Explore the vast range of titles available.

With the booming growth of artificial intelligence (AI), the traditional von Neumann computing architecture based on complementary metal oxide semiconductor devices are facing memory wall and power wall. Memristor based in-memory computing can potentially overcome the current bottleneck of computer and achieve hardware breakthrough. In this review, the recent progress of memory devices in material and structure design, device performance and applications are summarized. Various resistive switching materials, including electrodes, binary oxides, perovskites, organics, and two-dimensional materials, are presented and their role in the memristor are discussed. Subsequently, the construction of shaped electrodes, the design of functional layer and other factors influencing the device performance are analyzed. We focus on the modulation of the resistances and the effective methods to enhance the performance. Furthermore, synaptic plasticity, optical-electrical properties, the fashionable applications in logic operation and analog calculation are introduced. Finally, some critical issues such as the resistive switching mechanism, multi-sensory fusion, system-level optimization are discussed.

Small molecules and conjugated polymers, the two main types of organic materials used for optoelectronic and photonic devices, can be used in a number of applications including organic light-emitting diodes, photovoltaic devices, photorefractive devices and waveguides.

Exoskeletons have enormous potential to improve human locomotive performance 1 – 3 . However, their development and broad dissemination are limited by the requirement for lengthy human tests and handcrafted control laws 2 . Here we show an experiment-free method to learn a versatile control policy in simulation. Our learning-in-simulation framework leverages dynamics-aware musculoskeletal and exoskeleton models and data-driven reinforcement learning to bridge the gap between simulation and reality without human experiments. The learned controller is deployed on a custom hip exoskeleton that automatically generates assistance across different activities with reduced metabolic rates by 24.3%, 13.1% and 15.4% for walking, running and stair climbing, respectively. Our framework may offer a generalizable and scalable strategy for the rapid development and widespread adoption of a variety of assistive robots for both able-bodied and mobility-impaired individuals. A learning-in-simulation framework for wearable robots uses dynamics-aware musculoskeletal and exoskeleton models and data-driven reinforcement learning to bridge the gap between simulation and reality without human experiments to assist versatile activities.

IntroductionDespite successful recanalization with mechanical thrombectomy (MT) for acute anterior ischemic stroke (AAIS), the number of passes may impact clinical outcome.We analyzed the impact of more than three MT passes (>3) in a trial that evaluated contact aspiration (CA) versus stent retriever (SR) as the first-line technique in AAIS.MethodsWe included patients with mTICI 2b/3 recanalization after MT for isolated intracranial occlusions. The primary outcome was the percentage of patients with a 90-day modified Rankin Scale (mRS)≤2. Secondary outcomes included overall distribution of 90-day mRS, parenchymal hematoma on 24 hours' brain imaging (PH), and 90-day mortality.ResultsAmong the 281 patients included and even after adjustment on time to recanalization, significantly more patients with >3 passes had PH than patients with ≤3 passes in multivariate analysis (adjusted OR, 3.62; 95% CI, 1.55 to 8.44). When the analyses were stratified according to CA vs. SR, patients with >3 passes had a stronger risk of PH than patients with ≤3 passes, only in the SR first-line-treated group (adjusted OR, 9.24; 95% CI, 2.65 to 32.13) and not in the CA first-line-treated group (adjusted RR, 1.73; 95% CI, 0.57 to 5.19). A negative association of borderline significance (P=0.07) between >3 passes and favorable outcome was observed only in SR first-line-treated patients (adjusted OR, 0.33; 95% CI, 0.09 to 1.11).ConclusionsAfter three passes of SR and unlike for three passes of CA, there is an increased risk of PH and a trend toward a worse clinical outcome.

Many nanophotonic devices rely on the excitation of photonic resonances to enhance light-matter interaction. The understanding of the resonances is therefore of a key importance to facilitate the design of such devices. These resonances may be analyzed by use of the quasi-normal mode (QNM) theory. Here, we illustrate how QNM analysis may help study and design resonant nanophotonic devices. We will in particular use the QNM expansion of far-field quantities based on Riesz projection to design optical antennas.

Artificial intelligence-based medical devices (AIMDs) have emerged as transformative technologies in modern health care. However, comprehensive analysis of recent approval trends and characteristics of AIMDs in China remains limited. This study aimed to provide an up-to-date overview of AIMDs approved in China up to June 2025. We conducted a search of the Drugdataexpy database to identify AIMDs approved up to June 30, 2025, using artificial intelligence-related keywords in the \"structural composition\" and \"intended use\" fields. After manual verification and exclusion of non-AIMDs, we collected key characteristics, including name, manufacturer, approval date, risk class, clinical evaluation pathway, medical specialty, data source, review pathway, and algorithm type. Statistical analysis encompassed descriptive statistics and trend analysis. We used the Fisher exact test and Pearson chi-square test to assess the associations between risk class and categorical variables. A total of 154 AIMDs were identified since the first approval in 2020, with annual approvals increasing from 9 in 2020 to 45 in 2024 (a 49.53% compound annual growth rate), although the 20 approvals in the first half of 2025 suggest a potential moderation in pace. Most AIMDs (123/154, 79.9%) were categorized as class III, and the risk class was significantly associated with approval year (P=.03), manufacturer location (P=.03), and medical specialty (P=.004). Of the 123 class III devices, 19 (15.4%) were approved through innovation review, and 2 (1.6%) each were approved through priority and emergency approval. Deep learning was the dominant algorithm (143/154, 92.9%). Radiology dominated the field (106/154, 68.8%), with computed tomography serving as the primary data source (96/154, 62.3%), particularly for applications in pulmonary nodule detection and cardiovascular assessment. Clinical trials were the primary evaluation pathway for 76.6% (118/154) of all AIMDs. This approach was predominant for class III devices (116/123, 94.3%), whereas most class II devices (21/31, 67.7%) used a clinical exemption pathway. Market concentration was evident, with the top 4 manufacturers accounting for 38.3% (59/154) of all approvals and geographically clustered in major innovation hubs such as Beijing, Shanghai, Shenzhen, and Hangzhou. China's AIMD ecosystem is experiencing growth, heavily focused on radiology and computed tomography-based solutions within a risk-proportionate regulatory framework. The market is characterized by significant manufacturer and geographic concentration.

High profile device failures have highlighted the inadequacies of current regulation. Art Sedrakyan and colleagues call for a move to a graduated model of approval and suggest a framework to achieve this goal

Objective To evaluate the safety and comfort of ureteral stents with extraction strings during modified split-leg prone percutaneous nephrolithotomy (PCNL). Methods A prospective, single-centre study was conducted on 100 patients undergoing PCNL for unilateral upper urinary tract stones from April to August 2024. Patients were randomized into two groups: 50 with extraction-string stents and 50 without. Standardized surgical and postoperative protocols were followed. Primary outcomes included pain scores during stent removal, incidence of hematuria and flank pain, stent indwelling time, and cost analysis. Statistical analysis was performed using GraphPad Prism 9.5.0. Results The extraction-string group demonstrated significantly lower pain scores during stent removal (0.86 ± 0.62 vs. 5.23 ± 1.74, p  < 0.05) and shorter stent indwelling time (16.06 ± 4.48 vs. 60.54 ± 20.4 days, p  < 0.05). The incidence of hematuria (8 vs. 29 cases) and flank pain (7 vs. 22 cases) was notably lower in the extraction-string group ( p  < 0.05). Additionally, each patient in this group saved an average of 1,065 Yuan (145.78 USD) by avoiding cystoscopic removal. No significant differences were observed in postoperative hospitalization days or urinary irritation symptoms ( p  > 0.05). Conclusion The use of extraction-string ureteral stents during modified split-leg prone PCNL significantly reduces stent removal pain, enhances procedural convenience, lowers economic burden, and maintains a safety profile comparable to traditional methods. This innovative technique represents a clinically valuable advancement in PCNL surgery.