Annals of the New York Academy of Sciences最新文献

Extreme weather can cause urban groundwater levels (GWLs) to rise sharply, making anti-uplift performance critical for underground structures. We present a drainage-pressure relief anti-uplift technique (DPRAT) that integrates the Dupuit circular island model, the Thiem equation, and GWL distribution assumptions into an intelligent control system. The system activates automatically when the measured water head exceeds a design threshold, draining groundwater to relieve hydrostatic pressure on buried structures. Tests and simulations in Chengdu's expansive soil areas confirm that anti-uplift failure results primarily from buoyant forces and soil expansion. To ensure adequate safety margins, the target drainage level is calibrated to maintain system inactivity approximately 80% of the time under normal conditions. Four years of field monitoring demonstrate that DPRAT effectively maintains GWLs below the design datum during extreme rainfall events. A 50-year life cycle assessment reveals that DPRAT reduces cradle-to-grave carbon emissions by up to 97.5% compared with conventional uplift anchors, representing a substantial shift from high-energy construction methods to low-carbon alternatives.

The rapid proliferation of automotive augmented-reality head-up displays (AR-HUDs) in intelligent-driving systems has increased the need for rigorous multidimensional performance evaluation—a task complicated by data heterogeneity and cognitive uncertainty. To address this dual challenge, we propose a comprehensive decision-making method that combines Pythagorean fuzzy set and Hamacher confidence-weighted aggregation operator (PyHa-CWAO) and embeds it in a novel hierarchical evaluation framework that separately appraises the static basic information layer (SBIL) and the dynamic AR fusion layer (DAFL) before synthesizing the results. We further propose a hyperbolic tendency-confidence fused scoring model for the first time, which simultaneously captures subjective hesitation and confidence weights while amplifying interalternative discrimination. Combining DEMATEL for causal-relationship elicitation with VIKOR for multicriteria ranking yields the integrated PyHa-CWAO–DEMATEL–VIKOR decision model. A driving-simulation and eye-tracking experiment evaluated 10 AR-HUD prototypes (six DAFL, four SBIL) across 12 objective and subjective indicators. Compared with three established multicriteria decision-making methods, the proposed model improves ranking consistency by 14%, confirming its robustness and practical value. Overall, this work offers an accurate, resilient, and extensible decision tool for AR-HUD and broader extended reality interfaces, enriching fuzzy-decision theory and providing a methodological foundation for future virtual human−computer interaction optimization.

To address the challenge that existing lung nodule segmentation algorithms face in balancing high accuracy with a lightweight design, we propose LLNS-Net, a compact yet effective lung nodule segmentation network. In the feature-mining encoder, convolutional residual blocks operating at multiple scales extract both shallow and deep nodule information from CT images, while an efficient multiscale attention mechanism enriches semantic representations. A subsequent feature enhancement module explores and leverages correlations among the outputs of the encoder's submodules. Within this module, we introduce an enhanced mixed local channel attention (E-MLCA) mechanism and a reinforced multiscale feature module to further strengthen cross-scale feature learning. The decoder aggregates features from four decoding layers and applies subchannel enhancement to refine the segmentation map. This design improves boundary smoothness and more accurately preserves the true morphology of nodule regions. Experimental results demonstrate that, compared with mainstream methods such as HmsUnet, MSA-Unet, and H-vmunet, the intersection over union of LLNS-Net improves by 1.86%, 0.27%, and 7.62%, respectively. Additionally, the generated feature maps exhibit smoother boundaries and superior visual quality compared to those produced by leading medical image segmentation algorithms.

Alzheimer's disease (AD) dementia is associated with marked disruptions in resting-state eyes-closed electroencephalographic (rsEEG) rhythms, particularly in the periodic alpha band (8–12 Hz), suggesting impaired vigilance regulation. In contrast, the aperiodic rsEEG component, reflecting global cortical arousal, has been reported to remain unchanged. This exploratory study examined periodic and aperiodic EEG activity in patients with mild cognitive impairment due to AD (ADMCI) during transitions from quiet wakefulness to light sleep. EEG datasets (∼30 min) from 19 ADMCI patients and 18 matched cognitively unimpaired older adults (control) were analyzed. Vigilance stages were scored using a reduced version of Hori's system, distinguishing the alpha-dominant wakefulness stage and the theta-dominant light sleep (ripples) stage. EEG spectra were parameterized using the specparam algorithm. ADMCI participants showed reduced reactivity of individual alpha power between the wakefulness and ripples stages compared to the control group. Conversely, both groups exhibited comparable increases in fronto-central theta power and steepening of the aperiodic slope and offset. No group differences emerged in aperiodic exponent and offset, although statistical power was limited by modest sample size. Overall, EEG alpha rhythms reflecting vigilance regulation are disrupted in prodromal AD, while periodic and aperiodic signatures of sleep onset are relatively preserved, suggesting selective vulnerability of attentional thalamocortical systems.

High myopia is a serious global health issue, with particularly significant impacts on the central nervous system. However, the role of glymphatic system dysfunction and associated neural alterations in high myopia remains largely unexplored, and the underlying brain−eye pathological interactions are poorly understood. Our multimodal magnetic resonance imaging (MRI) study integrates diffusion tensor image analysis along the perivascular space (DTI-ALPS), choroid plexus volumetry, and regional homogeneity (ReHo) analyses to evaluate glymphatic function and neural activity changes in high myopia. Compared to controls, patients exhibited significantly reduced DTI-ALPS indices, enlarged choroid plexus volumes, and distinct ReHo alterations, including increased activity in the superior frontal gyrus, but decreased activity in the calcarine fissure and surrounding cortex, as well as the superior temporal gyrus. Notably, the DTI-ALPS index correlated positively with refractive error but negatively with axial length, whereas choroid plexus volume showed the opposite pattern. Furthermore, glymphatic dysfunction correlated with abnormal ReHo in key brain regions. These findings indicate a pathological cascade linking axial elongation, impaired glymphatic clearance, and disrupted neural synchronization—a pathophysiological state that may underlie the broader neurological risks associated with high myopia. This framework integrates multimodal evidence to elucidate brain−eye interactions, incorporating complementary insights derived from structural, diffusion, and functional MRI methodologies.

Child sexual abuse cases present unique challenges for conviction and exoneration due to the typical reliance on the child's testimony as the primary source of evidence. Yet, little empirical research has examined wrongful convictions and exonerations in such cases. In this study, we analyzed the largest sample of child sexual abuse exonerations to date (N = 326), drawing on cases documented by the National Registry of Exonerations between 1989 and 2024. Descriptive results revealed several patterns unique to child sexual abuse exonerations. Unlike other types of crimes, perjury or false accusation was the most frequent contributor to wrongful convictions, while mistaken eyewitness identifications and false confessions were rare. Physical and forensic evidence were presented at many of the original trials, yet in most cases this evidence was later determined to be false or misleading. Post-conviction witness recantation was the most common factor contributing to exoneration, whereas few exonerations were based on DNA evidence or actual perpetrator identification. Case narratives highlight the influential role of family members and authorities in prompting children's false allegations. These findings provide novel insights to inform investigative, prosecutorial, and adjudicative approaches in child sexual abuse cases and reduce risk for future miscarriages of justice.