gRNAde: A Geometric Deep Learning Pipeline for 3D RNA Inverse Design.

Fundamental to the diverse biological functions of RNA are its 3D structure and conformational flexibility, which enable single sequences to adopt a variety of distinct 3D states. Currently, computational RNA design tasks are often posed as inverse problems, where sequences are designed based on adopting a single desired secondary structure without considering 3D geometry and conformational diversity. In this tutorial, we present gRNAde, a geometric RNA design pipeline operating on sets of 3D RNA backbone structures to design sequences that explicitly account for RNA 3D structure and dynamics. gRNAde is a graph neural network that uses an SE (3) equivariant encoder-decoder framework for generating RNA sequences conditioned on backbone structures where the identities of the bases are unknown. We demonstrate the utility of gRNAde for fixed-backbone re-design of existing RNA structures of interest from the PDB, including riboswitches, aptamers, and ribozymes. gRNAde is more accurate in terms of native sequence recovery while being significantly faster compared to existing physics-based tools for 3D RNA inverse design, such as Rosetta.

Smartphone video imaging: A versatile, low-cost technology for food authentication.

This study presents a low-cost smartphone-based imaging technique called smartphone video imaging (SVI) to capture short videos of samples that are illuminated by a colour-changing screen. Assisted by artificial intelligence, the study develops new capabilities to make SVI a versatile imaging technique such as the hyperspectral imaging (HSI). SVI enables classification of samples with heterogeneous contents, spatial representation of analyte contents and reconstruction of hyperspectral images from videos. When integrated with a residual neural network, SVI outperforms traditional computer vision methods for ginseng classification. Moreover, the technique effectively maps the spatial distribution of saffron purity in powder mixtures with predictive performance that is comparable to that of HSI. In addition, SVI combined with the U-Net deep learning module can produce high-quality images that closely resemble the target images acquired by HSI. These results suggest that SVI can serve as a consumer-oriented solution for food authentication.

Intra-Individual Reproducibility of Automated Abdominal Organ Segmentation-Performance of TotalSegmentator Compared to Human Readers and an Independent nnU-Net Model.

The purpose of this study is to assess segmentation reproducibility of artificial intelligence-based algorithm, TotalSegmentator, across 34 anatomical structures using multiphasic abdominal CT scans comparing unenhanced, arterial, and portal venous phases in the same patients. A total of 1252 multiphasic abdominal CT scans acquired at our institution between January 1, 2012, and December 31, 2022, were retrospectively included. TotalSegmentator was used to derive volumetric measurements of 34 abdominal organs and structures from the total of 3756 CT series. Reproducibility was evaluated across three contrast phases per CT and compared to two human readers and an independent nnU-Net trained on the BTCV dataset. Relative deviation in segmented volumes and absolute volume deviations (AVD) were reported. Volume deviation within 5% was considered reproducible. Thus, non-inferiority testing was conducted using a 5% margin. Twenty-nine out of 34 structures had volume deviations within 5% and were considered reproducible. Volume deviations for the adrenal glands, gallbladder, spleen, and duodenum were above 5%. Highest reproducibility was observed for bones (- 0.58% [95% CI: - 0.58, - 0.57]) and muscles (- 0.33% [- 0.35, - 0.32]). Among abdominal organs, volume deviation was 1.67% (1.60, 1.74). TotalSegmentator outperformed the reproducibility of the nnU-Net trained on the BTCV dataset with an AVD of 6.50% (6.41, 6.59) vs. 10.03% (9.86, 10.20; p < 0.0001), most notably in cases with pathologic findings. Similarly, TotalSegmentator's AVD between different contrast phases was superior compared to the interreader AVD for the same contrast phase (p = 0.036). TotalSegmentator demonstrated high intra-individual reproducibility for most abdominal structures in multiphasic abdominal CT scans. Although reproducibility was lower in pathologic cases, it outperforms both human readers and a nnU-Net trained on the BTCV dataset.

Physics-informed neural networks for biopharmaceutical cultivation processes: Consideration of varying process parameter settings.

We present a new modeling approach for the study and prediction of important process outcomes of biotechnological cultivation processes under the influence of process parameter variations. Our model is based on physics-informed neural networks (PINNs) in combination with kinetic growth equations. Using Taylor series, multivariate external process parameter variations for important variables such as temperature, seeding cell density and feeding rates can be integrated into the corresponding kinetic rates and the governing growth equations. In addition to previous approaches, PINNs also allow continuous and differentiable functions as predictions for the process outcomes. Accordingly, our results show that PINNs in combination with Taylor-series expansions for kinetic growth equations provide a very high prediction accuracy for important process variables such as cell densities and concentrations as well as a detailed study of individual and combined parameter influences. Furthermore, the proposed approach can also be used to evaluate the outcomes of new parameter variations and combinations, which enables a saving of experiments in combination with a model-driven optimization study of the design space.

Prediction of directional solidification in freeze casting of biomaterial scaffolds using physics-informed neural networks.

Freeze casting, a manufacturing technique widely applied in biomedical fields for fabricating biomaterial scaffolds, poses challenges for predicting directional solidification due to its highly nonlinear behavior and complex interplay of process parameters. Conventional numerical methods, such as computational fluid dynamics (CFD), require adequate and accurate boundary condition knowledge, limiting their utility in real-world transient solidification applications due to technical limitations. In this study, we address this challenge by developing a physics-informed neural networks (PINNs) model to predict directional solidification in freeze-casting processes. The PINNs model integrates physical constraints with neural network predictions, requiring significantly fewer predetermined boundary conditions compared to CFD. Through a comparison with CFD simulations, the PINNs model demonstrates comparable accuracy in predicting temperature distribution and solidification patterns. This promising model achieves such a performance with only 5000 data points in space and time, equivalent to 250,000 timesteps, showcasing its ability to predict solidification dynamics with high accuracy. The study's major contributions lie in providing insights into solidification patterns during freeze-casting scaffold fabrication, facilitating the design of biomaterial scaffolds with finely tuned microstructures essential for various tissue engineering applications. Furthermore, the reduced computational demands of the PINNs model offer potential cost and time savings in scaffold fabrication, promising advancements in biomedical engineering research and development.

A Lesion-aware Edge-based Graph Neural Network for Predicting Language Ability in Patients with Post-stroke Aphasia.

We propose a lesion-aware graph neural network (LEGNet) to predict language ability from resting-state fMRI (rs-fMRI) connectivity in patients with post-stroke aphasia. Our model integrates three components: an edge-based learning module that encodes functional connectivity between brain regions, a lesion encoding module, and a subgraph learning module that leverages functional similarities for prediction. We use synthetic data derived from the Human Connectome Project (HCP) for hyperparameter tuning and model pretraining. We then evaluate the performance using repeated 10-fold cross-validation on an in-house neuroimaging dataset of post-stroke aphasia. Our results demonstrate that LEGNet outperforms baseline deep learning methods in predicting language ability. LEGNet also exhibits superior generalization ability when tested on a second in-house dataset that was acquired under a slightly different neuroimaging protocol. Taken together, the results of this study highlight the potential of LEGNet in effectively learning the relationships between rs-fMRI connectivity and language ability in a patient cohort with brain lesions for improved post-stroke aphasia evaluation.

The Application of Mask Region-Based Convolutional Neural Networks in the Detection of Nasal Septal Deviation Using Cone Beam Computed Tomography Images: Proof-of-Concept Study.

BACKGROUND: Artificial intelligence (AI) models are being increasingly studied for the detection of variations and pathologies in different imaging modalities. Nasal septal deviation (NSD) is an important anatomical structure with clinical implications. However, AI-based radiographic detection of NSD has not yet been studied. OBJECTIVE: This research aimed to develop and evaluate a real-time model that can detect probable NSD using cone beam computed tomography (CBCT) images. METHODS: Coronal section images were obtained from 204 full-volume CBCT scans. The scans were classified as normal and deviated by 2 maxillofacial radiologists. The images were then used to train and test the AI model. Mask region-based convolutional neural networks (Mask R-CNNs) comprising 3 different backbones-ResNet50, ResNet101, and MobileNet-were used to detect deviated nasal septum in 204 CBCT images. To further improve the detection, an image preprocessing technique (contrast enhancement [CEH]) was added. RESULTS: The best-performing model-CEH-ResNet101-achieved a mean average precision of 0.911, with an area under the curve of 0.921. CONCLUSIONS: The performance of the model shows that the model is capable of detecting nasal septal deviation. Future research in this field should focus on additional preprocessing of images and detection of NSD based on multiple planes using 3D images.

When and why does motor preparation arise in recurrent neural network models of motor control?

During delayed ballistic reaches, motor areas consistently display movement-specific activity patterns prior to movement onset. It is unclear why these patterns arise: while they have been proposed to seed an initial neural state from which the movement unfolds, recent experiments have uncovered the presence and necessity of ongoing inputs during movement, which may lessen the need for careful initialization. Here, we modeled the motor cortex as an input-driven dynamical system, and we asked what the optimal way to control this system to perform fast delayed reaches is. We find that delay-period inputs consistently arise in an optimally controlled model of M1. By studying a variety of network architectures, we could dissect and predict the situations in which it is beneficial for a network to prepare. Finally, we show that optimal input-driven control of neural dynamics gives rise to multiple phases of preparation during reach sequences, providing a novel explanation for experimentally observed features of monkey M1 activity in double reaching.

Revolution of Medical Review: The Application of Meta-Analysis and Convolutional Neural Network-Natural Language Processing in Classifying the Literature for Head and Neck Cancer Radiotherapy.

This study explored the application of meta-analysis and convolutional neural network-natural language processing (CNN-NLP) technologies in classifying literature concerning radiotherapy for head and neck cancer. It aims to enhance both the efficiency and accuracy of literature reviews. By integrating statistical analysis with deep learning, this research successfully identified key studies related to the probability of normal tissue complications (NTCP) from a vast corpus of literature. This demonstrates the advantages of these technologies in recognizing professional terminology and extracting relevant information. The findings not only improve the quality of literature reviews but also offer new insights for future research on optimizing medical studies through AI technologies. Despite the challenges related to data quality and model generalization, this work provides clear directions for future research.

This study examines how advanced technologies like meta-analysis and machine learning, specifically through Convolutional Neural Networks and Natural Language Processing (CNN-NLP), can revolutionize the way medical researchers review literature on radiotherapy for head and neck cancer. Typically, reviewing vast amounts of medical studies is time-consuming and complex. This paper showcases a method that combines statistical analysis and AI to streamline the process, enhancing the accuracy and efficiency of identifying crucial research. By applying these technologies, the researchers were able to sift through thousands of articles rapidly, pinpointing the most relevant ones without the extensive manual effort usually required. This approach not only speeds up the review process but also improves the quality of the information extracted, making it easier for medical professionals to keep up with the latest findings and apply them effectively in clinical settings. The findings of this study are promising, demonstrating that integrating AI with traditional review methods can significantly aid in managing the ever-growing body of medical literature, potentially leading to better treatment strategies and outcomes for patients suffering from head and neck cancer. Despite some challenges like data quality and the need for extensive computational resources, the study provides a forward path for using AI to enhance medical research and practice.

Message-Passing Monte Carlo: Generating low-discrepancy point sets via graph neural networks.

Discrepancy is a well-known measure for the irregularity of the distribution of a point set. Point sets with small discrepancy are called low discrepancy and are known to efficiently fill the space in a uniform manner. Low-discrepancy points play a central role in many problems in science and engineering, including numerical integration, computer vision, machine perception, computer graphics, machine learning, and simulation. In this work, we present a machine learning approach to generate a new class of low-discrepancy point sets named Message-Passing Monte Carlo (MPMC) points. Motivated by the geometric nature of generating low-discrepancy point sets, we leverage tools from Geometric Deep Learning and base our model on graph neural networks. We further provide an extension of our framework to higher dimensions, which flexibly allows the generation of custom-made points that emphasize the uniformity in specific dimensions that are primarily important for the particular problem at hand. Finally, we demonstrate that our proposed model achieves state-of-the-art performance superior to previous methods by a significant margin. In fact, MPMC points are empirically shown to be either optimal or near-optimal with respect to the discrepancy for low dimension and small number of points, i.e., for which the optimal discrepancy can be determined.

Encoding innate ability through a genomic bottleneck.

Animals are born with extensive innate behavioral capabilities, which arise from neural circuits encoded in the genome. However, the information capacity of the genome is orders of magnitude smaller than that needed to specify the connectivity of an arbitrary brain circuit, indicating that the rules encoding circuit formation must fit through a "genomic bottleneck" as they pass from one generation to the next. Here, we formulate the problem of innate behavioral capacity in the context of artificial neural networks in terms of lossy compression of the weight matrix. We find that several standard network architectures can be compressed by several orders of magnitude, yielding pretraining performance that can approach that of the fully trained network. Interestingly, for complex but not for simple test problems, the genomic bottleneck algorithm also captures essential features of the circuit, leading to enhanced transfer learning to novel tasks and datasets. Our results suggest that compressing a neural circuit through the genomic bottleneck serves as a regularizer, enabling evolution to select simple circuits that can be readily adapted to important real-world tasks. The genomic bottleneck also suggests how innate priors can complement conventional approaches to learning in designing algorithms for AI.

Simulation and time series analysis of responsive active Brownian particles (rABPs) with memory.

To realise the goals of active matter at the micro- and nano-scale, the next generation of microrobots must be capable of autonomously sensing and responding to their environment to carry out pre-programmed tasks. Memory effects are proposed to have a significant effect on the dynamics of responsive robotic systems, drawing parallels to strategies used in nature across all length-scales. Inspired by the integral feedback control mechanism by which Escherichia coli (E. coli) are proposed to sense their environment, we develop a numerical model for responsive active Brownian particles (rABP) in which the rABPs continuously react to changes in the physical parameters dictated by their local environment. The resulting time series, extracted from their dynamic diffusion coefficients, velocity or from their fluctuating position with time, are then used to classify and characterise their response, leading to the identification of conditional heteroscedasticity in their physics. We then train recurrent neural networks (RNNs) capable of quantitatively describing the responsiveness of rABPs using their 2D trajectories. We believe that our proposed strategy to determine the parameters governing the dynamics of rABPs can be applied to guide the design of microrobots with physical intelligence encoded during their fabrication.

Hippocampal-occipital connectivity reflects autobiographical memory deficits in aphantasia.

Aphantasia refers to reduced or absent visual imagery. While most of us can readily recall decade-old personal experiences (autobiographical memories, AM) with vivid mental images, there is a dearth of information about whether the loss of visual imagery in aphantasics affects their AM retrieval. The hippocampus is thought to be a crucial hub in a brain-wide network underlying AM. One important question is whether this network, especially the connectivity of the hippocampus, is altered in aphantasia. In the current study, we tested 14 congenital aphantasics and 16 demographically matched controls in an AM fMRI task to investigate how key brain regions (i.e. hippocampus and visual-perceptual cortices) interact with each other during AM re-experiencing. All participants were interviewed regarding their autobiographical memory to examine their episodic and semantic recall of specific events. Aphantasics reported more difficulties in recalling AM, were less confident about their memories, and described less internal and emotional details than controls. Neurally, aphantasics displayed decreased hippocampal and increased visual-perceptual cortex activation during AM retrieval compared to controls. In addition, controls showed strong negative functional connectivity between the hippocampus and the visual cortex during AM and resting-state functional connectivity between these two brain structures predicted better visualization skills. Our results indicate that visual mental imagery plays an important role in detail-rich vivid AM, and that this type of cognitive function is supported by the functional connection between the hippocampus and the visual-perceptual cortex.

Enhancing heart disease diagnosis through ECG image vectorization-based classification.

Heart disease is a major issue, and the severity of its effects can be reduced through early detection and prevention. ECG is an effective diagnostic tool. Automating ECG analysis increases the possibility of timely analysis and prediction of heart conditions, improving patient outcomes. The extraction of heart-related features further enhances the accuracy of ECG-based classification models, paving the way for more effective and efficient online detection and prevention of heart diseases. The image-vectorization technique suggested in this study produces a vector representation that precisely captures the distinctive features of the heart signal. It involves image cropping, erasing the ECG grid lines, and assigning pixels to distinguish the heart signal from the background. Compared to the feature vector produced by VGG16, the extracted feature vector is 589 times shorter than the feature vector produced by VGG16, which significantly decreased the amount of memory required, increased algorithm convergence, and required less computing power. The feature vector extracted using image-vectorization is used to create the training dataset, which is used to train artificial neural networks (ANNs). The results demonstrate that using image-vectorization techniques improved the performance of machine learning algorithms compared to using conventional feature extraction algorithms like CNNs and VGG16.

Artificial intelligence-powered electrochemical sensor: Recent advances, challenges, and prospects.

Integrating artificial intelligence (AI) with electrochemical biosensors is revolutionizing medical treatments by enhancing patient data collection and enabling the development of advanced wearable sensors for health, fitness, and environmental monitoring. Electrochemical biosensors, which detect biomarkers through electrochemical processes, are significantly more effective. The integration of artificial intelligence is adept at identifying, categorizing, characterizing, and projecting intricate data patterns. As the Internet of Things (IoT), big data, and big health technologies move from theory to practice, AI-powered biosensors offer significant opportunities for real-time disease detection and personalized healthcare. Still, they also pose challenges such as data privacy, sensor stability, and algorithmic bias. This paper highlights the critical advances in material innovation, biorecognition elements, signal transduction, data processing, and intelligent decision systems necessary for developing next-generation wearable and implantable devices. Despite existing limitations, the integration of AI into biosensor systems shows immense promise for creating future medical devices that can provide early detection and improved patient outcomes, marking a transformative step forward in healthcare technology.

Spectral integrated neural networks (SINNs) for solving forward and inverse dynamic problems.

This study introduces an innovative neural network framework named spectral integrated neural networks (SINNs) to address both forward and inverse dynamic problems in three-dimensional space. In the SINNs, the spectral integration technique is utilized for temporal discretization, followed by the application of a fully connected neural network to solve the resulting partial differential equations in the spatial domain. Furthermore, the polynomial basis functions are employed to expand the unknown function, with the goal of improving the performance of SINNs in tackling inverse problems. The performance of the developed framework is evaluated through several dynamic benchmark examples encompassing linear and nonlinear heat conduction problems, linear and nonlinear wave propagation problems, inverse problem of heat conduction, and long-time heat conduction problem. The numerical results demonstrate that the SINNs can effectively and accurately solve forward and inverse problems involving heat conduction and wave propagation. Additionally, the SINNs provide precise and stable solutions for dynamic problems with extended time durations. Compared to commonly used physics-informed neural networks, the SINNs exhibit superior performance with enhanced convergence speed, computational accuracy, and efficiency.

Machine Learning in the Diagnosis, Management, and Care of Patients with Low Back Pain: A Scoping Review of the Literature and Future Directions.

BACKGROUND CONTEXT: Low back pain (LBP) remains the leading cause of disability globally. In recent years, machine learning (ML) has emerged as a potentially useful tool to aid the diagnosis, management, and prognostication of LBP. PURPOSE: In this review, we assess the scope of ML applications in the LBP literature and outline gaps and opportunities. STUDY DESIGN/SETTING: A scoping review was performed in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. METHODS: Articles were extracted from the Web of Science, Scopus, PubMed, and IEEE Xplore databases. Title/abstract and full-text screening was performed by two reviewers. Data on model type, model inputs, predicted outcomes, and ML methods were collected. RESULTS: In total, 223 unique studies published between 1988-2023 were identified, with just over 50% focused on low-back-pain detection. Neural networks were used in 106 of these articles. Common inputs included patient history, demographics, and lab values (67% total). Articles published after 2010 were also likely to incorporate imaging data into their models (41.7% of articles). Of the 212 supervised learning articles identified, 168 (79.4%) mentioned use of a training or testing dataset, 116 (54.7%) utilized cross-validation, and 46 (21.7%) implemented hyperparameter optimization. Of all articles, only 8 included external validation and 9 had publicly available code. CONCLUSIONS: Despite the rapid application of ML in LBP research, a majority of articles do not follow standard ML best practices. Furthermore, over 95% of articles cannot be reproduced or authenticated due to lack of code availability. Increased collaboration and code sharing are needed to support future growth and implementation of ML in the care of patients with LBP.

Identifying defects and varieties of Malting Barley Kernels.

This study introduces a comprehensive approach for classifying individual malting barley kernels, involving dual-sided kernel imaging, a specifically designed image processing algorithm, an optimized deep neural network architecture, and a mechanical sorting system. The proposed method achieves precise classification into multiple classes, aligning with quality standards for malting material assessment. Throughout the study, various image analysis techniques were assessed, including traditional feature engineering, established transfer learning deep neural network architectures, and our custom-designed convolutional neural network tailored for barley kernel image analysis. Comparative analysis underscores the superior performance of our network model. The study reveals that our proposed deep learning network achieves a 94% accuracy in classifying barley kernel defects and varieties, outperforming well-established transfer learning models to complex architectures that attain 93% accuracy. Additionally, it surpasses the traditional machine learning approach involving feature extraction and support vector machine classifiers, which achieve accuracy below 90% in detecting defective kernels and below 70% in varietal classification. However, we also noted the traditional approach's advantage in morphological feature recognition. This observation guides new research toward integrating morphological feature extraction techniques with modern convolutional networks. This paper presents a deep neural network designed specifically for the analysis of cereal kernel images in two applications: defect and variety classification. It emphasizes the importance of standardizing kernel orientation and merging images from both sides of the kernel, and introduces a device for image acquisition that fulfills this need.

Artificial Intelligence (AI)-Enhanced Detection of Diabetic Retinopathy From Fundus Images: The Current Landscape and Future Directions.

Diabetic retinopathy (DR) remains a leading cause of vision loss worldwide, with early detection critical for preventing irreversible damage. This review explores the current landscape and future directions of artificial intelligence (AI)-enhanced detection of DR from fundus images. Recent advances in deep learning and computer vision have enabled AI systems to analyze retinal images with expert-level accuracy, potentially transforming DR screening. Key developments include convolutional neural networks achieving high sensitivity and specificity in detecting referable DR, multi-task learning approaches that can simultaneously detect and grade DR severity, and lightweight models enabling deployment on mobile devices. While these AI systems show promise in improving the efficiency and accessibility of DR screening, several challenges remain. These include ensuring generalizability across diverse populations, standardizing image acquisition and quality, addressing the "black box" nature of complex models, and integrating AI seamlessly into clinical workflows. Future directions in the field encompass explainable AI to enhance transparency, federated learning to leverage decentralized datasets, and the integration of AI with electronic health records and other diagnostic modalities. There is also growing potential for AI to contribute to personalized treatment planning and predictive analytics for disease progression. As the technology continues to evolve, maintaining a focus on rigorous clinical validation, ethical considerations, and real-world implementation will be crucial for realizing the full potential of AI-enhanced DR detection in improving global eye health outcomes.

A novel signal channel attention network for multi-modal emotion recognition.

Physiological signal recognition is crucial in emotion recognition, and recent advancements in multi-modal fusion have enabled the integration of various physiological signals for improved recognition tasks. However, current models for emotion recognition with hyper complex multi-modal signals face limitations due to fusion methods and insufficient attention mechanisms, preventing further enhancement in classification performance. To address these challenges, we propose a new model framework named Signal Channel Attention Network (SCA-Net), which comprises three main components: an encoder, an attention fusion module, and a decoder. In the attention fusion module, we developed five types of attention mechanisms inspired by existing research and performed comparative experiments using the public dataset MAHNOB-HCI. All of these experiments demonstrate the effectiveness of the attention module we addressed for our baseline model in improving both accuracy and F1 score metrics. We also conducted ablation experiments within the most effective attention fusion module to verify the benefits of multi-modal fusion. Additionally, we adjusted the training process for different attention fusion modules by employing varying early stopping parameters to prevent model overfitting.