The study, conducted by Li, Y., Li, R., Fan, Y., and their colleagues, represents a significant leap forward in materials engineering. By integrating sophisticated machine learning algorithms with inverse design principles, the team has bypassed traditional trial-and-error methods, exploring an expansive design space with remarkable efficiency and precision. This fusion of computational intelligence with bioinspired insights heralds a new era in smart materials development that could redefine how engineers and scientists approach the creation of next-generation composites.

Auxetic materials challenge the norms of mechanical response. Unlike conventional materials that thin out when stretched, auxetics expand laterally, providing enhanced energy absorption, fracture resistance, and indentation resilience. These traits make them ideal candidates for applications demanding robust yet adaptable materials, including aerospace components, biomedical implants, and wearable sensors. However, engineering composites that simultaneously optimize these properties while maintaining manufacturability has been a formidable challenge—until now.

Central to this breakthrough is the concept of inverse design, where the desired material properties guide the design process backward, enabling researchers to deduce the optimal micro- and nano-scale structural configurations to achieve specified mechanical responses. Traditionally, such inversion has been constrained by limited computational resources and the complexity of material behaviors. The introduction of machine learning has shattered these barriers, offering a scalable and nuanced predictive framework that captures the intricate, nonlinear interactions within layered composites.

The research team employed a suite of machine learning models capable of assimilating vast datasets derived from both experimental measurements and high-fidelity simulations. These models iteratively refined the composite structure parameters—such as layer thickness, orientation, and constituent material properties—to iteratively converge on configurations exhibiting peak auxetic performance. This data-driven paradigm not only accelerates the discovery process but also unveils new design principles rooted in natural, biological analogs.

Bioinspiration played a vital role, as the team drew on evolutionary-honed architectures found in natural materials like nacre, bone, and plant cell walls. By mimicking hierarchical layering and strategic interfacial bonding patterns, the researchers created composites that synergize strength, flexibility, and auxetic response. This biomimetic strategy, amplified by machine learning, enabled the generation of novel structures that outperform conventionally designed materials in critical mechanical metrics.

One of the most striking achievements of the study is the demonstration of composites with tunable auxetic behavior, wherein the degree of negative Poisson’s ratio can be precisely modulated depending on specific application needs. This versatility stems from the ability of the machine learning framework to explore multidimensional design landscapes efficiently, identifying subtle trade-offs and synergies between competing structural factors. This marks a departure from monolithic, fixed-property materials toward adaptive composites.

The implications extend beyond mechanical properties alone. The inverse design methodology facilitates the exploration of multifunctional materials capable of integrating auxetic performance with other desirable attributes, such as thermal stability, electrical conductivity, and self-healing capabilities. This holistic optimization could revolutionize sectors ranging from wearable electronics to soft robotics, where integrated performance dictates feasibility and success.

Moreover, the researchers underscore the scalability and manufacturability of their bioinspired designs. By incorporating constraints reflecting real-world fabrication techniques, the machine learning models generate practically viable structures, significantly narrowing the gap between computational innovation and industrial application. This approach addresses a perennial bottleneck in advanced materials development—translating theoretical designs into tangible products.

The study’s comprehensive dataset and open-source machine learning frameworks invite further exploration and community-driven advancements. This democratization of design tools fosters collaboration across disciplines, encouraging material scientists, engineers, and computer scientists to co-develop next-generation composites. The transparent sharing of design principles also accelerates education and innovation pipelines worldwide.

Furthermore, the adaptability of the methodology promises new frontiers in customizing material behaviors to tailor-fit diverse environmental and operational contexts. For instance, engineers can now envision composites specifically engineered for variable loading conditions in aerospace environments or personalized implants optimized for patient-specific biomechanical demands. Such precision engineering was previously unattainable due to computational and experimental constraints.

In summary, this research exemplifies the transformative power of integrating artificial intelligence with biomimetic materials science. The machine learning-enabled inverse design framework offers an unprecedented route to engineer layered composite materials with maximized auxetic performance, pushing the boundaries of what is mechanically achievable. It sets a new standard for the rational design of smart materials, promising to impact myriad industries and inspire future scientific breakthroughs.

As the research community continues to refine these techniques, the convergence of biology, materials science, and machine learning heralds a paradigm shift towards intelligent, adaptive, and multifunctional materials. The strategies unveiled by Li and colleagues not only solve longstanding challenges in composite design but also open new vistas for innovation at the nexus of digital and physical material realms.

This visionary approach aligns with emerging trends in materials informatics and digital twinning, where digital replicas of physical systems enable real-time optimization and predictive maintenance. The incorporation of machine learning in inverse design scenarios accelerates the feedback loop between design, testing, and deployment, facilitating rapid prototyping and iterative improvements.

Ultimately, the study delivers a compelling blueprint for harnessing nature-inspired structures through modern computational tools, embodying the synthesis of tradition and technology. It reflects an exciting frontier where engineering ingenuity, computational power, and biological wisdom converge to create materials that were once thought impossible.

The combination of rigorous scientific methodology, interdisciplinary collaboration, and technological innovation showcased in this research underscores not only the present capabilities but also the future potential of AI-assisted materials science. The impact on both academic research and industrial manufacturing could be profound, fostering smarter, safer, and more sustainable material solutions for the challenges of tomorrow.

Article Title: Machine learning-enabled inverse design of bioinspired layered composite structures with maximum auxetic performance

Article References:
Li, Y., Li, R., Fan, Y. et al. Machine learning-enabled inverse design of bioinspired layered composite structures with maximum auxetic performance. Commun Eng (2025). https://doi.org/10.1038/s44172-025-00557-5

Image Credits: AI Generated

Historically, achieving printed transistors with submicrometre channel lengths necessitated complex chemical processes or extensive post-processing, which inherently restricted their applicability in practical settings. This new methodology, however, appears to sidestep such convolutions, allowing for direct fabrication without the need for chemical alterations or physical interventions after printing. This not only streamlines the production process but also opens up possibilities for using the technology in diverse applications, thereby enhancing its overall versatility.

Capillary flow printing operates by exploiting the natural tendency of liquids to move through narrow spaces. In this innovative approach, researchers have successfully utilized this property to print various conducting, semiconducting, and insulating inks onto a plethora of substrate materials including silicon, Kapton, and paper. The research team demonstrated that the technology is not merely limited to a single type of substrate but can adapt to various surfaces, expanding the range of potential applications significantly.

When examining the performance of the fabricated carbon nanotube thin-film transistors, the results were nothing short of impressive. The devices exhibited on-currents of 1.12 mA mm⁻¹ when operating through a back-gated configuration on a Si/SiO₂ substrate. Furthermore, when the transistors were side-gated using an ion gel on Kapton, the on-current registered at a respectable 490 µA mm⁻¹. Such results indicate that the electrical performance of these printed transistors is quite competitive, potentially rivaling traditionally manufactured counterparts.

Moreover, a key feature of this research is the assessment of mechanical resilience. The transistors printed on Kapton substrates displayed exceptional properties in terms of mechanical bending and sweep rate resilience. This characteristic is particularly vital, as flexible electronics are increasingly becoming integral components in personal electronics and wearable technologies. The ability to maintain performance integrity under physical stress is a crucial metric for evaluating the feasibility of these printed devices in real-world applications.

The capillary flow printing approach not only makes the printed devices functionally innovative but also positions itself as a sustainable solution within the context of manufacturing. Traditional methods of fabricating transistors are often resource-intensive, requiring significant energy, materials, and time. By contrast, capillary flow printing offers a more efficient route to producing high-quality transistors, which could lead to reduced waste and a lower carbon footprint in the long run.

Furthermore, the implications of this breakthrough extend beyond individual devices and could reshape entire electronic systems. As the ability to create highly functional, submicrometre transistors that are both flexible and robust becomes more accessible, industries involved in electronics manufacturing may find themselves at a pivotal junction where cost-effective, scalable solutions become standard practice.

Analyzing the broader impact, one cannot overlook how this technology could facilitate the development of next-generation electronics. As high-performance printed transistors become a reality, we might witness a significant leap in the capabilities of flexible and portable electronic devices. This is particularly promising for applications in areas such as health monitoring, where lightweight and flexible electronics can enhance patient comfort and device functionality.

In addition to health technology, this innovation has potential implications in sectors like automotive and consumer electronics. Imagine the prospects of flexible displays that could bend and adapt to various forms without compromising their functionalities, or intelligent wearable devices embedded with these high-performance transistors, continuously monitoring and responding to user needs. The versatile applications of capillary flow printed transistors could redefine product design and consumer experiences in multiple domains.

While the research undoubtedly demonstrates exciting possibilities, the team acknowledges the continuous challenges that must be addressed for widespread adoption. Scaling the technology for mass production while maintaining the same high-performance standards will be a critical next step. Collaborative efforts between academia and industry may accelerate this process, ultimately leading to commercialization opportunities that capitalize on this innovative printing technique.

Looking ahead, it will be essential to explore further advancements in ink formulations, substrate compatibility, and printing techniques to expand the range of materials and applications. With continued research and development, we may witness an evolution within the realm of printed electronics that aligns with future demands and technological aspirations.

As the world increasingly gravitates towards miniaturization and portability in electronics, capillary flow printing emerges as a vital player. This new technique not only breaks the barriers imposed by conventional printing technologies but also paves the way for an exciting future where printed electronics can seamlessly blend into our daily lives. The forthcoming years will undoubtedly unveil more such innovations, with the potential to revolutionize how we interact with technology and redefine what is possible in the fascinating field of electronics.

In conclusion, the study conducted by Smith et al. elucidates a pivotal moment in the domain of printed electronics. The successful demonstration of capillary flow printing for fabricating submicrometre carbon nanotube transistors signals a substantial leap forward. As this promising technology continues to evolve, the obstacles of traditional manufacturing processes may soon become relics of the past, providing exciting new avenues for research, development, and ultimately, commercialization in the dynamic landscape of electronics.

Subject of Research: Capillary flow printing of submicrometre carbon nanotube thin-film transistors.

Article Title: Capillary flow printing of submicrometre carbon nanotube transistors.

Article References:

Smith, B.N., Albarghouthi, F.M., Doherty, J.L. et al. Capillary flow printing of submicrometre carbon nanotube transistors.
Nat Electron (2025). https://doi.org/10.1038/s41928-025-01470-7

Image Credits: AI Generated

DOI:

Keywords: carbon nanotubes, thin-film transistors, capillary flow printing, flexible electronics, submicrometre fabrication, printed electronics, sustainable manufacturing, high-performance devices, innovative technology, electronic systems.

The spider web has long fascinated scientists and engineers due to its uniquely isotropic mechanical properties. Its radial and spiral threads distribute forces evenly, enabling the detection of multidirectional stresses with high sensitivity. Capitalizing on these natural characteristics, the research team devised a sensor array mimicking the web’s concentric and radial fiber layout, enabling omnidirectional strain detection—a capability that traditional linear or unidirectional strain sensors lack.

Central to this innovation is the use of Ti3C2Tx, a highly conductive material belonging to the MXene family. MXenes have emerged as promising candidates for flexible electronics due to their excellent electrical conductivity, mechanical flexibility, and surface chemistry. The team formulated a conductive ink from Ti3C2Tx flakes, which was then precisely deposited onto substrates using advanced 3D printing techniques. The combination of MXene’s superior properties and additive manufacturing allowed the fabrication of delicate sensor elements arranged in a bioinspired spider web pattern.

The fabrication process leverages 3D printing’s additive nature, enabling intricate geometries with high resolution and reproducibility. This method not only replicates the spider web’s complex weave but also ensures consistent sensor performance across the array. The printed sensor lines form interconnected pathways that respond to strain-induced deformation by altering their electrical resistance, a response that is then captured and interpreted.

Detecting strain in multiple directions poses a significant challenge because traditional sensors typically measure deformation along a single axis. The isotropic nature of the spider web design means the sensor array exhibits uniform sensitivity to strain regardless of direction, providing a rich dataset from complex mechanical stimuli. However, decoding such multidimensional signals requires sophisticated data processing to distinguish between strain magnitude and orientation.

To address this, the researchers integrated a multi-class, multi-output neural network model to perform signal decoupling. The neural network was trained to analyze electrical resistance changes across the entire sensor array and to infer the precise direction and magnitude of the applied strain. This approach transforms raw sensor data into actionable information, enabling accurate and real-time monitoring of complex mechanical interactions.

Experimental evaluation of the sensor array revealed a gauge factor (GF) of 26.3 within a strain range of 0 to 10%, which signifies a high sensitivity to strain-induced resistance changes. The gauge factor is a critical benchmark for strain sensors, representing the ratio of relative change in electrical resistance to mechanical strain. A high GF indicates the sensor can detect subtle strain variations, essential for applications requiring precise mechanical feedback.

Beyond sensitivity, the sensor array demonstrated exemplary accuracy in identifying strain parameters. Under various surface stimuli, the neural network achieved approximately 97% correctness in distinguishing both the magnitude and directional components of applied strain. This high classification accuracy is pivotal for applications such as wearable health devices and intelligent robotics, where precise motion detection and feedback are necessary.

One promising application for this bioinspired sensor technology lies in human motion monitoring. The human body produces complex, multidirectional strains during everyday activities, demanding sensors capable of capturing nuanced deformations. The spider web-inspired array’s isotropic sensing and robust decoding algorithm offer a significant advantage in wearable devices, facilitating detailed analysis of joint and muscle movements.

Moreover, the multifaceted strain detection ability can revolutionize smart robotic systems. Robots often require intricate tactile sensing and force feedback for precise manipulation tasks. Integrating such sensor arrays may enhance robotic skin, enabling machines to perceive multidirectional stresses and adjust their motions accordingly for safer and more efficient operation.

The research also underscores the reliability and repeatability of the sensor array’s performance. By combining the stable electrochemical properties of MXene materials with the robust architecture of the spider web pattern, the device maintains consistent responsiveness even after multiple cycles of mechanical deformation. This durability is crucial for real-world applications where sensors must operate under repetitive and dynamic loads.

Furthermore, the compatibility of the sensor fabrication process with flexible substrates and printing technologies suggests scalability and cost-effectiveness. Utilizing 3D printing allows rapid prototyping and customization, enabling tailored sensor designs for specific applications without the need for complex lithographic processes. This flexibility accelerates the translation of laboratory innovations into market-ready products.

In summary, this study effectively bridges natural biological structures with cutting-edge materials science and machine learning to realize an omnidirectional strain sensor array with exceptional sensitivity and signal processing capabilities. The integration of MXene conductive ink, spider web-inspired design, and neural network-based data decoupling forms a powerful platform for next-generation strain sensing technologies.

As wearable health monitoring devices proliferate and robotics advance toward more intuitive human-machine interactions, such sensor systems are poised to become indispensable. Their ability to accurately discern multifaceted mechanical signals promises to enhance device responsiveness, user experience, and safety across numerous fields. This research thus charts a promising path to multifarious applications in smart textiles, prosthetics, and interactive robotic skins.

The convergence of biomimicry, novel nanomaterials, additive manufacturing, and deep learning epitomizes future directions in sensor development. By drawing inspiration from something as deceptively simple yet functionally complex as spider webs, scientists have demonstrated how multidisciplinary approaches can resolve longstanding challenges in precision sensing, underscoring the profound insights nature continues to provide.

Subject of Research: Omnidirectional strain sensor array inspired by spider web structures using MXene-based conductive ink and neural network signal processing.

Article Title: (Not provided)

News Publication Date: (Not provided)

Web References: (Not provided)

References: (Not provided)

Image Credits: (Not provided)

Keywords: Omnidirectional strain sensor, spider web structure, MXene (Ti3C2Tx), 3D printing, neural network, signal decoupling, gauge factor, wearable sensors, intelligent robotics, flexible electronics.

Hydrogels, renowned for their remarkable ability to swell and shrink based on changes in their environment, have found applications across various fields, from flexible electronics to targeted drug delivery. One particularly compelling challenge in the design of hydrogel-based devices lies in effectively managing the opening and closing of pores to trap or release substances, such as drug particles. Typically, hydrogels that employ circular pores are hampered by inconsistent actuation behaviors and slow response times, which can jeopardize their efficiency and effectiveness in critical applications.

The team’s approach introduces polygonal pores, integrated with origami-inspired hinge and facet architectures, to enable a more responsive and reliable actuation mechanism. Professor Yoon states, “Our design allows for a facet-driven folding strategy where the alignment of the hinges at the vertices dictates how the facets will move during the pore’s swelling and shrinking processes.” This control over the actuation process marks a significant leap forward, allowing researchers to predict and program the behavior of the hydrogel pores with greater accuracy.

Applications of this technology extend into the medical field, where the system demonstrates pH-triggered release of microparticles. The researchers note that by finely tuning the environment’s pH levels, they can orchestrate a staged release of microparticles from the hydrogel pores. This capability is particularly advantageous for drug delivery systems, where targeting specific areas of the body marked by variable pH can minimize systemic effects and enhance therapeutic efficacy.

The degree to which these polymeric materials can maintain their structural integrity is also impressive. Notably, the polygonal pores are reported to retain a remarkable 90% of their original shape even after repeated cycles of swelling and shrinking. This property not only underscores the reliability of the innovative design but also opens the door for various potential applications where durability is a key requirement.

Additionally, the researchers have ventured into the domain of information encryption using these hydrogels. They created a unique mixed matrix consisting of both square and circular hydrogel pores, filled with fluorescent particles. This setup capitalizes on the distinct closing behaviors of the different pore shapes, which allows for hiding and revealing patterns—an exciting frontier in the field of secure information storage and communication.

The implications of this facet-driven folding strategy are vast and varied. It not only presents improvements for current hydrogel technologies but also lays the foundation for future breakthroughs in numerous areas, including lab-on-a-chip systems and next-generation soft robotics. Such advancements could revolutionize how we think about drug delivery, environmental monitoring, and even smart textiles.

As researchers continue to explore the full potential of these innovative hydrogel structures, it becomes increasingly clear that they could usher in a new era of smart materials. These materials could seamlessly integrate into both everyday applications and sophisticated technological solutions, enhancing functionality while reducing risks associated with conventional systems.

In summary, the work carried out by Professor Yoon and his team exemplifies the intersection of creativity and scientific inquiry, showcasing how concepts from fields such as origami can be harnessed to solve complex engineering challenges. This research not only significantly enhances the precision of hydrogel actuation but also expands the horizons of what is possible with soft materials in various technological and medical applications.

As these scientists prepare to share their findings with the broader scientific community, the anticipation surrounding their work grows, with the hope that more researchers will adopt this innovative strategy to further enhance the capabilities and applications of hydrogels in the future. The trajectory suggested by their initial results indicates a bright future for facet-driven hydrogel technologies, potentially changing the landscape of soft materials science for the better.

The insights gained from this research may lead to refined methodologies and protocols for developing highly tunable hydrogel systems that can be applied in diverse industries, from healthcare to environmental remediation. Overall, the implications of this innovative approach signal a significant milestone in the development of next-generation materials, and the scientific community is eager to witness its impact.

Subject of Research: Reversible actuation of hydrogel pores
Article Title: Facet-driven folding for precise control of hydrogel pore actuation
News Publication Date: 30-Jun-2025
Web References: DOI link
References: Journal Matter
Image Credits: Seoul National University of Science and Technology (SEOULTECH)

Keywords

Hydrogels, polymer engineering, drug delivery, origami-inspired structures, soft materials, pH-responsive systems, information encryption, advanced manufacturing, biomedical applications, smart materials.

Tendons are essential connective tissues that transmit forces between muscle and bone, enabling movement and stability. Their unique structural characteristics feature a gradient of mechanical properties, transitioning from soft, compliant muscle attachments to stiff, durable bone insertions. Replicating this mechanical gradient within synthetic hydrogels has long been a challenge due to the intrinsic limitations of uniform polymer networks, which often fail to balance strength, elasticity, and biocompatibility simultaneously. Zhu and colleagues tackled this problem head-on by engineering hydrogels that regenerate this gradation through controlled physical crosslinking methods, circumventing traditional chemical crosslinkers that might compromise biological compatibility.

At the core of this innovation is a physical crosslinking technique that manipulates polymer chain interactions without introducing covalent bonds, thereby preserving reversibility, self-healing capability, and dynamic responsiveness under physiological conditions. The team employed a meticulous assembly of polymer components whose interactions are fine-tuned to generate regions with distinct mechanical stiffness. This gradient structure emulates the natural transition in tendon tissues, improving cellular integration and mechanical performance under dynamic loading situations, pivotal for functional tissue regeneration.

In practical terms, the hydrogel’s strength arises from dual physical crosslinking domains that include hydrogen bonding and hydrophobic association. These non-covalent interactions provide a balance of stability and flexibility, allowing the material to withstand substantial mechanical stress while maintaining elasticity. Zhu et al. demonstrated that the physical crosslinks serve as sacrificial bonds, dissipating energy efficiently during deformation, a property crucial for mimicking the fatigue resistance of tendons subjected to repetitive strain cycles.

Moreover, the formation of a mechanical gradient within the hydrogel matrix is orchestrated by spatially controlling the crosslinking density. This gradient not only augments the toughness and strength of the material but also guides cell migration and differentiation within the scaffold. Tendon cells, or tenocytes, are notoriously sensitive to mechanical cues, and this engineered substrate supplies a biomimetic environment that enhances their alignment and proliferation, thereby accelerating the healing process in tendon injuries.

The implications of this research extend beyond tendon repair. By advancing the methodology of physical crosslinking to create tunable mechanical gradients, the study opens avenues for designing hydrogels tailored to other complex tissues that exhibit heterogeneous structures, such as cartilage, ligaments, and even interfaces in flexible electronic systems. The dynamic nature of the physical bonds allows these hydrogels to interface seamlessly with biological systems, potentially serving as bioelectronic platforms where mechanical compliance and electrical function coexist.

From a materials science perspective, the research stands out by leveraging supramolecular chemistry principles to impart both structure and function. The reversible interactions offer self-healing properties that conventional hydrogels lack, granting longevity and durability in physiological environments. Experimental data showcased by the team confirms remarkable recovery of mechanical properties post-damage, suggesting that such hydrogels could enhance implant lifespan and reduce the frequency of surgical interventions.

The team’s methodological approach involved synthesizing a copolymer system integrated with moieties capable of hydrogen bonding and hydrophobic interactions, carefully calibrating monomer ratios to dictate mechanical outputs. Mechanical testing revealed elastic moduli spanning across the physiological range of natural tendons, alongside impressive tensile strength, surpassing many previously reported hydrogel systems. Additionally, cyclic loading experiments demonstrated outstanding resilience and negligible hysteresis, indicative of the effective energy dissipation mechanisms facilitated by physical crosslinks.

Biocompatibility assays conducted in vitro confirmed that the hydrogel environment supports tenocyte viability and does not elicit adverse inflammatory responses. Cell culture studies further revealed organized extracellular matrix deposition, a hallmark of functional tendon regeneration. These findings underscore the potential of mechanically gradient hydrogels as scaffolds that not only restore physical continuity but actively participate in tissue remodeling and repair.

Critically, the absence of permanent chemical crosslinkers reduces cytotoxic risks and simplifies fabrication, offering an accessible platform for clinical translation. The use of physical crosslinking also allows for facile tuning of mechanical gradients by modulating environmental stimuli such as temperature or ionic strength, which could enable personalized therapeutic designs matching patient-specific tissue mechanics.

This study represents a significant stride in biomaterials innovation where mechanical competence meets biological activity. The team’s results, published in the highly respected journal npj Flexible Electronics, set a new benchmark for hydrogel design targeted toward soft tissue repair, particularly for challenging anisotropic tissues like tendons. By addressing both the mechanical and biological aspects simultaneously, these hydrogels manifest the ideal scaffold that has eluded researchers for years in regenerative medicine.

Future directions envisioned by Zhu and colleagues involve integrating electrical conductivity into these mechanically gradient hydrogels, thereby coupling biomechanical and electrophysiological functionalities. This synergy is especially relevant in bioelectronic medicine, where flexible platforms capable of biomechanical support and electrical signaling could revolutionize treatments for musculoskeletal disorders and neural interfacing.

In conclusion, the development of high-strength, mechanically gradient hydrogels via physical crosslinking constitutes a versatile and impactful innovation for tissue repair, holding promise to significantly improve clinical outcomes for tendon injuries. This work offers an inspiring blueprint for the field of flexible biomaterials, illustrating how careful manipulation of polymer physics can bridge synthetic constructs with natural tissue mechanics, paving the way for next-generation medical therapies and smart implantable devices.

Subject of Research: High-strength mechanically gradient hydrogels designed via physical crosslinking for tendon-mimetic tissue repair.

Article Title: High-strength mechanically gradient hydrogels via physical crosslinking for tendon-mimetic tissue repair.

Article References:
Zhu, H., Wang, C., Yang, Y. et al. High-strength mechanically gradient hydrogels via physical crosslinking for tendon-mimetic tissue repair. npj Flex Electron 9, 53 (2025). https://doi.org/10.1038/s41528-025-00430-7

Image Credits: AI Generated