A comprehensive understanding of microstructural evolution and moisture distribution under ambient drying conditions is essential for evaluating structural performance. This study examines the drying behaviour and microstructural changes in partially water-penetrated concrete under ambient drying conditions. Concrete specimens with varying water-to-cement ratios and curing conditions were prepared, subjected to partial water penetration, and subsequently exposed to ambient drying. Hyperspectral imaging (HSI) was employed to visualise moisture profiles during drying, complemented by X-ray diffraction and mercury intrusion porosimetry to analyse microstructural changes before and after exposure. The results demonstrate the successful application of HSI for visualising moisture distribution in concrete during drying. Porosity reduction was observed post-drying, with the region further from the drying surface exhibiting a significant change. These findings offer valuable insights into the drying behaviour and microstructural evolution of partially water-penetrated concrete, contributing to the assessment of its durability performance. • Hyperspectral imaging (HSI) to obtain moisture profiles of concrete was explored. • Partially water-penetrated concrete under ambient drying was examined via HSI. • Concrete samples were varied by water-to-cement ratio and curing condition. • Microstructural changes and moisture distributions due to drying were observed.

How do citizens respond to political violence during election campaigns? Previous research suggests such events can increase support for the victim's party through a “rally ‘round the flag” effect, where the public unites against perceived threats, or an “empathy effect,” where individuals sympathize with the victim. Yet it remains unclear whether these effects vary across contexts. This study examines regional variation in public opinion following the 2022 assassination of former Japanese Prime Minister Shinzo Abe, which occurred two days before the national election. Using over 640,000 daily responses from a Voting Advice Application (VAA), we estimate changes in support for Abe's Liberal Democratic Party (LDP). Applying the generalized synthetic control method (GSCM), we find no significant change in public opinion in the area surrounding the assassination site. By contrast, in Abe's home region, support for the LDP temporarily increased and support for rights restrictions declined, suggesting that psychological proximity to the victim briefly influenced public opinion. However, surveys conducted 8 to 9 months later reveal no persistent regional effects, suggesting that neither physical nor psychological proximity to the victim produced lasting changes in political attitudes.

Moyamoya disease (MMD) has a strong genetic basis, with the rare RNF213 p.Arg4810Lys variant (rs112735431) representing a major risk factor, while the broader genetic architecture and disease-relevant vascular cell types remain incompletely understood. We conducted a genome-wide association study in Japanese individuals (n=47 656; 401 MMD cases and 47 255 controls). Population-level features at MMD risk loci were examined by regional allele frequency and haplotype analyses. We performed single-nucleus RNA-seq of superficial temporal arteries from patients with MMD (n=3). Cell type-specific enrichment of genome-wide association study signals was assessed using the Single-Cell Disease Relevance Score. Endothelial signatures were validated by integration with publicly available single-cell data sets from controls (n=5) and immunohistochemistry for candidate markers (n=1). Beyond rs112735431, we identified a genome-wide significant signal in the HDAC9-TWIST1 region (rs12530920; P=3.3×10-14; odds ratio, 1.77). Conditional analysis on rs112735431 revealed a protective RNF213 missense variant, p.Asp1331Gly (rs8074015; P=3.7×10-9; odds ratio, 0.53), whose minor allele was mutually exclusive with rs112735431-A on haplotypes. Population analysis revealed geographic variation and extended haplotype structure of the rs112735431-A allele in Japan. Single-nucleus RNA-seq identified a mesenchymal-like endothelial cell (MEC) population with selective FN1 expression. Genome-wide association study-prioritized disease genes were strongly enriched in MECs. MECs showed mesenchymal pathway activation with a regulatory program distinct from canonical endothelial states. The proportion of MECs was markedly increased in MMD (72% versus 28% in controls), and fibronectin 1 (FN1) expression in endothelial regions was confirmed by immunohistochemistry. Our findings identify a protective RNF213 p.Asp1331Gly variant (rs8074015) that is mutually exclusive with the known rs112735431-A allele. Genetic risk converges on an MEC state markedly expanded in MMD.

We propose interpolation-based algorithms that enable convolutional and transposed convolutional layers to operate with arbitrary (including non-integer) strides. A primary motivation for the proposed algorithms is to maintain a consistent temporal resolution when adapting deep neural networks (DNNs) to different sampling frequencies (SFs). To handle untrained SFs, we previously introduced SF-independent (SFI) convolutional layers, which adjust kernel weights in accordance with the target SF. However, achieving full consistency across SFs also requires the proportional adjustment of the stride, which results in non-integer values in many practical cases. Conventional algorithms for convolutional layers cannot handle such strides directly, and commonly used approaches (e.g., stride rounding or signal resampling) lead to performance degradation. To solve this problem, we propose a feature-domain interpolation framework that constructs continuous-time representations of intermediate features. This enables sampling at arbitrary stride intervals without modifying the network architecture. Through music source separation experiments, we show that the proposed algorithms maintain a strong performance across a range of SFs, including those where the stride becomes non-integer. Our analysis reveals that the proposed algorithms are robust to the choice of interpolation method and are especially effective for sources containing pitched sounds.