Accurate and rapid prediction of structural responses to seismic loading is critical for ensuring structural safety. Recently, there has been active research focusing on the application of deep learning techniques, including Physics-Informed Neural Networks (PINNs) and Long Short-Term Memory (LSTM) networks, to predict the dynamic behavior of structures. While these methods have shown promise, each comes with distinct limitations. PINNs offer physical consistency but struggle with capturing long-term temporal dependencies in nonlinear systems, while LSTMs excel in learning sequential data but lack physical interpretability. To address these complementary limitations, this study proposes a hybrid LSTM-PINN model, combining the temporal learning ability of LSTMs with the physics-based constraints of PINNs. This hybrid approach allows the model to capture both nonlinear, time-dependent behaviors and maintain physical consistency. The proposed model is evaluated on both single-degree-of-freedom (SDOF) and multi-degree-of-freedom (MDOF) structural systems subjected to the El-Centro ground motion. For validation, the 1940 El-Centro NS earthquake record was used, and the ground acceleration data were normalized and discretized for numerical simulation. The proposed LSTM-PINN is trained under the same conditions as the conventional PINN models (e.g., same optimizer, learning rate, and loss structure), but with fewer training epochs, to evaluate learning efficiency. Prediction accuracy is quantitatively assessed using mean error and mean squared error (MSE) for displacement, velocity, and acceleration, and results are compared with PINN-only models (PINN-1, PINN-2). The results show that LSTM-PINN consistently achieves the most stable and precise predictions across the entire time domain. Notably, it outperforms the baseline PINNs even with fewer training epochs. Specifically, it achieved up to 50% lower MSE with only 10,000 epochs, compared to the PINN's 50,000 epochs, demonstrating improved generalization through temporal sequence learning. This study empirically validates the potential of physics-guided time-series AI models for dynamic structural response prediction. The proposed approach is expected to contribute to future applications such as real-time response estimation, structural health monitoring, and seismic performance evaluation.

Shark skin exhibits a well-defined multilayered architecture, consisting of three-dimensional denticles and an underlying dermal layer, which contributes to its passive drag reduction. However, the active drag reduction mechanisms of this interface remain largely unexplored. In this study, the Marangoni effect potentially arising from the active secretion of mucus on shark skin is investigated. A 3D-printed swimming robot with a porous substrate and a biomimetic shark denticle structure is developed. By introducing surfactants into the porous substrate and adjusting denticle arrangements, on-demand propulsion and controlled swimming trajectories are achieved. A superhydrophobic surface is fabricated on the swimming robot, which reduces water resistance and enhances propulsion. Moreover, denticles with a 30° attack angle demonstrate optimal propulsion performance in both Marangoni-driven hydrodynamics and aerodynamics. This study suggests that the secretion of mucus on shark skin may facilitate active drag reduction via the Marangoni effect, offering novel insights into the biomimetic structural design of autonomous swimming robots.

Biomimetic lattice structures, inspired by natural architectures such as bone, coral, mollusk shells, and Euplectella aspergillum, have gained increasing attention for their exceptional strength-to-weight ratios, energy absorption, and deformation control. These properties make them ideal for advanced engineering applications in aerospace, biomedical devices, and structural impact protection. This study presents a comprehensive bibliometric analysis of global research on biomimetic lattice structures published between 2020 and 2025, aiming to identify thematic trends, collaboration patterns, and underexplored areas. A curated dataset of 3685 publications was extracted from databases like PubMed, Dimensions, Scopus, IEEE, Google Scholar, and Science Direct and merged together. After the removal of duplication and cleaning, about 2226 full research articles selected for the bibliometric analysis excluding review works, conference papers, book chapters, and notes using Cite space, VOS viewer version 1.6.20, and Bibliometrix R packages (4.5. 64-bit) for mapping co-authorship networks, institutional affiliations, keyword co-occurrence, and citation relationships. A significant increase in the number of publications was found over the past year, reflecting growing interest in this area. The results identify China as the most prolific contributor, with substantial institutional support and active collaboration networks, especially with European research groups. Key research focuses include additive manufacturing, finite element modeling, machine learning-based design optimization, and the performance evaluation of bioinspired geometries. Notably, the integration of artificial intelligence into structural modeling is accelerating a shift toward data-driven design frameworks. However, gaps remain in geometric modeling standardization, fatigue behavior analysis, and the real-world validation of lattice structures under complex loading conditions. This study provides a strategic overview of current research directions and offers guidance for future interdisciplinary exploration. The insights are intended to support researchers and practitioners in advancing next-generation biomimetic materials with superior mechanical performance and application-specific adaptability.

The early detection of Alzheimer's disease (AD) in cognitively healthy individuals remains a major preclinical challenge. EEG is a promising tool that has shown effectiveness in detecting AD risk. Task-related EEG has been rarely used in Alzheimer's disease research, as most studies have focused on resting-state EEG. An interpretable deep learning framework-Interpretable Convolutional Neural Network (InterpretableCNN)-was utilized to identify AD-related EEG features. EEG data were recorded during three cognitive task conditions, and samples were labeled based on APOE genotype and polygenic risk scores. A 100-fold leave-p%-subjects-out cross-validation (LPSO-CV) was used to evaluate model performance and generalizability. The model achieved an ROC AUC of 60.84% across the tasks and subjects, with a Kappa value of 0.22, indicating fair agreement. Interpretation revealed a consistent focus on theta and alpha activity in the parietal and temporal regions-areas commonly associated with AD pathology. Task-related EEG combined with interpretable deep learning can reveal early AD risk signatures in healthy individuals. InterpretableCNN enhances transparency in feature identification, offering a valuable tool for preclinical screening.

Biomimetic design for engineering applications may suggest the optimal performance of engineering devices. In this work the passive/pure pitching characteristics of a hydrofoil are investigated experimentally with and without a pair of biomimetic fin strips placed symmetrically on the two sides of the foil leading edge. The work is performed in a recirculating water channel at low Reynolds numbers (Re) with a range of 1300 ≤ Re ≤ 3200. Using high-speed videography and Particle Image Velocimetry (PIV), the pitching characteristics and wakes are visualized. Passive pitching characteristics, i.e., the pitching amplitude and pitching frequency of the hydrofoils, are investigated based on their trailing edge movement. Significant improvement in both pitching frequency and amplitudes are observed for the foil with fin strips compared to the baseline simple foil. Comparing the pitching characteristics of the two foils, it is observed that the hydrofoil with biomimetic fin strips exhibits 25% and 21% higher pitching amplitude and pitching frequency, respectively, compared to that of the baseline at comparable Reynolds numbers. The initiation of pitching for the finned foil is also observed at comparatively low Reynolds numbers. The wake is also studied using time mean and fluctuating velocity profiles obtained using PIV.

Biomimetic robotics aims to replicate biological movement, perception, and cognition, drawing inspiration from nature to develop robots with enhanced adaptability, flexibility, and intelligence. The integration of artificial intelligence has significantly advanced the control mechanisms of biomimetic robots, enabling real-time learning, optimization, and adaptive decision-making. This review systematically examines AI-driven control strategies for biomimetic robots, categorizing recent advancements and methodologies. First, we review key aspects of biomimetic robotics, including locomotion, sensory perception, and cognitive learning inspired by biological systems. Next, we explore various AI techniques-such as machine learning, deep learning, and reinforcement learning-that enhance biomimetic robot control. Furthermore, we analyze existing AI-based control methods applied to different types of biomimetic robots, highlighting their effectiveness, algorithmic approaches, and performance compared to traditional control techniques. By synthesizing the latest research, this review provides a comprehensive overview of AI-driven biomimetic robot control and identifies key challenges and future research directions. Our findings offer valuable insights into the evolving role of AI in enhancing biomimetic robotics, paving the way for more intelligent, adaptive, and efficient robotic systems.

Nature's principles offer design references for additive manufacturing (AM), enabling structures that achieve remarkable efficiency through hierarchical organization rather than material excess. This perspective article proposes a framework for integrating biomimetic principles into AM beyond morphological mimicry, focusing on functional adaptation and sustainability. By emulating biological systems like nacre, spider silk, and bone, AM utilizes traditional geometric replication to embed multifunctionality, responsiveness, and resource efficiency. Recent advances in the fields of 4D printing, soft robotics, and self-morphing systems demonstrate how time-dependent behaviors and environmental adaptability can be engineered through bioinspired material architectures. However, challenges in scalable fabrication, dynamic material programming, and true functional emulation (beyond morphological mimicry) necessitate interdisciplinary collaboration. In this context, the synthesis of biological intelligence with AM technologies offers sustainable, high-performance solutions for aerospace, biomedical, and smart infrastructure applications, once challenges related to material innovation and standardization are overcome.

Hydrogels are ECM-mimicking three-dimensional (3D) networks that are widely used in biomedical applications; however, conventional natural and synthetic polymer-based hydrogels present limitations such as poor mechanical strength, limited bioactivity, and low reproducibility. Self-assembling peptides (SAPs) offer a promising alternative, as they can form micro- and nanostructured hydrogels through non-covalent interactions and allow precise control over their biofunctionality, mechanical properties, and responsiveness to biological cues. Through rational sequence design, SAPs can be engineered to exhibit tunable mechanical properties, controlled degradation rates, and multifunctionality, and can dynamically regulate assembly and degradation in response to specific stimuli such as pH, ionic strength, enzymatic cleavage, or temperature. Furthermore, SAPs have been successfully incorporated into conventional hydrogels to enhance cell adhesion, promote matrix remodeling, and provide a more physiologically relevant microenvironment. In this review, we summarize recent advances in SAP-based hydrogels, particularly focusing on their novel biofunctional properties such as anti-inflammatory, antimicrobial, and anticancer activities, as well as bioimaging capabilities, and discuss the mechanisms by which SAP hydrogels function in biological systems.

The liver is a vital organ responsible for a broad range of metabolic functions, including glucose and lipid metabolism, detoxification, and protein synthesis. Its structural complexity, characterized by hexagonal hepatic lobules composed of diverse parenchymal and non-parenchymal cell types, supports its broad spectrum of physiological activities. Traditional in vitro liver models have contributed significantly to our understanding of hepatic biology and the development of therapies for liver-related diseases. However, static culture systems fail to replicate the dynamic in vivo microenvironment, particularly the continuous blood flow and shear stress that are critical for maintaining hepatocyte function and metabolic zonation. Recent advances in microphysiological systems (MPS) incorporating dynamic fluid flow have addressed these limitations by providing more physiologically relevant platforms for modeling liver function. These systems offer improved fidelity for applications in drug screening, toxicity testing, and disease modeling. Furthermore, the integration of liver MPS with other organ models in multi-organ-on-chip platforms has enabled the investigation of inter-organ crosstalk, enhancing the translational potential of in vitro systems. This review summarizes recent progress in the development of dynamic liver MPS, highlights their biomedical applications, and discusses future directions for creating more comprehensive and predictive in vitro models.

In this paper, a biologically congruent parallel mechanism (PM) inspired by the masticatory system of human beings has been proposed to recreate complete chewing behaviours in three-dimensional space. The mechanism is featured by direct constraints from the base (DCFB) to its end effector at two higher kinematic pairs (HKPs), which greatly raise its topological complexity. Meanwhile, friction effects occur at HKPs and actuators, causing wear and then reducing motion accuracy. Regarding these, an inverse dynamic model that can raise the computational efficiency and the modelling fidelity is proposed, being prepared to be applied to realise accurate real-time motion and/or force control. In it, Euler parameters are employed to express the motions of the constrained end effector, and Newton-Euler's law is applied, which can conveniently incorporate friction effects at both HKPs and actuators into the dynamic model. Numerical results show that the time consumption of the model using Euler parameters is only approximately 23% of that of the model using Euler angles, and friction effects significantly increase the model's nonlinearity. Further, from the comparison between the models of the target PM and its counterpart free of DCFB, these constraints sharply raise the modelling complexity in terms of the transformation between Euler parameters and Euler angles in the end effector and the computational cost of inverse dynamics.