Search Results (812)

Neuromorphic hardware systems allow efficient implementation of artificial neural networks (ANNs) across various applications that demand high data throughput, reduced physical size, and low energy consumption. Field-Programmable Gate Arrays (FPGAs) possess inherent features that can be aligned with these requirements. However, implementing ANNs on FPGAs also presents challenges, including the computation of the neuron activation functions, due to the balance between resource constraints and numerical precision. This paper proposes a resource-efficient hardware approximation method for the sigmoid function, utilizing a combination of first- and second-degree polynomial functions. The method aims mainly to minimize the approximation error. This paper also evaluates the obtained results against existing techniques and discusses their significance. The experimental results showed that, although the proposed method mainly aimed to minimize the approximation error, it also had lower hardware resource usage than several of the most closely related works. Using 16-bit fixed-point number representation, the absolute mean error was $1.66\times {10}^{-3}$ by using 0.04% of the logic blocks and 3.21% of the DSP blocks in a Ciclone V 5CGXFC7C7F23C8 FPGA Device. Full article

The human respiratory Central Pattern Generator (CPG) is a complex and tightly regulated network of neurons responsible for the automatic rhythm of breathing. Among the brain nuclei involved in respiratory control, excitatory neurons within the PreBotzinger Complex (PreBötC) are both necessary and sufficient for generating this rhythmic activity. Although several models of the PreBötC circuit have been proposed, a comprehensive analysis of network behavior in response to physiologically relevant external inputs remains limited. In this study, we present a computational model of the PreBötC consisting of 1000 excitatory neurons, divided into two functional subgroups: the rhythm-generating population and the pattern-forming population. To enable real-time closed-loop simulations, we employed parallelized multi-process computing to accelerate network simulation. The network, composed of asynchronous neurons, could produce bursting activity at a eupneic breathing frequency of 0.22 Hz, which could also reproduce the rapid and stable chemoreception of breathing activated in response to hypercapnia. Additionally, it successfully replicated rapid and stable respiratory responses to elevated carbon dioxide levels (hypercapnia), mediated through simulated chemoreception. External inputs from a carbon dioxide sensor were used to modulate the network activity, allowing the implementation of a real-time respiratory control system. These results demonstrate that a network of asynchronous, non-bursting neurons can emulate the behavior of the respiratory CPG and its modulation by external stimuli. The proposed model represents a step toward developing a closed-loop controller for breathing regulation. Full article

Background: Bone marrow-derived mesenchymal stem cell exosomes (EXOs) are attractive in biotechnology and biomedical research, as they possess natural cell-targeting properties and can cross biological barriers by influencing the SDF-1/CXCR4 axis. Lipid calcium phosphate (LCP) consists of a calcium phosphate core and an asymmetric phospholipid bilayer containing abundant Ca²⁺ ions. AMD3100 modification of targeted LCP (T-LCP) can achieve targeted delivery to ischemic lesions via specific binding to CXCR4 receptors on various neuronal cell surfaces. Methods: Herein, a fused membrane formulation that simultaneously possesses EXO characteristics and enables targeted modification with AMD3100 was produced. The characteristics of biologically derived EXOs, artificially designed T-LCP, and the fused membrane formulation, including targeted delivery and gene loading efficiency, were then compared. Results: The fusion of artificially designed T-LCP with EXOs of natural origin is feasible and combines the advantages of both to achieve more prominent targeted delivery effects. Conclusions: MiRNA210-based gene therapy was effective in this study and provides a strategy for therapeutic efficacy in delivery systems with different targeting efficiencies. Full article

Low-Dimensional Nanohybrids (LDNHs) have emerged as potent multifunctional platforms for neurosensing and neuromodulation, providing elevated spatial-temporal precision, versatility, and biocompatibility. This review examines the intersection of LDNHs with artificial intelligence, brain–computer interfaces (BCIs), and closed-loop neurotechnologies, highlighting their transformative potential in personalized neuro-nano-medicine. Utilizing stimuli-responsive characteristics, optical, thermal, magnetic, and electrochemical LDNHs provide real-time feedback-controlled manipulation of brain circuits. Their pliable and adaptable structures surpass the constraints of inflexible bioelectronics, improving the neuronal interface and reducing tissue damage. We also examined their use in less invasive neurological diagnostics, targeted therapy, and adaptive intervention systems. This review delineates recent breakthroughs, integration methodologies, and fundamental mechanisms, while addressing significant challenges such as long-term biocompatibility, deep-tissue accessibility, and scalable manufacturing. A strategic plan is provided to direct future research toward clinical use. Ultimately, LDNHs signify a transformative advancement in intelligent, tailored, and closed-loop neurotechnologies, integrating materials science, neurology, and artificial intelligence to facilitate the next era of precision medicine. Full article

Traumatic brain injury (TBI) remains a major global health problem, contributing significantly to morbidity and mortality worldwide. Despite advances in understanding its complex pathophysiology, current therapeutic strategies are insufficient in addressing the long-term cognitive, emotional, and neurological impairments. While the primary mechanical injury is immediate and unavoidable, the secondary phase involves a cascade of biological processes leading to neuroinflammation, blood–brain barrier (BBB) disruption, and systemic immune activation. The heterogeneity of patient responses underscores the urgent need for reliable biomarkers and targeted interventions. Emerging evidence highlights the gut–brain axis as a critical modulator of the secondary phase, with microbiota-derived extracellular vesicles (MEVs) representing a promising avenue for both diagnosis and therapy. MEVs can cross the intestinal barrier and BBB, carrying biomolecules that influence neuronal survival, synaptic plasticity, and inflammatory signaling. These properties make MEVs promising biomarkers for early detection, severity classification, and prognosis in TBI, while also offering therapeutic potential through modulation of neuroinflammation and promotion of neural repair. MEV-based strategies could enable tailored interventions based on the individual’s microbiome profile, immune status, and injury characteristics. The integration of multi-omics with artificial intelligence is expected to fully unlock the diagnostic and therapeutic potential of MEVs. These approaches can identify molecular subtypes, predict outcomes, and facilitate real-time clinical decision-making. By bridging microbiology, neuroscience, and precision medicine, MEVs hold transformative potential to advance TBI diagnosis, monitoring, and treatment. This review also identifies key research gaps and proposes future directions for MEVs in precision diagnostics and gut microbiota-based therapeutics in neurotrauma care. Full article

Repairing the central nervous system (CNS) remains one of the most difficult obstacles to overcome in translational neurosciences. This is due to intrinsic growth inhibitors, extracellular matrix issues, the glial scar–form barrier, chronic neuroinflammation, and epigenetic silencing. The purpose of this review is to bring together findings from recent developments in genome editing and computational approaches, which center around the possible convergence of clustered regularly interspaced short palindromic repeats (CRISPR) platforms and artificial intelligence (AI), towards precision neuroregeneration. We wished to outline possible ways in which CRISPR-based systems, including but not limited to Cas9 and Cas12 nucleases, RNA-targeting Cas13, base and prime editors, and transcriptional regulators such as CRISPRa/i, can be applied to potentially reactivate axon-growth programs, alter inhibitory extracellular signaling, reprogram or lineage transform glia to functional neurons, and block oncogenic pathways in glioblastoma. In addition, we wanted to highlight how AI approaches, such as single-cell multi-omics, radiogenomic prediction, development of digital twins, and design of adaptive clinical trials, will increasingly be positioned to act as system-level architects that allow translation of complex datasets into predictive and actionable therapeutic approaches. We examine convergence consumers in spinal cord injury and adaptive neuro-oncology and discuss expanse consumers in ischemic stroke, Alzheimer’s disease, Parkinson’s disease, and rare neurogenetic syndromes. Finally, we discuss the ethical and regulatory landscape around beyond off-target editing and genomic stability of CRISPR, algorithmic bias, explainability, and equitable access to advanced neurotherapies. Our intent was not to provide a comprehensive inventory of possibilities but rather to provide a conceptual tool where CRISPR acts as a molecular manipulator and AI as a computational integrator, converging to create pathways towards precision neuroregeneration, personalized medicine, and adaptive neurotherapeutics that are ethically sound. Full article

Autonomous robotic manipulators in industrial environments face significant challenges, including time-varying payloads, multi-source disturbances, and real-time computational constraints. Traditional model predictive control frameworks degrade by over 40% under model uncertainties, while conventional adaptive techniques exhibit convergence times incompatible with industrial cycles. This work presents a hybrid adaptive model predictive control framework integrating edge artificial intelligence with dual-stage parameter estimation for 6-DoF industrial manipulators. The approach combines recursive least squares with a resource-optimized neural network (three layers, 32 neurons, <500 KB memory) designed for industrial edge deployment. The system employs innovation-based adaptive forgetting factors, providing exponential convergence with mathematically proven Lyapunov-based stability guarantees. Simulation validation using the Fanuc CR-7iA/L manipulator demonstrates superior performance across demanding scenarios, including precision laser cutting and obstacle avoidance. Results show 52% trajectory tracking RMSE reduction (0.022 m to 0.012 m) under 20% payload variations compared to standard MPC, while achieving sub-5 ms edge inference latency with 99.2% reliability. The hybrid estimator achieves 65% faster parameter convergence than classical RLS, with 18% energy efficiency improvement. Statistical significance is confirmed through ANOVA (F = 24.7, p < 0.001) with large effect sizes (Cohen’s d > 1.2). This performance surpasses recent adaptive control methods while maintaining proven stability guarantees. Hardware validation under realistic industrial conditions remains necessary to confirm practical applicability. Full article

In artificial neural networks, the activation dynamics of non-trainable variables are strongly coupled to the learning dynamics of trainable variables. During the activation pass, the boundary neurons (e.g., input neurons) are mapped to the bulk neurons (e.g., hidden neurons), and during the learning pass, both bulk and boundary neurons are mapped to changes in trainable variables (e.g., weights and biases). For example, in feedforward neural networks, forward propagation is the activation pass and backward propagation is the learning pass. We show that a composition of the two maps establishes a duality map between a subspace of non-trainable boundary variables (e.g., dataset) and a tangent subspace of trainable variables (i.e., learning). In general, the dataset-learning duality is a complex nonlinear map between high-dimensional spaces. We use duality to study the emergence of criticality, or the power-law distribution of fluctuations of the trainable variables, using a toy and large models at learning equilibrium. In particular, we show that criticality can emerge in the learning system even from the dataset in a non-critical state, and that the power-law distribution can be modified by changing either the activation function or the loss function. Full article

Chronic inflammation plays a crucial role in the pathogenesis and progression of neurodegenerative diseases such as Alzheimer’s disease (AD) and Parkinson’s disease (PD), where sustained neuroinflammatory responses contribute to neuronal damage and functional decline. Recent advances in nanomedicine offer novel therapeutic strategies aimed at modulating inflammation, with a focus on targeting the gut–brain axis, a key mediator in the interplay between systemic inflammation and neurodegeneration. Artificial intelligence (AI) has emerged as a transformative tool in this context, facilitating the integration of large, complex datasets to better understand the intricate relationship between gut microbiota dysbiosis, chronic neuroinflammation, the exposome (cumulative impact of lifelong environmental exposures), and disease manifestation. AI-driven approaches and integrating exposome data with AI enable deeper insights into exposure–microbiome–inflammation interactions, enhance our understanding of the inflammatory pathways involved, support the development of predictive models for disease progression, and optimize the delivery of nanomedicine-based therapeutics. Additionally, AI applications in neuroimaging and personalized therapy planning have shown promise in addressing both motor and non-motor symptoms. This review provides a comprehensive synthesis of current knowledge, highlighting the convergence of AI, nanomedicine, and chronic inflammation in neurodegenerative disease care. Full article

Ship stiffened panels, typically flat plates reinforced with various types of stiffeners, form a substantial part of a ship’s structure and are susceptible to resonance, especially in areas such as the after peak structure, engine room, and accommodation compartments. These vibrations are primarily excited by main engine and propeller forces. Conventional methods such as finite element analysis (FEA) and plate theory are widely used to estimate vibration frequencies, but they are time-consuming and computationally intensive when applied to numerous stiffened panels. This study proposes a machine learning approach using an artificial neural network (ANN) algorithm to efficiently predict the vibration frequencies of ship stiffened panels. A crude oil tanker is chosen as the case study, and FEA is conducted to generate the vibration frequency and mass data for panels across critical regions. The input layer features for the ANN include panel area, thickness, number and area of stiffeners, fluid density, number of fluid contact sides, and overall structural stiffness. The ANN model predicts two outputs: the fundamental vibration frequency and the mass of the panel structure. To evaluate the model performance, hyperparameters such as the number of hidden neurons are optimized. The results indicate that the ANN achieves accurate predictions while significantly reducing the time and resources required compared with conventional methods. This approach offers a promising tool for accelerating the local vibration analysis process in ship structural design. Full article

Prototype-Based Networks (PBNs) are inherently interpretable architectures that facilitate understanding of model outputs by analyzing the activation of specific neurons—referred to as prototypes—during the forward pass. The learned prototypes serve as transformations of the input space into a latent representation that more effectively encapsulates the main characteristics shared across data samples, thereby enhancing classification performance. Crucially, these prototypes can be decoded and projected back into the original input space, providing direct interpretability of the features learned by the network. While this characteristic marks a meaningful advancement toward the realization of fully interpretable artificial intelligence systems, our findings reveal that prototype representations can be deliberately or inadvertently manipulated without compromising the superficial appearance of explainability. In this study, we conduct a series of empirical investigations that demonstrate this phenomenon, framing it as a structural paradox potentially intrinsic to the architecture or its design, which may represent a significant robustness challenge for explainable AI methodologies. Full article

A neural model was developed to predict the distribution of ZnO nanoparticles obtained by electrochemical synthesis. It is a three-layer multilayer perceptron (MLP) artificial neural network (ANN) with five neurons in the input layer, eight neurons in the hidden layer, and one neuron in the output layer. This network has a hyperbolic tangent activation function for the neurons in the hidden layer and an exponential activation function for the neuron in the output layer. The input (independent) variables are particle size (nm), solvent type, and temperature (°C), and the output (dependent) variable is fraction share (%). The best neural model (ann08) has a root mean square error (RMSE) 0.84% for the training subset, 0.98% for the testing subset, and 1.27% for the validation subset. The RMSE values are therefore small, which enables practical use of the ANN model. Full article

An information-cognitive concept of a predicting method for obtaining specialized human–computer interaction (HCI) objects’ perception subjectivization results, based on impact factors analysis, with the use of multilayer perceptron (MP) artificial neural networks (ANNs), has been developed. The main purpose of the developed method is to increase the level of intellectualization and automation of research into relevant processes of HCI objects’ perception subjectivization, especially in the context of software products’ comprehensive support processes. The method is based on the developed conceptual models and the developed mathematical model, as well as a specialized developed algorithm. Results prediction is carried out on the basis of a preliminary in-depth analysis of a set of unique direct chains (UDCs) of neurons of the relevant encapsulated MP ANN, built on the basis of researching the results of isolated influences of each of the previously declared impact factors and further comparing the present direct chain (of each separate investigated modeling case) with UDCs from the aforementioned sets. As an example of practical approbation of the developed method, the appropriate practically applied problem of identifying a member of the support team, whose multifactor portrait is as close as possible to the corresponding multifactor portrait of a given client’s user, has been resolved. Full article

The precise identification, classification, sorting, and rapid and accurate quality assessment of soybean seeds are extremely important in terms of the continuity of agricultural production, varietal purity, seed processing, protein extraction, and food safety. Currently, commonly used methods for the identification and quality assessment of soybean seeds include morphological analysis, chemical analysis, protein electrophoresis, liquid chromatography, spectral analysis, and image analysis. The use of image analysis and artificial intelligence is the aim of the presented research, in which a method for the automatic classification of soybean varieties, the assessment of the degree of damage, and the identification of geometric features of soybean seeds based on numerical models obtained using a 3D scanner has been proposed. Unlike traditional two-dimensional images, which only represent height and width, 3D imaging adds a third dimension, allowing for a more realistic representation of the shape of the seeds. The research was conducted on soybean seeds with a moisture content of 13%, and the seeds were stored in a room with a temperature of 20–23 °C and air humidity of 60%. Individual soybean seeds were scanned to create 3D models, allowing for the measurement of their geometric parameters, assessment of texture, evaluation of damage, and identification of characteristic varietal features. The developed 3D-CNN network model comprised an architecture consisting of an input layer, three hidden layers, and one output layer with a single neuron. The aim of the conducted research is to design a new, three-dimensional 3D-CNN architecture, the main task of which is the classification of soybean seeds. For the purposes of network analysis and testing, 22 input criteria were defined, with a hierarchy of their importance. The training, testing, and validation database of the SB3D-NET network consisted of 3D models obtained as a result of scanning individual soybean seeds, 100 for each variety. The accuracy of the training process of the proposed SB3D-NET model for the qualitative classification of 3D models of soybean seeds, based on the adopted criteria, was 95.54%, and the accuracy of its validation was 90.74%. The relative loss value during the training process of the SB3D-NET model was 18.53%, and during its validation process, it was 37.76%. The proposed SB3D-NET neural network model for all twenty-two criteria achieves values of global error (GE) of prediction and classification of seeds at the level of 0.0992. Full article

Species of the Ferula genus are known for their traditional medicinal applications against diverse illnesses. Our previous study was the first to suggest the cholinesterase inhibitory activity of Ferula persica L. However, the neuroprotective efficacy of therapeutic molecules is often limited by their ability to cross the blood–brain barrier (BBB) and reach the brain. In the present study, the BBB permeability of the main molecules present in the aerial parts and roots of F. persica L. extracted under optimum conditions was assessed using two well-established methods: the parallel artificial membrane permeability assay (PAMPA) and the HBMEC cell culture in vitro model. The results demonstrated a high permeability of several neuroprotective compounds, such as apigenin, diosmetin, and α-cyperone. Additionally, the neuroprotective potential of F. persica extracts was evaluated using SH-SY5Y neuron-like cells exposed to different insults, including oxidative stress (H₂O₂), excitotoxicity (L-glutamate), and Aβ1-42 peptide toxicity. However, none of the obtained extracts provided significant protection. This study highlights the importance of in vitro cell culture models for a better understanding of BBB permeability mechanisms and reports the tentative identification of newly formed sulfated metabolites derived from the metabolism of ferulic acid, apigenin, and diosmetin by HBMEC cells. Full article