrently update all views' discriminative representation learning by leveraging a specific graph attention strategy induced by multiview prototype learning. On both public and in-house datasets of real-patient dental models, our method significantly outperformed state-of-the-art (SOTA) deep learning methods for teeth semantic segmentation. In addition, extended experimental results suggest the applicability of HiCA to other general 3-D shape segmentation tasks. The code is available at https://github.com/ladderlab-xjtu/HiCA.

PMID:38848227 | DOI:10.1109/TNNLS.2024.3404276

IEEE Trans Neural Syst Rehabil Eng. 2024;32:2096-2105. doi: 10.1109/TNSRE.2024.3394738.

ABSTRACT

Sleep staging serves as a fundamental assessment for sleep quality measurement and sleep disorder diagnosis. Although current deep learning approaches have successfully integrated multimodal sleep signals, enhancing the accuracy of automatic sleep staging, certain challenges remain, as follows: 1) optimizing the utilization of multi-modal information complementarity, 2) effectively extracting both long- and short-range temporal features of sleep information, and 3) addressing the class imbalance problem in sleep data. To address these challenges, this paper proposes a two-stream encode-decoder network, named TSEDSleepNet, which is inspired by the depth sensitive attention and automatic multi-modal fusion (DSA2F) framework. In TSEDSleepNet, a two-stream encoder is used to extract the multiscale features of electrooculogram (EOG) and electroencephalogram (EEG) signals. And a self-attention mechanism is utilized to fuse the multiscale features, generating multi-modal saliency features. Subsequently, the coarser-scale construction module (CSCM) is adopted to extract and construct multi-resolution features from the multiscale features and the salient features. Thereafter, a Transformer module is applied to capture both long- and short-range temporal features from the multi-resolution features. Finally, the long- and short-range temporal features are restored with low-layer details and mapped to the predicted classification results. Additionally, the Lovász loss function is applied to alleviate the class imbalance problem in sleep datasets. Our proposed method was tested on the Sleep-EDF-39 and Sleep-EDF-153 datasets, and it achieved classification accuracies of 88.9% and 85.2% and Macro-F1 scores of 84.8% and 79.7%, respectively, thus outperforming conventional traditional baseline models. These results highlight the efficacy of the proposed method in fusing multi-modal information. This method has potential for application as an adjunct tool for diagnosing sleep disorders.

PMID:38848223 | DOI:10.1109/TNSRE.2024.3394738