Search Results (932)

Spiking Neural Networks (SNNs), with their event-driven and energy-efficient characteristics, have shown great promise in processing data from neuromorphic sensors. However, the sparse and non-stationary nature of event-based data poses significant challenges to optimization, particularly when using conventional algorithms such as AdamW, which assume smooth gradient dynamics. To address this limitation, we propose MemristiveAdamW, a novel algorithm that integrates memristor-inspired dynamic adjustment mechanisms into the AdamW framework. This optimization algorithm introduces three biologically motivated modules: (1) a direction-aware modulation mechanism that adapts the update direction based on gradient change trends; (2) a memristive perturbation model that encodes history-sensitive adjustment inspired by the physical characteristics of memristors; and (3) a memory decay strategy that ensures stable convergence by attenuating perturbations over time. Extensive experiments are conducted on two representative event-based datasets, Prophesee NCARS and GEN1, across three SNN architectures: Spiking VGG-11, Spiking MobileNet-64, and Spiking DenseNet-121. Results demonstrate that MemristiveAdamW consistently improves convergence speed, classification accuracy, and training stability compared to standard AdamW, with the most significant gains observed in shallow or lightweight SNNs. These findings suggest that memristor-inspired optimization offers a biologically plausible and computationally effective paradigm for training SNNs on event-driven data. Full article

The growing demand for energy efficiency and sustainability in smart buildings necessitates advanced AI-driven methods for adaptive control and predictive maintenance. This study explores the application of Spiking Neural Networks (SNNs) to event-driven processing, real-time anomaly detection, and edge computing-based optimization in building automation. In contrast to conventional deep learning models, SNNs provide low-power, high-efficiency computation by mimicking biological neural processes, making them particularly suitable for real-time, edge-deployed decision-making. The proposed SNN based on Reward-Modulated Spike-Timing-Dependent Plasticity (STDP) and Bayesian Optimization (BO) integrates occupancy and ambient condition monitoring to dynamically manage assets such as appliances while simultaneously identifying anomalies for predictive maintenance. Experimental evaluations show that our BO-STDP-SNN framework achieves notable reductions in both energy consumption by 27.8% and power requirements by 70%, while delivering superior accuracy in anomaly detection compared with CNN, RNN, and LSTM based baselines. These results demonstrate the potential of SNNs to enhance the efficiency and resilience of smart building systems, reduce operational costs, and support long-term sustainability through low-latency, event-driven intelligence. Full article

Background/Objectives: Synaptic plasticity is fundamental to learning and memory, yet classical models such as Hebbian learning and spike-timing-dependent plasticity often overlook the distributed and wave-like nature of neural activity. We present a computational framework grounded in Scale Relativity Theory (SRT), which describes neural propagation along fractal geodesics in a non-differentiable space-time. The objective is to link nonlinear wave dynamics with the emergence of structured memory representations in a biologically plausible manner. Methods: Neural activity was modeled using nonlinear Schrödinger-type equations derived from SRT, yielding complex wave solutions. Synaptic plasticity was coupled through a reaction–diffusion rule driven by local activity intensity. Simulations were performed in one- and two-dimensional domains using finite difference schemes. Analyses included spectral entropy, cross-correlation, and Fourier methods to evaluate the organization and complexity of the resulting synaptic fields. Results: The model reproduced core neurobiological features: localized potentiation resembling CA1 place fields, periodic plasticity akin to entorhinal grid cells, and modular tiling patterns consistent with V1 orientation maps. Interacting waveforms generated interference-dependent plasticity, modeling memory competition and contextual modulation. The system displayed robustness to noise, gradual potentiation with saturation, and hysteresis under reversal, reflecting empirical learning and reconsolidation dynamics. Cross-frequency coupling of theta and gamma inputs further enriched trace complexity, yielding multi-scale memory structures. Conclusions: Wave-driven dynamics in fractal space-time provide a hypothesis-generating framework for distributed memory formation. The current approach is theoretical and simulation-based, relying on a simplified plasticity rule that omits neuromodulatory and glial influences. While encouraging in its ability to reproduce biological motifs, the framework remains preliminary; future work must benchmark against established models such as STDP and attractor networks and propose empirical tests to validate or falsify its predictions. Full article

This study provides an in-depth analysis of the learning dynamics of multichannel optical neurons based on spatial solitons generated in lithium niobate crystals. Single-node and multi-node configurations with different topological complexities (3 × 3, 4 × 4, and 5 × 5) were compared, assessing how the number of channels, geometry, and optical parameters affect the speed and efficiency of learning. The simulations indicate that single-node neurons achieve the desired imbalance more rapidly and with lower energy expenditure, whereas multi-node structures require higher intensities and longer timescales, yet yield a greater variety of responses, more accurately reproducing the functional diversity of biological neural tissues. The results highlight how the plasticity of these devices can be entirely modulated through optical parameters, paving the way for fully optical photonic neuromorphic networks in which memory and computation are co-localized, with potential applications in on-chip learning, adaptive routing, and distributed decision-making. Full article

The current study presents a multi-class, image-based classification of eight morphologically similar macroscopic Earthstar fungal species (Astraeus hygrometricus, Geastrum coronatum, G. elegans, G. fimbriatum, G. quadrifidum, G. rufescens, G. triplex, and Myriostoma coliforme) using deep learning and explainable artificial intelligence (XAI) techniques. For the first time in the literature, these species are evaluated together, providing a highly challenging dataset due to significant visual overlap. Eight different convolutional neural network (CNN) and transformer-based architectures were employed, including EfficientNetV2-M, DenseNet121, MaxViT-S, DeiT, RegNetY-8GF, MobileNetV3, EfficientNet-B3, and MnasNet. The accuracy scores of these models ranged from 86.16% to 96.23%, with EfficientNet-B3 achieving the best individual performance. To enhance interpretability, Grad-CAM and Score-CAM methods were utilised to visualise the rationale behind each classification decision. A key novelty of this study is the design of two hybrid ensemble models: EfficientNet-B3 + DeiT and DenseNet121 + MaxViT-S. These ensembles further improved classification stability, reaching 93.71% and 93.08% accuracy, respectively. Based on metric-based evaluation, the EfficientNet-B3 + DeiT model delivered the most balanced performance, with 93.83% precision, 93.72% recall, 93.73% F1-score, 99.10% specificity, a log loss of 0.2292, and an MCC of 0.9282. Moreover, this modeling approach holds potential for monitoring symbiotic fungal species in agricultural ecosystems and supporting sustainable production strategies. This research contributes to the literature by introducing a novel framework that simultaneously emphasises classification accuracy and model interpretability in fungal taxonomy. The proposed method successfully classified morphologically similar puffball species with high accuracy, while explainable AI techniques revealed biologically meaningful insights. All evaluation metrics were computed exclusively on a 10% independent test set that was entirely separate from the training and validation phases. Future work will focus on expanding the dataset with samples from diverse ecological regions and testing the method under field conditions. Full article

While backpropagation (BP) has long served as the cornerstone of training deep neural networks, it relies heavily on strict differentiation logic and global gradient information, lacking biological plausibility. In this paper, we systematically present a novel neural network training paradigm that depends solely on signal propagation, which we term Backward Signal Propagation (BSP). The core idea of this framework is to reinterpret network training as a symmetry-driven process of discovering inverse causal relationships. Starting from symmetry principles, we define symmetric differential equations and leverage their inherent properties to implement a learning mechanism analogous to differentiation. Furthermore, we introduce the concept of causal distance, a core invariant that bridges the forward propagation and inverse learning processes. It quantifies the influence strength between any two elements in the network, leading to a generalized form of the chain rule. With these innovations, we achieve precise, pointwise adjustment of model parameters. Unlike traditional BP, the BSP method enables parameter updates based solely on local signal features. This work offers a new direction toward efficient and biologically plausible learning algorithms. Full article

Recent advances in plant electrophysiology and machine learning suggest that bioelectric signals in plants may encode environmentally relevant information beyond physiological processes. In this study, we present a novel framework to analyse waveforms from real-time bioelectrical potentials recorded in vascular plants. Using a multi-channel electrophysiological monitoring system, we acquired continuous data from Vitis vinifera samples in a vineyard plantation under natural conditions. Plants were in different health conditions: healthy; under the infection of Flavescence dorée; plants in recovery from the same disease; and dead stumps. These signals were used as input features for an ensemble of complex machine learning models, including recurrent neural networks, trained to infer short-term meteorological parameters such as temperature and humidity. The models demonstrated predictive capabilities, with accuracy comparable to sensor-based benchmarks between one and two degree Celsius for temperature, particularly in forecasting rapid weather transitions. Feature importance analysis revealed plant-specific electrophysiological patterns that correlated with ambient conditions, suggesting the existence of biological pre-processing mechanisms sensitive to microclimatic fluctuations. This bioinspired approach opens new directions for developing plant-integrated environmental intelligence systems, offering passive and biologically rooted strategies for ultra-local forecasting—especially valuable in remote, sensor-sparse, or climate-sensitive regions. Our findings contribute to the emerging field of plant-based sensing and biomimetic environmental monitoring, expanding the role of flora to biosensors, useful in Earth system observation tasks. Full article

Background/Objectives: The evolution of gene expression responses is a critical component of population adaptation to variable environments. Predicting how DNA sequence influences expression is challenging because the genotype-to-phenotype map is not well resolved for cis-regulatory elements, transcription factor binding, regulatory interactions, and epigenetic features, not to mention how these factors respond to the environment. Methods: We tested if flexible machine learning models could learn some of the underlying cis-regulatory genotype-to-phenotype map to predict expression response to a specific environment. We tested this approach using cold-responsive transcriptome profiles in five Arabidopsis thaliana natural accessions. Results: We first tested for evidence that cis regulation plays a role in environmental response, finding 14 and 15 motifs that were significantly enriched within the up- and downstream regions of cold-responsive differentially regulated genes (DEGs). We next applied convolutional neural networks (CNNs), which learn de novo cis-regulatory motifs in DNA sequences to predict expression response to cold. We found that CNNs predicted differential expression with moderate accuracy, with evidence that predictions were hindered by the biological complexity of regulation and the large potential regulatory code. Conclusions: Overall, approaches for predicting DEGs between specific environments based only on proximate DNA sequences require further development. It may be necessary to incorporate additional biological information into models to generate accurate predictions that will be useful to population biologists. Full article

Maternal behavior encompasses a range of biologically driven responses whose expression and duration vary across species. Maternal responses rely on robust adaptive changes in the female brain, enabling mothers to engage in caregiving, nourishing, and offspring protection. Morphological and functional changes in the maternal brain enhance sensitivity to offspring cues, eliciting maternal behaviors, rewarding responses, and social processing stimuli essential for parenting. Maternal behavior comprises a range of biological responses that extend beyond basic actions, reflecting a complex, evolutionarily shaped neurobiological adaptation. These behaviors can be broadly categorized into direct behaviors, which are explicitly aimed at the care of the offspring, and indirect behaviors that, overall, ensure the protection, nourishment, and survival of the newborn. The secretion of main neuropeptide hormones, such as oxytocin (OT), prolactin (PRL), and placental lactogens (PLs), during the peripartum period, is relevant for inducing and regulating maternal responses to offspring cues, including suckling behavior. Although PRL is primarily associated with reproductive and parental functions in vertebrates, it also modulates distinct neural functions during pregnancy that extend from lactogenesis to adult neurogenesis, neuroprotection, and neuroplasticity, all of which contribute to preparing the maternal brain for motherhood and parenting interactions. Parvocellular OT-containing neurons in the paraventricular nucleus (PVN) and in the anterior hypothalamic nucleus (AHN) project axon collaterals to the medial preoptic area, which, in turn, projects to the nucleus accumbens (NACC) and lateral habenula (lHb) via the retrorubral field (RRF) and the ventral tegmental area (VTA), which mediate the motivational aspects of maternal responses to offspring cues. The reshaping process of the brain and neural networks implicated in motherhood depends on several factors, such as up- and downregulation of neuronal gene expression of bioactive peptide hormones (i.e., OT, PRL, TIP-39, galanin, spexin, pituitary adenylate cyclase-activating polypeptide (PACAP), corticotropin-releasing hormone (CRH), peptide receptors, and transcription factors (i.e., c-fos and pSTAT)) in target neurons in hypothalamic nuclei, mesolimbic areas, the hippocampus, and the brainstem, which, overall, regulate the expression of maternal behavior to offspring cues, as shown in postpartum female rodents. In this review, we describe the modulatory neuropeptides, the neural networks underlying peptide transmission systems, and cell signaling involved in parenthood. We highlight the dysregulation of neuropeptide hormones and their receptors in the central nervous system in relation to psychiatric disorders. Full article

Heat stress has a major impact on dairy cow health and productivity, especially during early lactation. Conventional heat stress monitoring methods frequently rely on single indicators, such as the temperature–humidity index (THI), which may miss subtle physiological and metabolic responses. This study presents a novel threshold-based classification framework that integrates biologically meaningful combinations of environmental, behavioural, and physiological variables to detect early-stage heat stress responses in dairy cows. Six composite heat stress conditions (C1–C6) were developed using real-time THI, milk temperature, reticulorumen pH, rumination time, milk lactose, and milk fat-to-protein ratio. The study applied and assessed five supervised machine learning models (Partial Least Squares Discriminant Analysis (PLS-DA), Support Vector Machine (SVM), Random Forest (RF0, Neural Network (NN), and an Ensemble approach) trained on daily datasets gathered from early-lactation dairy cows fitted with intraruminal boluses and monitored through milking parlour sensor systems. The dataset comprised approximately 36,000 matched records from 200 cows monitored over 60 days. The highest classification performance was observed for RF and NN models, particularly under C1 (THI > 73 and milk temperature > 38.6 °C) and C6 (THI > 74 and milk temperature > 38.7 °C), with AUC values exceeding 0.90. SHAP analysis revealed that milk temperature, THI, rumination time, and milk lactose were the most informative features across conditions. This integrative approach enhances precision livestock monitoring by enabling individualised heat stress risk classification well before clinical or production-level consequences emerge. Full article

Introduction: Stroke remains a major cause of disability worldwide, and precise identification of stroke lesions is essential for prognosis and rehabilitation planning. Machine learning has emerged as a powerful tool for automating stroke lesion segmentation and outcome prediction; however, these tasks are often studied in isolation. The two strategies are inherently interdependent since segmentation provides lesion-based features that directly inform prediction models. Methods: This narrative review synthesises studies published between 2010 and 2024 on the application of machine learning in stroke lesion segmentation and recovery forecasting. A total of 23 relevant studies were reviewed, including 10 focused on lesion segmentation and 13 on recovery prediction. Results: Convolutional Neural Networks (CNNs), including architectures such as U-Net, have improved segmentation accuracy on the Anatomical Tracings of Lesions After Stroke (ATLAS) V2 dataset; however, dataset bias and inconsistent evaluation metrics limit comparability. Integrating imaging-derived lesion characteristics with clinical features improves predictive accuracy at a higher level. Furthermore, semi-supervised and self-supervised methods enhanced performance where annotated datasets are scarce. Discussion: The review highlights the interdependence between segmentation and outcome prediction. Reliable segmentation provides biologically meaningful features that underpin recovery forecasting, while prediction tasks validate the clinical relevance of segmentation outputs. This bidirectional relationship underlines the need for unified pipelines integrating lesion segmentation with outcome prediction. Future research can improve generalisability and foster clinically robust models by advancing semi-supervised and self-supervised learning, bridging the gap between automated image analysis and patient-centred prognosis. Conclusion: Accurate lesion segmentation and outcome prediction should be viewed not as separate goals but as mutually reinforcing components of a single pipeline. Progress in segmentation strengthens recovery forecasting, while predictive modelling emphasises the clinical importance of segmentation outputs. This interdependence provides a pathway for developing more effective, generalisable, and relevant AI-driven stroke care tools. Full article

Optical biosensors play a crucial role in the field of biological detection by converting biological signals into optical signals for detection. Among them, Surface Plasmon Resonance (SPR) optical biosensors have become a research hotspot in this field due to their significant advantage of high sensitivity. Weyl Semimetals (WSMs), as a type of three-dimensional topological material with unique electronic structures and other properties, exhibit potential applications in the field of SPR sensing. Against this background, we designed a terahertz (THz) high-sensitivity SPR phase biosensor with a KR structure based on WSMs. When applied in gas sensing scenarios, the phase detection sensitivity of this sensor can reach 22,402$°$/RIU, showing a significant improvement compared to traditional SPR biosensors. Moreover, we found that the Weyl node separation distance and twist angle of WSMs have obvious effects on sensitivity regulation. Additionally, we optimized the sensitivity and structural parameters of this structure using a neural network-based deep learning algorithm. We expect that this proposed scheme can provide a feasible reference for the field of biological sensing. Full article

The Eastern Mediterranean Sea has become a hotspot for biological invasions, with thermophilic species like Lagocephalus sceleratus (silver-cheeked toadfish) and Pterois miles (devil firefish) posing significant ecological and socioeconomic threats. Machine learning models (support vector machine and neural network) were developed to predict species establishment, demonstrating high predictive accuracy. SHapley Additive exPlanations analyses further highlighted the relative influence of environmental predictors. Nominal logistic regression identified bottom temperature and salinity as the key environmental drivers for the establishment of these species, with thresholds of 16.38 °C and 39.14 psu for P. miles and 15.84 °C and 39.09 psu for L. sceleratus. Forecasts through 2035, generated using the Prophet model, have predicted warming bottom temperatures but declining salinity levels, creating variable conditions for invasion. Long-term suitability was assessed by comparing forecasted conditions against thresholds, revealing that salinity and chlorophyll a consistently fall below suitable levels for both species. L. sceleratus showed stable suitability with occasional declines, while P. miles exhibited greater variability. These findings underscore the importance of fine-scale benthic data and integrated modeling approaches for early detection and adaptive management of invasive species in Mediterranean ecosystems. The study provides clear thresholds to guide ongoing environmental monitoring and emphasizes the need for continuous assessments to anticipate future invasion risks under changing climatic conditions. Full article

The widespread application of organoids in disease modeling and drug development is significantly constrained by challenges in automated quantitative analysis. In bright-field microscopy images, organoids exhibit complex characteristics, including irregular morphology, blurred boundaries, and substantial scale variations, largely stemming from their dynamic self-organization that mimics in vivo tissue development. Existing convolutional neural network-based methods are limited by fixed receptive fields and insufficient modeling of inter-channel relationships, making them inadequate for detecting such evolving biological structures. To address these challenges, we propose a novel detection head, termed Interactive Dual-Perception Head (IDP-Head), inspired by hierarchical perception mechanisms in the biological visual cortex. Integrated into the RTMDet framework, IDP-Head comprises two bio-inspired components: a Large-Kernel Global Perception Module (LGPM) to capture global morphological dependencies, analogous to the wide receptive fields of cortical neurons, and a Progressive Channel Synergy Module (PCSM) that models inter-channel semantic collaboration, echoing the integrative processing of multi-channel stimuli in neural systems. Additionally, we construct a new organoid detection dataset to mitigate the scarcity of annotated data. Extensive experiments on both our dataset and public benchmarks demonstrate that IDP-Head achieves a 5-percentage-point improvement in mean Average Precision (mAP) over the baseline model, offering a biologically inspired and effective solution for high-fidelity organoid detection. Full article

Reliable prediction of chemical–protein interactions (CPIs) remains a key challenge in drug discovery, especially under sparse or noisy biological data. We present MM-TCoCPIn, a Multi-Modal Topology-aware Chemical–Protein Interaction Network that integrates three causally grounded modalities—network topology, biomedical semantics, and a 3D protein structure—into an interpretable graph learning framework. The model processes topological features via a CTC (Comprehensive Topological Characteristics)-based encoder, literature-derived semantics via SciBERT (Scientific Bidirectional Encoder Representations from Transformers), and structural geometry via a GVP-GNN (Geometric Vector Perceptron Graph Neural Network) applied to AlphaFold2 contact graphs. Evaluation on datasets from STITCH, STRING, and PubMed shows that MM-TCoCPIn achieves state-of-the-art performance (AUC = 0.93, F1 = 0.92), outperforming uni-modal baselines. Importantly, ablation and counterfactual analyses confirm that each modality contributes distinct biological insight: topology ensures robustness, semantics enhance recall, and structure sharpens precision. This framework offers a scalable and causally interpretable solution for CPI modeling, bridging the gap between predictive accuracy and mechanistic understanding. Full article