Strain Parameters Research Articles

Global and local feature extraction based on convolutional neural network residual learning for MR image denoising.

Given the different noise distribution information of global and local magnetic resonance (MR) images, this study aims to extend the current work on convolutional neural networks that preserve global structure and local details in MR image denoising tasks. This study proposed a parallel and serial network for denoising 3D MR images, called 3D-PSNet. We use the residual depthwise separable convolution block to learn the local information of the feature map, reduce the network parameters, and thus improve the training speed and parameter efficiency. In addition, we consider the feature extraction of the global image and utilize residual dilated convolution to process the feature map to expand the receptive field of the network and avoid the loss of global information. Finally, we combine both of them to form a parallel network. What's more, we integrate reinforced residual convolution blocks with dense connections to form serial network branches, which can remove redundant information and refine features to further obtain accurate noise information. The peak signal-to-noise ratio, structural similarity index measure, and root mean square error metrics of 3D-PSNet are as high as 47.79%, 99.81%, and 0.40%, respectively, achieving competitive denoising effect on three public datasets. The ablation experiments demonstrated the effectiveness of all the designed modules regarding all the evaluated metrics in both datasets. The proposed 3D-PSNet takes advantage of multi-scale receptive fields, local feature extraction and residual dense connections to more effectively restore the global structure and local fine features in MR images, and is expected to help doctors quickly and accurately diagnose patients' conditions.

Evaluation of the strength of the truck tractor’s frame under emergency braking conditions

Authors carried out a theoretical study of the force impact during emergency braking on the strength of the frame of a truck tractor as a part of a road train with a semi-trailer in order to build a model of the dependence of the safety factor index of the frame on the geometric and weight parameters of the car, semi-trailer, transported cargo and the type of the pavement. Using D'Alembert's principle, a generalized calculation scheme of the frame of the semi-trailer and the truck unit after reaching a steady deceleration during emergency braking while maintaining straight-line movement and adhesion of all wheels to pavement was made. Through a static calculation, the dependences for the unknown reactions of the car frame constraints were obtained, its dangerous cross-section was established, and the condition for the static strength of the longeron was determined. Using this condition and the obtained ratios for active and reactive loads, a model of the dependence of the safety margin coefficient of the car frame on the road train parameters was built. By the example of a road train consisting of a KrAZ-6446 truck tractor and a VARZ-6006 semi-trailer the application of the constructed model for evaluating the strength of the car frame and determining the permissible safe zone location of the center of gravity of different weights cargo in the plane of the road train is demonstrated. The method proposed in this work and the built model can be used to study the strength of the frames of truck tractors and semi-trailers of other manufacturers and types and to establish safe schemes for the placement and fastening of loads to ensure the operational reliability of the vehicle and the safety and security of cargo transportation.

Dilated multilevel fused network for virus classification using transmission electron microscopy images

Previous studies have demonstrated significant performance in the field of virus classification; however, they focused on the classification of a small number of virus classes, with a maximum of 16 classes. To address this limitation, this study aims to create a deep learning-based network that outperforms the state-of-the-art (SOTA) models for the classification of 22 different virus classes with the fewest possible trainable parameters. We introduce an automatic identification system for virus classes based on our classification-driven retrieval framework. The proposed dilated multilevel fused network (DMLF-Net) utilizes the multilevel feature fusion concept within a network to exploit more abstract features for microscopic data analysis. A multi-stage training strategy was applied to achieve optimal model convergence without overfitting the training data.We evaluated the performance of the DMLF-Net on three open databases including two virus datasets and one bacteria species dataset. The results demonstrated an accuracy of 89.89%, a weighted harmonic mean of precision and recall (F1-score) of 83.39%, and an area under the curve (AUC) of 92.50% for the 1st virus dataset. For the 2nd virus dataset, the accuracy was 80.70%, the F1-score was 81.20%, and the AUC was 86.20%. For the 3rd bacteria species dataset, the accuracy was 95.93% and the F1-score was 96.24%. DMLF-Net outperforms SOTA methods in terms of classification accuracy while utilizing nearly 5.3 times fewer trainable parameters (25.5 million) compared to the second-best model, visual geometry group (VGG)16 (134.3 million).

An enzyme-inspired specificity in deep learning model for sleep stage classification using multi-channel PSG signals input: Separating training approach and its performance on cross-dataset validation for generalizability

Numerous automatic sleep stage classification systems have been developed, but none have become effective assistive tools for sleep technicians due to issues with generalization. Four key factors hinder the generalization of these models are instruments, montage of recording, subject type, and scoring manual factors. This study aimed to develop a deep learning model that addresses generalization problems by integrating enzyme-inspired specificity and employing separating training approaches. Subject type and scoring manual factors were controlled, while the focus was on instruments and montage of recording factors. The proposed model consists of three sets of signal-specific models including EEG-, EOG-, and EMG-specific model. The EEG-specific models further include three sets of channel-specific models. All signal-specific and channel-specific models were established with data manipulation and weighted loss strategies, resulting in three sets of data manipulation models and class-specific models, respectively. These models were CNNs. Additionally, BiLSTM models were applied to EEG- and EOG-specific models to obtain temporal information. Finally, classification task for sleep stage was handled by ‘the-last-dense’ layer. The optimal sampling frequency for each physiological signal was identified and used during the training process. The proposed model was trained on MGH dataset and evaluated using both within dataset and cross-dataset. For MGH dataset, overall accuracy of 81.05 %, MF1 of 79.05 %, Kappa of 0.7408, and per-class F1-scores: W (84.98 %), N1 (58.06 %), N2 (84.82 %), N3 (79.20 %), and REM (88.17 %) can be achieved. Performances on cross-datasets are as follows: SHHS1 200 records reached 79.54 %, 70.56 %, and 0.7078; SHHS2 200 records achieved 76.77 %, 66.30 %, and 0.6632; Sleep-EDF 153 records gained 78.52 %, 72.13 %, and 0.7031; and BCI-MU (local dataset) 94 records achieved 83.57 %, 82.17 %, and 0.7769 for overall accuracy, MF1, and Kappa respectively. Additionally, the proposed model has approximately 9.3 M trainable parameters and takes around 26 s to process one PSG record. The results indicate that the proposed model demonstrates generalizability in sleep stage classification and shows potential as a feasibility tool for real-world applications. Additionally, enzyme-inspired specificity effectively addresses the challenges posed by varying montage of recording, while the identified optimal frequencies mitigate instrument-related issues.

Growth of Lactiplantibacillus plantarum BG112 in Batch and Continuous Culture with Camellia sinensis as Prebiotic

This work aimed to study the effect of Camellia sinensis extract (CSExt) as a particular growth promoter of Lactiplantibacillus plantarum (LP) in batch and continuous production processes. Growth conditions were 1% (v/v) inoculum, pHC = 6.5, 1% of dissolved oxygen (D.O.), 37 °C, and 150 rpm in a 0.2 L bioreactor using a commercial MRS broth (de Man, Rogosa, and Sharpe) and 1% (v/v) or 10% (v/v) CSExt according to the experimental design. In batch experiments, the maximum specific growth rate and the affinity constant increased with the increase in CSExt. In continuous culture, biomass production increased significantly with the addition of 1% (w/v) CSExt at 0.15 (1/h). Kinetic parameters adjusted were similar to those reported in the literature. Substrate affinity and the specific growth rate increased significantly in the presence of CSExt in batch and continuous cultures. Based on the results, prebiotics from plant extracts may function as growth promoters in specific physiological stages. This is the first report showing the change in kinetic parameters of a probiotic strain growing in crude plant extracts.

Incremental prognostic value of left atrial strain in apical hypertrophic cardiomyopathy: a cardiovascular magnetic resonance study.

This study aimed to evaluate the prognostic value of left atrial (LA) strain in patients with apical hypertrophic cardiomyopathy (ApHCM), as assessed by cardiac magnetic resonance (CMR) imaging. Four hundred and five consecutive patients with ApHCM who underwent CMR examination were retrospectively included. The study endpoint included all-cause death, heart transplant, aborted sudden cardiac death, hospitalization for heart failure, stroke, and new-onset atrial fibrillation (AF). After a median follow-up of 97 months, 75 patients (18.5%) reached the endpoint. Patients were divided into two groups based on the median LA reservoir strain of 29.4%. The group with lower LA reservoir strain had thicker maximum wall thickness, greater late gadolinium enhancement extent, and smaller end-diastolic volume index, stroke volume index, and cardiac index (all p < 0.02). For LA parameters, this subgroup showed greater diameter and volume index and worse ejection fraction, reservoir, conduit, and booster strain (all p < 0.001). In the multivariable model, age (HR 1.88, 95% CI: 1.06-3.31, p = 0.030), baseline AF (HR 2.95, 95% CI: 1.64-5.28, p < 0.001), LA volume index (LAVi) (HR 2.07, 95% CI: 1.21-3.55, p = 0.008) and LA reservoir strain (HR 2.82, 95% CI: 1.51-5.26, p = 0.001) were all associated with the outcome. Adding LAVi and LA reservoir strain in turn to the multivariable model (age and baseline AF) resulted in significant improvements in model performance (p < 0.001). In ApHCM patients, LA reservoir strain is independently associated with cardiovascular risk events and has an incremental prognostic value. Left atrial reservoir strain measured by cardiac magnetic resonance is highly correlated with the prognosis of apical hypertrophic cardiomyopathy and has potential incremental value in the prognosis of major adverse cardiac events. Left atrial (LA) strain parameters may be useful for risk stratification and treatment of apical hypertrophic cardiomyopathy (ApHCM). Apical hypertrophic cardiomyopathy (ApHCM) is independently associated with LA morphology and function. Cardiac MR examination, especially its feature-tracking technology, provides the possibility to prognosticate ApHCM at an early stage.

Optimizing poultry audio signal classification with deep learning and burn layer fusion

This study introduces a novel deep learning-based approach for classifying poultry audio signals, incorporating a custom Burn Layer to enhance model robustness. The methodology integrates digital audio signal processing, convolutional neural networks (CNNs), and the innovative Burn Layer, which injects controlled random noise during training to reinforce the model's resilience to input signal variations. The proposed architecture is streamlined, with convolutional blocks, densely connected layers, dropout, and an additional Burn Layer to fortify robustness. The model demonstrates efficiency by reducing trainable parameters to 191,235, compared to traditional architectures with over 1.7 million parameters. The proposed model utilizes a Burn Layer with burn intensity as a parameter and an Adamax optimizer to optimize and address the overfitting problem. Thorough evaluation using six standard classification metrics showcases the model's superior performance, achieving exceptional sensitivity (96.77%), specificity (100.00%), precision (100.00%), negative predictive value (NPV) (95.00%), accuracy (98.55%), F1 score (98.36%), and Matthew’s correlation coefficient (MCC) (95.88%). This research contributes valuable insights into the fields of audio signal processing, animal health monitoring, and robust deep-learning classification systems. The proposed model presents a systematic approach for developing and evaluating a deep learning-based poultry audio classification system. It processes raw audio data and labels to generate digital representations, utilizes a Burn Layer for training variability, and constructs a CNN model with convolutional blocks, pooling, and dense layers. The model is optimized using the Adamax algorithm and trained with data augmentation and early-stopping techniques. Rigorous assessment on a test dataset using standard metrics demonstrates the model's robustness and efficiency, with the potential to significantly advance animal health monitoring and disease detection through audio signal analysis.

Visuospatial Skills Explain Differences in the Ability to Use Propulsion Biofeedback Post-stroke.

Visual biofeedback can be used to help people post-stroke reduce biomechanical gait impairments. Using visual biofeedback engages an explicit, cognitively demanding motor learning process. Participants with better overall cognitive function are better able to use visual biofeedback to promote locomotor learning; however, which specific cognitive domains are responsible for this effect are unknown. We aimed to understand which cognitive domains were associated with performance during acquisition and immediate retention when using visual biofeedback to increase paretic propulsion in individuals post-stroke. Participants post-stroke completed cognitive testing, which provided scores for different cognitive domains, including executive function, immediate memory, visuospatial/constructional skills, language, attention, and delayed memory. Next, participants completed a single session of paretic propulsion biofeedback training, where we collected treadmill-walking data for 20min with biofeedback and 2min without biofeedback. We fit separate regression models to determine if cognitive domain scores, motor impairment (measured with the lower-extremity Fugl-Meyer), and gait speed could explain propulsion error and variability during biofeedback use and recall error during immediate retention. Visuospatial/constructional skills and motor impairment best-explained propulsion error during biofeedback use (adjusted R 2 =0.56, P =0.0008), and attention best-explained performance variability (adjusted R 2 =0.17, P =0.048). Language skills best-explained recall error during immediate retention (adjusted R 2 =0.37, P =0.02). These results demonstrate that specific cognitive domain impairments explain variability in locomotor learning outcomes in individuals with chronic stroke. This suggests that with further investigation, specific cognitive impairment information may be useful to predict responsiveness to interventions and personalize training parameters to facilitate locomotor learning.

A new lightweight framework based on knowledge distillation for reducing the complexity of multi-modal solar irradiance prediction model

The inherent uncertainty of solar energy brings great difficulties to the grid connection and short-term energy planning and dispatching. Deep learning method makes it possible to predict the short-term solar energy with its powerful learning ability, but its complex model structure and huge trainable parameters bring great difficulties to the practical deployment. Therefore, this paper proposes a lightweight framework based on knowledge distillation strategy, which greatly reduces the complexity of multi-modal solar irradiance prediction model meanwhile ensuring an acceptable accuracy, facilitating the practical deployment. Firstly, a teacher model with multi-modal structure and good accuracy is built based on ResNet18-Informer. Then, the lightweight model is obtained by the proposed lightweight framework depending on the knowledge of teacher model. The comparisons of various models and the optimal settings of knowledge distillation are analyzed. Results show that the lightweight model can reduce the trainable parameters, inference time, and GPU memory by 97.7%, 52.5%, and 36.3%, respectively. The normalized root mean square error is reduced by 24.87% compared with the same structure model but without knowledge distillation, verifying the superiority of the proposed framework. The soft loss using the light loss with the ratio of 0.3 can obtain the best training results for the lightweight model. The structure with 3 residual blocks and 3 LSTM layers is proved to be the best for the lightweight model in the solar irradiance prediction task.

The effects of myocardial bridging on two-dimensional myocardial strain during dobutamine stress echocardiography.

Myocardial bridging (MB) is a common anatomic variant in coronary arteries with unclear functional significance. We evaluated regional myocardial strain by speckle tracking during dobutamine stress echocardiography (DSE) in patients with MB in the left anterior descending coronary artery (LAD). We studied 11 patients with MB in the LAD and no obstructive coronary artery disease (CAD), 7 patients without MB, but obstructive CAD in the LAD, and 12 controls without MB or obstructive CAD. MB was defined as either > 1mm (superficial) or > 2mm (deep) intramyocardial course of the LAD in coronary CT angiography. Regional longitudinal, radial and circumferential strains and strain rates as well as post-systolic strain index (PSI) were measured at rest, peak stress, and early recovery (1min after stress). Strain parameters during DSE were similar in the myocardium distal to MB and other myocardial regions of the same patients as well as the LAD territory in controls. However, patients with obstructive CAD showed impaired LS and strain rate as well as increased PSI at peak stress. None of the MB was associated with systolic compression in invasive coronary angiography and strain parameters were similar between superficial and deep MB. Stress myocardial blood flow by positron emission tomography correlated with LS and RS at peak stress in the myocardium distal to MB (r = -0.73, p = 0.03, and r = 0.64, p = 0.04, respectively). Myocardial strain is not reduced during DSE in patients with MB in the LAD and no significant systolic compression.

Fetal Cardiovascular Magnetic Resonance Feature Tracking Myocardial Strain Analysis in Congenital Heart Disease

BackgroundCardiovascular magnetic resonance (CMR) is an emerging imaging modality for assessing anatomy and function of the fetal heart in congenital heart disease (CHD). This study aimed to evaluate myocardial strain using fetal CMR feature tracking (FT) in different subtypes of CHD. MethodsFetal CMR FT analysis was retrospectively performed on four-chamber cine images acquired with Doppler US gating at 3 Tesla. Left ventricular (LV) global longitudinal strain (GLS), LV global radial strain (GRS), LV global longitudinal systolic strain rate (SR), and right ventricular (RV) GLS were quantified using a dedicated software optimized for fetal strain analysis. Analysis was performed in normal fetuses and different CHD subtypes (d-Transposition of the great arteries (dTGA), hypoplastic left heart syndrome (HLHS), coarctation of the aorta (CoA), tetralogy of Fallot (TOF), RV-dominant atrioventricular septal defect (AVSD), and critical pulmonary stenosis or atresia (PS/PA)). Analyses of variance (ANOVA) with Tukey post-hoc test was used for group comparisons. ResultsA total of 60 fetuses were analyzed (8/60 (13%) without CHD, 52/60 (87%) with CHD). Myocardial strain was successfully assessed in 113/120 ventricles (94%). Compared to controls, LV GLS was significantly reduced in fetuses with HLHS (-18.6±2.7% vs. -6.2±5.6%; p<0.001) and RV-dominant AVSD (-18.6±2.7% vs. -7.7±5.0%; p=0.003) and higher in fetuses with CoA (-18.6±2.7% vs. -25.0±4.3%; p=0.038). LV GRS was significantly reduced in fetuses with HLHS (25.7±7.5% vs. 11.4±9.7%; p=0.024). Compared to controls, RV GRS was significantly reduced in fetuses with PS/PA (-16.1±2.8% vs. -8.3±4.2%; p=0.007). Across all strain parameters, no significant differences were present between controls and fetuses diagnosed with dTGA and TOF. ConclusionsFetal myocardial strain assessment with CMR FT in CHD is feasible. Distinct differences are present between various types of CHD, suggesting potential implications for clinical decision-making and prognostication in fetal CHD.

Barley disease recognition using deep neural networks

Plant disease negatively impacts food production and quality. It is crucial to detect and recognise plant diseases correctly. Traditional approaches do not offer a rapid and comprehensive management system for detecting plant diseases. Deep learning techniques (DL) have achieved encouraging results in discriminating patterns and anomalies in visual samples. This ability provides an effective method to diagnose any plant disease symptoms automatically. However, one of the limitations of recent studies is that in-field disease detection is underexplored, so developing a model that performs well for in-field samples is necessary. The objective of this study is to develop and investigate DL techniques for in-field disease detection of barley (Hordeum vulgare L.), one of the main crops in Australia, given visual samples captured at barley trials using a consumer-grade RGB camera. Consequently, A dataset was captured from test-bed trials across multiple paddocks infected with three diseases: net form net blotch (NFNB), spot form net blotch (SFNB), and scald, in various weather conditions. The collected data, 312 images (6000 × 4000 pixels), are divided into patches of 448 × 448 pixels, which are manually annotated into four classes: no-disease, scald, NFNB and SFNB. Finally, the data was augmented using random rotation and flip to increase the dataset size. The generated barley disease dataset is then applied to several well-known pre-trained DL networks such as DenseNet, ResNet, InceptionV3, Xception, and MobileNet as the network backbone. Given limited data, these methods can be trained to detect anomalies in visual samples. The results show that MobileNet, Xception, and InceptionV3 performed well in barley disease detection. On the other hand, ResNet showed poor classification ability. Moreover, Augmenting the data improves the performance of DL networks, particularly for underperforming backbones like ResNet, and mitigates the limited data access for these data-intensive networks. The augmentation step improved MobileNet performance by approximately 6 %. MobileNet achieved the highest accuracy of 98.63 % (the average of the three diseases) in binary classification and an accuracy of 93.50 % in multi-class classification. Even though classifying SFNB and NFNB is challenging in the early stages, MobileNet achieved the minimum misclassification rate among the two diseases. The results show the efficiency of this model in diagnosing barley diseases using complex data collected from the field environment. In addition, the model is lighter and comprises fewer trainable parameters. Consequently, MobileNet is suitable for small training datasets, reducing data acquisition costs.

An hetero-modal deep learning framework for medical image synthesis applied to contrast and non-contrast MRI

Some pathologies such as cancer and dementia require multiple imaging modalities to fully diagnose and assess the extent of the disease. Magnetic resonance imaging offers this kind of polyvalence, but examinations take time and can require contrast agent injection. The flexible synthesis of these imaging sequences based on the available ones for a given patient could help reduce scan times or circumvent the need for contrast agent injection. In this work, we propose a deep learning architecture that can perform the synthesis of all missing imaging sequences from any subset of available images. The network is trained adversarially, with the generator consisting of parallel 3D U-Net encoders and decoders that optimally combines their multi-resolution representations with a fusion operation learned by an attention network trained conjointly with the generator network. We compare our synthesis performance with 3D networks using other types of fusion and a comparable number of trainable parameters, such as the mean/variance fusion. In all synthesis scenarios except one, the synthesis performance of the network using attention-guided fusion was better than the other fusion schemes. We also inspect the encoded representations and the attention network outputs to gain insights into the synthesis process, and uncover desirable behaviors such as prioritization of specific modalities, flexible construction of the representation when important modalities are missing, and modalities being selected in regions where they carry sequence-specific information. This work suggests that a better construction of the latent representation space in hetero-modal networks can be achieved by using an attention network.

Distinguishing word identity and sequence context in DNA language models

Transformer-based large language models (LLMs) are very suited for biological sequence data, because of analogies to natural language. Complex relationships can be learned, because a concept of "words" can be generated through tokenization. Training the models with masked token prediction, they learn both token sequence identity and larger sequence context. We developed methodology to interrogate model learning, which is both relevant for the interpretability of the model and to evaluate its potential for specific tasks. We used DNABERT, a DNA language model trained on the human genome with overlapping k-mers as tokens. To gain insight into the model′s learning, we interrogated how the model performs predictions, extracted token embeddings, and defined a fine-tuning benchmarking task to predict the next tokens of different sizes without overlaps. This task evaluates foundation models without interrogating specific genome biology, it does not depend on tokenization strategies, vocabulary size, the dictionary, or the number of training parameters. Lastly, there is no leakage of information from token identity into the prediction task, which makes it particularly useful to evaluate the learning of sequence context. We discovered that the model with overlapping k-mers struggles to learn larger sequence context. Instead, the learned embeddings largely represent token sequence. Still, good performance is achieved for genome-biology-inspired fine-tuning tasks. Models with overlapping tokens may be used for tasks where a larger sequence context is of less relevance, but the token sequence directly represents the desired learning features. This emphasizes the need to interrogate knowledge representation in biological LLMs.

Deep learning method for finding eigenpairs in Sturm-Liouville eigenvalue problems

Solving the eigenvalue problem for differential equations in inhomogeneous media poses a significant challenge across diverse scientific fields. While classical finite difference methods and finite element methods have produced numerous outcomes, they heavily rely on discretizing the computational domain, which can introduce complexities and limitations. In this study, we present an unsupervised neural network approach tailored for finding eigenpairs in Sturm-Liouville eigenvalue problems within inhomogeneous media. Our method introduces eigenvalues as trainable parameters, crafts a novel cost function, incorporates an adaptive hyper-parameter tuning strategy, and sequentially trains the eigenpairs. The simplicity, accuracy, and interpretability of our approach significantly expand its applicability across various domains. The method we present in this paper can easily tackle boundary value conditions with derivatives, resulting in orthogonal eigenfunctions. This is a very important advantage of deep learning methods that has not yet been noticed. Quantitative estimation of eigenpairs is given for the Sturm-Liouville eigenvalue problems. Furthermore, we extend the proposed methodology to tackle two-dimensional cases, periodic scenarios, demonstrating its versatility and broad potential. For more information see https://ejde.math.txstate.edu/Volumes/2024/53/abstr.html

CLESSR-VC: Contrastive learning enhanced self-supervised representations for one-shot voice conversion

One-shot voice conversion (VC) has attracted more and more attention due to its broad prospects for practical application. In this task, the representation ability of speech features and the model’s generalization are the focus of attention. This paper proposes a model called CLESSR-VC, which enhances pre-trained self-supervised learning (SSL) representations through contrastive learning for one-shot VC. First, SSL features from the 23rd and 9th layers of the pre-trained WavLM are adopted to extract content embedding and SSL speaker embedding, respectively, to ensure the model’s generalization. Then, the conventional acoustic feature mel-spectrograms and contrastive learning are introduced to enhance the representation ability of speech features. Specifically, contrastive learning combined with the pitch-shift augmentation method is applied to disentangle content information from SSL features accurately. Mel-spectrograms are adopted to extract mel speaker embedding. The AM-Softmax and cross-architecture contrastive learning are applied between SSL and mel speaker embeddings to obtain the fused speaker embedding that helps improve speech quality and speaker similarity. Both objective and subjective evaluation results on the VCTK corpus confirm that the proposed VC model has outstanding performance and few trainable parameters.

Abnormal left atrial strain and left atrial stiffness index are associated with adverse outcomes in children with cardiomyopathies: a pilot study

Conventional diastolic dysfunction parameters seem to be imperfect when applied to the pediatric cardiomyopathy population. The aim of this pilot study was to search for novel echocardiographic parameters associated with adverse outcomes in children with the most common cardiomyopathies. Fifty-six patients with pediatric cardiomyopathies (28 with dilated, 21 with hypertrophic, 7 with left ventricular non-compaction cardiomyopathy) and 28 healthy subjects were included in the study. Left atrial reservoir (LASr), conduit (LAScd) and contraction (LASct) strain, left atrial stiffness index (LASI), as well as conventional diastolic dysfunction parameters were measured using echocardiography. Adverse outcomes were defined as heart failure (including heart transplant) and arrhythmic endpoints. Patients with adverse outcomes presented with significantly lower LASr (16.68% ± 8.64% vs. 33.97% ± 9.99%, p-value < 0.001), lower LAScd (− 10.37% ± 5.83% vs. − 25.50% ± 9.24%, p-value < 0.001) and higher values of LASI (0.69 [IQR 0.34; 1.11] vs. 0.21 [IQR 0.16; 0.31], p-value < 0.001). LASr < 20%, LAScd ≥ − 12%, and LASI ≥ 0.26 were all associated with reduced survival. LASr, LAScd and LASI seem to be promising parameters in predicting adverse outcomes in the most common pediatric cardiomyopathies. Left atrial strain parameters and LASI are helpful in differentiating healthy control subjects from children with hypertrophic and dilated cardiomyopathies.

Hand Surgery Training Structure in Germany - Evaluation of an Online Questionnaire

The duration of the training for hand surgery in Germany was reduced from three to two years in 2020, while other European Countries introduced hand surgery as a separate surgical specialty. Therefore, the structure of hand surgery training in Germany is under intense, ongoing discussion. This study aimed to evaluate the current situation of hand surgery training in Germany. Two separate online questionnaires were developed for self-assessment of residents and teachers in German hand surgery services. The questionnaire contained 29 questions for residents and 24 questions for teachers. Demographic data, educational tools, surgical training parameters and satisfaction were assessed from May to July 2023. The survey was based on the former training time of three years. The online questionnaire was sent to the members and associate members of the German Society for Hand Surgery via email. 148 participants answered the questionnaire, including 54 residents and 94 teachers. Overall, more female residents were assessed, whereas the majority of teachers was male. Hand surgery training is mostly conducted in high-level hospitals (tertiary care p=0.012, quaternary care p<0.001). Residents use digital teaching media significantly more often than teachers do (p<0.001). In contrast, teachers use journals (p=0.006) and clinical visits (p=0.024) significantly more often for their training than residents do. Trainers and trainees mostly agreed about time points of surgical milestones during surgical training that should be performed independently. Both trainers and trainees were largely satisfied with the practical and theoretical training in hand surgery. The importance of digital media in hand surgery training will increase in the near future. Therefore, digital media should be more frequently included in hand surgery training protocols. The study revealed that the majority of residents as well as teachers would appreciate the introduction of an independent hand surgery specialty, which would most likely improve the hand surgery competence in Germany.

Long-term reprojection loss for self-supervised monocular depth estimation in endoscopic surgery

Aim: Depth information plays a key role in enhanced perception and interaction in image-guided surgery. However, it is difficult to obtain depth information with monocular endoscopic surgery due to a lack of reliable cues for perceiving depth. Although there are reprojection loss-based self-supervised learning techniques to estimate depth and pose, the temporal information from the adjacent frames is not efficiently utilized to handle occlusion in surgery. Methods: We design long-term reprojection loss (LT-RL) self-supervised monocular depth estimation techniques by integrating longer temporal sequences into reprojection to learn better perception and to address occlusion artifacts in image-guided laparoscopic and robotic surgery. For this purpose, we exploit four temporally adjacent source frames before and after the target frame, where conventional reprojection loss uses two adjacent frames. The pixels that are visible in the target frame but occluded in the immediate two adjacent frames will produce the inaccurate depth but a higher chance to appear in the four adjacent frames during the calculation of minimum reprojection loss. Results: We validate LT-RL on the benchmark surgical datasets of Stereo correspondence and reconstruction of endoscopic data (SCARED) and Hamlyn to compare the performance with other state-of-the-art depth estimation methods. The experimental results show that our proposed technique yields 2%-4% better root-mean-squared error (RMSE) over the baselines of vanilla reprojection loss. Conclusion: Our LT-RL self-supervised depth and pose estimation technique is a simple yet effective method to tackle occlusion artifacts in monocular surgical video. It does not add any training parameters, making it flexible for integration with any network architecture and improving the performance significantly.

Meta-Attention Deep Learning for Smart Development of Metasurface Sensors.

Optical metasurfaces with pronounced spectral characteristics are promising for sensor applications. Currently, deep learning (DL) offers a rapid manner to design various metasurfaces. However, conventional DL models are usually assumed as black boxes, which is difficult to explain how a DL model learns physical features, and they usually predict optical responses of metasurfaces in a fuzzy way. This makes them incapable of capturing critical spectral features precisely, such as high quality (Q) resonances, and hinders their use in designing metasurface sensors. Here, a transformer-based explainable DL model named Metaformer for the high-intelligence design, which adopts a spectrum-splitting scheme to elevate 99% prediction accuracy through reducing 99% training parameters, is established. Based on the Metaformer, all-dielectric metasurfaces based on quasi-bound states in the continuum (Q-BIC)for high-performance metasensing are designed, and fabrication experiments are guided potently. The explainable learning relies on spectral position encoding and multi-head attention of meta-optics features, which overwhelms traditional black-box models dramatically. The meta-attention mechanism provides deep physics insights on metasurface sensors, and will inspire more powerful DL design applications on other optical devices.