Browse Articles

Structure-based identification of compounds with potential as selective BLT1 antagonists

An advanced multi-attribute decision-making model for Urban transportation planning based on complex intuitionistic fuzzy sets with hierarchical parameters

Mapping Plasmodium transitions and interactions in the Anopheles female

Abstract The human malaria parasite, Plasmodium falciparum , relies exclusively on Anopheles mosquitoes for transmission. Once ingested during blood feeding, most parasites die in the mosquito midgut lumen or during epithelium traversal 1 . How surviving ookinetes interact with midgut cells and form oocysts remains poorly understood, yet these steps are essential to initiate a remarkable growth process culminating in the production of thousands of infectious sporozoites 2 . Here, using single-cell RNA sequencing of both parasites and mosquito cells across different developmental stages and metabolic conditions, we unveil key transitions and mosquito–parasite interactions that occur in the midgut. Functional analyses uncover processes that regulate oocyst growth and identify the Plasmodium transcription factor PfSIP2 as essential for sporozoite infection of human hepatocytes. Combining shared mosquito–parasite barcode analysis with confocal microscopy, we reveal that parasites preferentially interact with midgut progenitor cells during epithelial crossing, potentially using their basal location as an exit landmark. Additionally, we show tight connections between extracellular late oocysts and surrounding muscle cells that may ensure parasite adherence to the midgut. We confirm our major findings in several mosquito–parasite combinations, including field-derived parasites. Our study provides fundamental insight into the molecular events that characterize previously inaccessible biological transitions and mosquito–parasite interactions, and identifies candidates for transmission-blocking strategies.

Comprehensive evaluation of sound absorption performance in porous concrete pavement with crushed stone base course

Abstract This study comprehensively evaluates the sound absorption performance of porous concrete pavements with crushed stone base layers of varying thicknesses and particle sizes. The noise reduction potential of pavement structures was assessed using the standing wave tube method, and overall sound absorption was quantified via the full-frequency domain average sound absorption coefficient. Results indicate that the presence of a crushed stone base layer substantially enhances sound absorption, with thicker bases providing greater improvements. While particle size and surface layer type also influence absorption, their effects are markedly smaller than that of base course thickness. These findings suggest that optimizing base layer thickness, combined with high-porosity or fine-aggregate surface layers, can effectively improve the acoustic performance of porous concrete pavements, providing practical guidance for noise mitigation in road design.

Mortality impacts of rainfall and sea-level rise in a developing megacity

Bottom-up design of Ca2+ channels from defined selectivity filter geometry

Double stochastic resonance in stink bug sexual communication

Deeply divergent human exposure to food crises across socioeconomic pathways

Prognostic value of the Braden scale for short-term mortality in critically ill patients with traumatic brain injury

SRAM based Gaussian noise generation for post quantum cryptography

A two-stage dc–dc converter with high voltage gain and reduced current ripple for efficient PV energy harvesting

Interpretable multi-model deep learning framework for automated four-class diagnosis of ocular toxoplasmosis using fundus imaging

Associations of serum CA 15–3 with interstitial lung disease, PM/Scl100 antibodies and other disease characteristics in systemic sclerosis

Decoding cell-class specific roles of non-coding variants in human retina

AI discovers learning algorithm that outperforms those designed by humans

The design of vanadium–titanium co-doped zircon pigments and color mechanism

Developing a conceptual framework for adolescent disaster resilience education in Iran

Adaptive finite-time fault-tolerant control scheme of UAV against combined faults

Built environment disparities are amplified during extreme weather recovery

Dynamic context-aware multi-modal deep learning for longitudinal prediction of Parkinson’s disease progression

Abstract Accurately forecasting the progression of Parkinson’s disease (PD) motor symptoms in early-to-moderate stages is essential for timely intervention and personalized patient care but remains challenging due to heterogeneous and longitudinal symptom evolution. We present a novel dynamic context-aware multi-modal deep learning framework that predicts future motor symptom severity by integrating advanced voice biomarkers with signal processing techniques, clinical progression features, demographic metadata, and semantically enriched patient summary embeddings derived from comprehensive clinical narratives via state-of-the-art natural language processing. Leveraging bidirectional LSTMs augmented with multi-head self-attention, our architecture captures complex temporal dependencies while preventing information leakage. To ensure robust evaluation despite limited sample size (42 patients), we implemented repeated 5-fold cross-validation at the patient level (8 repetitions, 40 total folds), substantially exceeding standard evaluation rigor. Our approach achieves exceptional performance ( $$\hbox {R}^2$$ = 0.9925 ± 0.0027, RMSE = 0.67 ± 0.19, MAE = 0.50 ± 0.15) with all 40 folds achieving $$\hbox {R}^2$$ > 0.989, significantly outperforming classical machine learning baselines ( $$p < 1 \times 10^{-5}$$ and 0.002785) and all previously published methods on this dataset. Cross-validated ablation studies (240 total model trainings across 6 configurations) reveal that clinical features establish a strong baseline ( $$\hbox {R}^2$$ = 0.9887 ± 0.0043), while text embeddings provide the largest incremental gain (3.82% RMSE reduction). Voice biomarkers contribute modestly to accuracy (2.72%) but substantially enhance stability (10-fold lower variability). The full multi-modal model achieves optimal performance (7.50% RMSE reduction vs. clinical-only) with the lowest variability (CV = 0.27%), demonstrating that dynamic cross-modal fusion enhances both accuracy and robustness. These findings, validated through 40 independent evaluations with each patient tested 8 times, demonstrate that integrating engineered temporal dynamics and contextual embeddings through advanced temporal modeling enables accurate longitudinal predictions of early-to-moderate PD progression. Complete code and implementation details are publicly available to ensure reproducibility.