Annals of the New York Academy of Sciences最新文献

Panic-buying incidents triggered by public emergencies are often accompanied by panic emotions, which are prone to spread in both time and space, causing serious negative impacts on social stability and economic development. Therefore, we took stimuli–organism–response (SOR) as our framework to comprehensively consider the effects of internal and external factors and to clarify the chain evolution process of stimulus panic-buying behavior. Second, based on an improved infectious disease model using the Markov chain and combined with cellular automata methods, we studied the spatiotemporal evolution of group panic-buying behavior. Specifically, we focused on the evolutionary process of panic-buying across temporal sequences and within macroscopic geographical spaces such as urban communities. Our simulations yielded compelling insights: (1) the intensity of panic-buying information has a significant stimulating effect on panic-buying behavior. Interestingly, previous encounters with such buying experiences can act as a soothing balm, alleviating panic emotions. (2) The groups that participate in and stop buying both exhibit a clustered distribution in space, forming several clustered buying areas in space. Individuals at the center of the buying area often participate in the buying first, while those at the edge are more likely to stop buying. (3) Partition control implemented in the buying area can help mitigate the cascading effects of fear-driven purchasing behavior and increase its rate of decline. In addition, we also verified the effectiveness of the model using a case study in China, and discuss the model's theoretical and practical significance, limitations, and future research directions.

In women of reproductive age, metabolic dysfunction-associated steatotic liver disease (MASLD) is associated with increased risks of gestational complications and adverse pregnancy outcomes. In addition, gestational MASLD may affect the lifelong health trajectory of the offspring. The live-streamed exercise intervention in pregnant women with metabolic dysfunction–associated steatotic liver disease (Live-MASLD) trial aims to investigate the efficacy of a live-streamed exercise intervention in ameliorating MASLD progression and improving maternal−infant health outcomes. A two-arm randomized controlled trial will be conducted across two centers. Pregnant individuals with MASLD in their first trimester will be recruited and randomly allocated to a control group or an intervention group. The control group will receive routine prenatal healthcare, while the intervention group will undergo a supervised, live-streamed exercise program during the second and third trimesters. Assessments will occur at baseline, second trimester, third trimester, delivery, and 42 days postpartum. Process evaluation will be conducted to allow for a health economic evaluation of the intervention. The primary outcome will describe between-group differences in changes to the hepatic steatosis index (LiSA) as quantitatively measured by visual transient elastography (ViTE). The findings from the Live-MASLD study will provide scientific evidence and implementation guidance for the clinical management of pregnant women with MASLD.

This systematic review synthesizes 69 original studies (2019–2025) to evaluate the transformative potential of functional near-infrared spectroscopy (fNIRS) in ADHD research. As a portable, motion-tolerant neuroimaging tool, fNIRS enables robust measurement of cortical hemodynamic activity during cognitive tasks. We first consolidate the specifications of fNIRS devices employed in ADHD studies. Next, we discuss the neural markers derived from fNIRS data—including hemodynamic response function features, functional connectivity metrics, the beta coefficients of general linear model, graph theory measures, amplitude of low-frequency fluctuations, and multiscale entropy—alongside artificial intelligence (AI) algorithms achieving high diagnostic accuracy. Critically, we demonstrate fNIRS's utility in objectively monitoring treatment response, as evidenced by prefrontal cortex normalization and posterior activation modulation following interventions. To realize personalized diagnostics and therapeutics, future research should prioritize: (1) wearable fNIRS systems for ecological monitoring, (2) multimodal AI frameworks integrating fNIRS with behavioral/genetic data, and (3) standardized protocols validated in large-scale cohorts.

Current interactive information visualization systems often fail to adequately incorporate cultural elements, limiting their effectiveness in cross-cultural communication and reducing user engagement across diverse global audiences. This research presents a comprehensive methodology that integrates artificial intelligence technologies with cultural elements to create more engaging and culturally appropriate interactive visualization systems. The proposed approach includes an intelligent cultural element identification algorithm achieving 91.2% recognition accuracy, an adaptive visualization generation framework that automatically incorporates cultural preferences, and real-time interface optimization mechanisms that dynamically adjust to user behavior patterns. Experimental validation with 600 participants from diverse cultural backgrounds demonstrated significant improvements: 23.5% increase in user satisfaction, 18.3% reduction in cognitive load, and 31.7% enhancement in user engagement compared to traditional visualization approaches. The multidimensional evaluation revealed superior performance across aesthetic appeal, functional usability, and cultural appropriateness metrics. This research contributes theoretical frameworks for understanding cultural influences on visualization perception and provides practical guidelines for implementing culturally informed, AI-driven design systems in multicultural environments.

Dynamic obstacle avoidance remains a key challenge in robotic arm motion planning, as traditional algorithms struggle to balance adaptive decision-making with precise trajectory generation in unstructured environments. We present a hierarchical motion planning framework that combines proximal policy optimization (PPO) with rapidly exploring random tree star (RRT*), trained using a curriculum learning paradigm. PPO learns global obstacle avoidance strategies through progressively difficult training scenarios, while RRT* refines local trajectories to compensate for PPO's limitations in fine motor control. A multiobjective reward function—incorporating step-efficiency terms and artificial potential field principles—balances exploration and exploitation through tailored penalties and rewards. In dynamic obstacle scenarios, the proposed method achieves an 87.6% success rate, outperforming standalone PPO and existing hybrid reinforcement learning approaches. This framework offers a practical solution for dynamic obstacle avoidance with broader applicability to high-dimensional autonomous manipulation tasks.

Conventional lower-limb exoskeletons often rely on lateral supports, which can shift the user's center of mass and induce gait asymmetry, thereby limiting practical deployment. This study preliminarily validated a rear-supported exoskeleton with a three-layer torque–impedance control strategy in healthy adults, providing initial evidence for potential rehabilitation applications. We designed the rear-supported lower-limb exoskeleton and evaluated its performance during level walking and on an 8.5° incline in six participants using synchronized surface electromyography and kinematic data. Results showed that rectus femoris root mean square of muscle activation decreased by 20.62% on level ground and 36.38% on the slope, while the Gait Symmetry Index improved by 8.82% and 14.37%, respectively. The combination of the rear-supported structure and hierarchical control reduced muscular demand and enhanced bilateral gait symmetry, suggesting a potential design paradigm for assistive walking and future rehabilitation.