Browse Articles

Radar-based inspiratory-to-expiratory time ratio estimation: a validation study

Abstract Respiration is a key indicator of health and wellbeing, with metrics such as respiratory rate (RR), inspiratory time (TI), expiratory time (TE), and the inspiratory-to-expiratory time (I:E) ratio offering insights into conditions ranging from acute life-threatening and chronic diseases to symptom management. While traditional methods already measure these parameters with high accuracy, they still require contact-based sensors, limiting their practicality for continuous monitoring. This study evaluates radar as a non-contact alternative by validating multiple radar-derived respiratory metrics against impedance pneumography measurements in 30 healthy volunteers at rest. Synchronous recordings from both modalities were analysed to assess agreement across methods using descriptive statistics, scatter plots, modified Bland-Altman plots, and equivalence testing (TI: ±0.3 s, TE: ±0.3 s, RR: ±2 brpm, I:E ratio: ±0.2). Equivalence testing indicated high correlation ( p  ≤ 0.001***) across all metrics, with 81.8% (TI), 77.6% (TE), 97.2% (RR), and 85.7% (I:E ratio) of values within predefined bounds. These findings highlight radar’s potential for continuous respiratory monitoring, particularly in medical fields where minimizing patient burden is essential as in palliative, post anaesthesia, and intensive care settings.

Phase II Clinical Study of Adebrelimab and Bevacizumab Combined With Cisplatin/Carboplatin in Patients With Triple-Negative Breast Cancer With Brain Metastases (ABC Study)

PURPOSE Brain metastases (BMs) of triple-negative breast cancer (TNBC) are lethal, often associated with a limited life span and lack of effective antitumor agents. Here we reported a triple combination therapy consisting of adebrelimab, bevacizumab, and cisplatin/carboplatin in BMs of TNBC. METHODS This phase II clinical trial involved patients with TNBC with active BMs. Participants were administered with adebrelimab, bevacizumab, and cisplatin/carboplatin until disease progression or unacceptable toxic effects. The primary end point was the objective response rate in CNS (CNS-ORR) according to the Response Assessment in Neuro-Oncology BMs criteria, and the secondary end points included the clinical benefit rate in CNS (CNS-CBR), progression-free survival (PFS), overall survival (OS), the first progression site, and safety. RESULTS A total of 35 patients were enrolled and treated, and the median lines of previous treatment were 2 (range, 0-4). The confirmed CNS-ORR was 77.1% (27/35, 95% CI, 59.9 to 89.6), and the CNS-CBR was 80.0% (28/35, 95% CI, 63.1 to 91.6). The median overall PFS was 8.3 months (95% CI, 5.8 to 11.5), whereas the median CNS-PFS was 10.3 months (95% CI, 7.4 to 14.3) and the median OS was 21.1 months (95% CI, 13.2 to not reached). Among the 28 patients who progressed, progression was intracranial-only in 32.1% (9/28) patients, extracranial-only in 35.7% (10/28) patients, and both in 32.1% (9/28) patients. The incidence of grade ≥3 treatment-related adverse events was 65.7% (23/35). Treatment-related serious adverse events occurred in five patients (14.3%), and no treatment-related deaths were reported. CONCLUSION The combination of adebrelimab, bevacizumab, and cisplatin/carboplatin was the first regimen to demonstrate promising intracranial antitumor activity and prolonged PFS and CNS-PFS, along with a manageable safety profile, warranting further investigation.

A spam detection model based on the discriminative TF-IDF belief rule base

Abstract Novel spam with rapidly evolving content faces a scarcity of labeled data in its early stages. Yet, current detection models rely heavily on large datasets and high-dimensional features, leading to poor generalization and opaque decisions when data is scarce. This opacity hinders error tracing and limits their use in early threat detection and response. The belief rule base (BRB), as an expert system, demonstrates effective learning under small-sample conditions, and its rule-based reasoning mechanism provides decision interpretability. However, high-dimensional features may cause combination explosion. To address these issues, a BRB spam detection model based on the Discriminative term frequency-inverse document frequency (TF-IDF) method (DTI-BRB) is proposed in this paper. By discriminating whether terms are more indicative of ham or spam, the Discriminative TF-IDF method converts raw text into low-dimensional features, thereby effectively resolving the combination explosion problem inherent in the traditional BRB model. Through two case studies under small-sample conditions, the effectiveness of the proposed model is validated. With only 200 samples, it achieves accuracies of 91.5% and 95.5% in the two cases, respectively, exhibiting excellent predictive performance and interpretability.

Critical Dynamics in the Association Cortex Predict Higher Intelligence in Typically Developing Children

Neuronal network models have indicated that the so-called critical dynamics facilitate efficient information processing, while criticality disruptions were linked to neuropathology through excitation/inhibition (E/I) imbalances. However, there is limited empirical evidence for a relationship between critical brain dynamics and cognition in healthy children and adolescents. Here, we investigate how these dynamics relate to intelligence in a developing cohort. We recorded eyes-open resting EEG in 128 children (6–19 years, 72 female) and quantified near-critical dynamics in the alpha-band using functional excitation/inhibition ratio ( f E/I) and in nonoscillatory activity using the 1/ f aperiodic exponent of the power spectrum. We devised models relating intelligence to f E/I and 1/ f exponent across seven Yeo7 functional brain networks ranked from lower-order sensorimotor to higher-order association networks. We observed significant correlations between f E/I and 1/ f exponent and IQ in association cortices, in contrast to sensorimotor cortices. Children in the high-IQ group had f E/I ratios closer to the theoretical critical value of 1 in association cortices compared with the low-IQ group. The association–sensorimotor axis rank moderated the associations between 1/ f exponent and IQ, these associations decreasing on a gradient across the hierarchy of the Yeo7 networks. Age and rank moderated the f E/I–IQ association, with the association–sensorimotor effect size gradient most visible in adolescents. Together, the results suggest that individual variation in criticality-sensitive biomarkers in association networks may be linked to IQ differences in an age-dependent manner, consistent with the hypothesis that developmental modulation of critical dynamics across the cortical hierarchy may support more efficient cognitive processing.

Tailored Phosphate Leaving Groups Direct Pathway-Dependent Self-Assembly

Physical model study on the mechanism of floor heave for the deep-buried roadway excavated in soft rock of gently inclined thin strata

Preventing chick culling in the poultry industry with a new biomarker for rapid in ovo gender screening

Abstract Chicken eggs are one of the most consumed foods worldwide. However, the practice of chicken culling in the poultry industry involves unnecessary animal suffering and finding a way to put an end to this has become a societal priority. One approach that has been propagated as acceptable is based on the selection of female eggs early in the incubation process and the devitalization of the male eggs. It is with this objective in mind that we searched for a biomarker for early gender screening in eggs. Applying an untargeted mass spectrometry approach, we profiled allantoic fluid of different day-old eggs and identified the feature 3-[(2-aminoethyl)sulfanyl]butanoic acid (ASBA) as a strong biomarker for in-ovo gender prediction for day-9 old embryos. In the present work, we describe the identification of ASBA as a new biomarker in allantoic fluid for gender screening and the optimization of a high throughput assay using acoustic droplet ejection-mass spectrometry (ADE-MS). Special attention is given to the optimization of ADE-MS compatible liquid handling and the development of the data processing to ensure a reliable gender prediction. We have been able to accurately determine the gender of day-9 eggs in a cohort of 154 samples with a prediction accuracy of 95.5%, with a throughput of 1800 samples per hour for the prototype, which may vary in production systems.

Optimization of drawing parameters based on top-coal flow law in thick-seam caving mining

Cutaneous Alternating Current Stimulation Can Cause a Phasic Modulation of Speech Perception

Segregating important stimuli from distractors is crucial for successful speech perception. Neural activity synchronized to speech, also termed “neural speech tracking,” is thought to be instrumental for this purpose. However, whether neural tracking of targets and distractors both play a similarly important role for speech perception in a setting with multiple competing speakers is rarely examined. In 61 human participants (30 males, 31 females), we used transcranial alternating current stimulation (tACS) to presumably manipulate neural tracking of two simultaneously presented sequences of rhythmic speech while participants attended to one of them. A random temporal relationship between speech streams allowed us to disentangle effects of tACS on target and distractor processing and to examine their combined effect on a behavioral measure of speech perception. We found that the phase relation between tACS and both target and distracting speech modulated word report accuracy to a similar degree. This effect was observed during bilateral tACS over auditory regions and the inferior frontal gyrus and, importantly, in a control “shunt” group that received near-identical cutaneous stimulation but ∼50% reduced brain stimulation. These results imply that, although tACS phase modulates the perception of both target and distracting speech, the cutaneous stimulation accompanying tACS can induce phase effects in speech perception that resemble those observed with conventional tACS. Our finding illustrates the urgent need to control for cutaneous stimulation in tACS studies.

Research and implementation of intelligent clothing personalized customization system based on deep learning

Abstract This study presents an intelligent personalized garment customization system that integrates deep learning methodologies. The system employs a microservices architecture to unify four core modules: body measurement data extraction, style preference learning, virtual try-on visualization, and design recommendation generation. We propose a novel CNN-Transformer-GAN architecture, specifically tailored for personalized garment design tasks, achieving exceptional accuracy. Experimental results demonstrate that the system attains a mean absolute error (MAE) of 0.38 cm in body measurement, an accuracy of 87.4% in style matching, and a response time of 285 ms. Compared to existing approaches, the proposed system improves measurement accuracy by 38.7% and delivers visualization quality comparable to metaverse-based systems. To evaluate user experience, we conducted two complementary studies: (1) a controlled single-blind user study with 120 participants, which yielded satisfaction scores between 4.42 and 4.65 across recommendation accuracy, interface usability, and visualization quality; and (2) a large-scale deployment test involving 250 users, which reported an average overall satisfaction score of 4.55 out of 5.0. This research helps bridge the gap between artificial intelligence and personalized fashion design, advancing resource-efficient customization and better alignment with consumer needs in the apparel industry. By integrating state-of-the-art deep learning techniques with responsive user preference modeling, the system offers an innovative solution for intelligent garment customization.

Post-Ictal Sleep Changes in Human Focal Epilepsy

Bidirectional interactions between sleep, seizures, and epilepsy remain incompletely understood. Evidence from animal models and people with focal epilepsy suggest that seizures may engage mechanisms of memory consolidation during post-ictal sleep to reinforce and strengthen synaptic connections within the pathological networks that generates seizures, termed seizure-related consolidation (SRC). Human studies of post-ictal sleep changes supportive of SRC, however, are limited by small sample size and restricted observations of post-ictal sleep. We investigated the interplay between seizures and sleep by analyzing sleep–wake and seizure catalogs derived from continuous local field potential (LFP) recordings in 11 people (6 males and 5 females) with drug-resistant focal epilepsy implanted with novel investigational devices and living in their natural environments. Our findings demonstrate that post-ictal rapid-eye-movement sleep duration is reduced, whereas slow-wave sleep duration, slow-wave LFP spectral power, and waveform slope are increased compared with inter-ictal nights without preceding seizures. The most significant changes localize to the epileptogenic networks generating the participants’ habitual seizures. These results reveal parallels between SRC and physiological memory consolidation, providing novel insights into the potential role of post-ictal sleep in strengthening epileptic neural engrams and may have implications for targeted disruption of post-ictal sleep and SRC in focal epilepsy.

Human stem cell-derived neurons establish functional inhibitory–excitatory cortical circuits in a chimeric transplantation model

Transcriptional remodeling of cardiomyocytes and fibroblasts during post-myocardial infarction recovery

Abstract Myocardial infarction (MI) results from reduced coronary blood flow, leading to oxygen deprivation and impaired systolic and diastolic function, which increases the risk of cardiac arrhythmias. Various cardiac cell types respond to this stress to preserve heart function, but the precise, cell-type-specific mechanisms remain poorly understood. To investigate these responses, we performed single-nucleus RNA sequencing (snRNA-seq) on left ventricular tissue from mouse hearts at baseline (Day 0) and at 1 and 4 weeks post-MI. This enabled us to characterize transcriptional changes across major cardiac cell types. We observed significant shifts in the transcriptional states of cardiomyocytes (CMs) and fibroblasts (FBs) cell populations following MI. CMs showed a major transcriptional modulation from healthy to diseased state during early chronic phase of post-MI, however, the recovery phenotype was observed during the late chronic phase, suggesting a natural compensatory response of CMs against the ischemic stress. FBs exhibited dynamic transcriptional changes consistent with roles in post-MI healing and fibrosis. In addition, inferred alterations in cell-cell communication networks highlighted changes in intercellular signaling pathways, shedding light on disrupted crosstalk in the injured heart. Together, our findings provide a comprehensive transcriptional landscape of cardiac cell populations, especially CMs and FBs, following MI and identify potential molecular targets for therapeutic intervention.

Causal Evidence for Prefrontal-Motor Coupling in Reward-Responsive Goal-Directed Behavior

Interventions for major depressive disorder (MDD) using noninvasive brain stimulation often target the left lateral prefrontal cortex to reduce symptoms of anhedonia. However, the electrophysiological mechanism by which symptoms of anhedonia arise from dysfunction of the lateral prefrontal cortex is poorly understood. Furthermore, multiple constructs related to reward processing within the positive valence systems have been found to be blunted with anhedonia. To disentangle components of the positive valence systems, human participants with an active episode of MDD performed an adaptive version of the expenditure of effort for reward task (Adaptive-EEfRT) and three constructs of reward processing were investigated: effort motivation, reward valuation, and reward-responsive goal-directed behavior. Individual differences analysis revealed that symptoms of anhedonia were most strongly related to blunted reward-responsive goal-directed behavior. By recording electroencephalography (EEG), spectral analysis revealed that the amplitude of left prefrontal delta oscillations (2–3 Hz) increased with each reward construct and beta oscillations decreased over the motor cortex contralateral to the hand used for effort exertion. Phase–amplitude delta-beta coupling between the left prefrontal cortex and contralateral motor cortex was significantly increased for each reward construct. When cross-frequency transcranial alternating current stimulation (tACS) was delivered to 26 females and 9 males to mimic prefrontal-motor delta-beta coupling, only reward-responsive goal-directed behavior was improved relative to placebo and active-control tACS. These findings provide causal evidence that the left prefrontal cortex supports reward-responsive goal-directed behavior via cross-frequency coupling. Interventions for anhedonia that target the left prefrontal cortex might be improved by frequency-specific targeting of delta-beta coupling.

Self-supervised non-dominated sorted model for co-clustering

Abstract Co-clustering is widely used for data analysis that independently reveals the clustering structures of rows and columns while also identifying their inter-relationships, which renders it more informative than conventional one-way clustering methods. Co-clustering is to not only cluster the samples and features of original data, but also mine the relationship between samples and features, and this is naturally a multi-objective problem. However, researchers frequently utilize the method of single-objective optimization to solve the co-clustering issue, while disregarding its multi-objective nature, and the side information in the original data is also ignored. To address these problems, we propose a self-supervised non-dominated sorted model for co-clustering (SNSC), which is represented by a group of multi-objective functions. The model not only perfectly aligns with the multi-objective nature of co-clustering tasks but also utilizes the supervised information in the original data. The objective function group consists of four objective functions acting on the original data and similarity matrix respectively. The heuristic initialization method with self-supervised properties is used in conjunction with the random initialization method, which improves the efficiency of the model and reduces the likelihood of converging to local optima. The overall model remains unsupervised, as all the supervised information is derived from the original data. Further, the algorithm for the SNSC model is designed by using the idea of the genetic algorithm, which is theoretically supported, and the complexity analysis of the algorithm is given. Finally, experiments on 12 datasets and 5 comparison algorithms show that the SNSC algorithm has significant advantages.

An IoT-enabled CRNN framework for secure wearable sensor-based activity recognition in physical education

Fuzzy-logic-controlled DVR for enhancing the fault resilience of wind energy conversion systems

PCSK9 inhibitors patterns of use in France from nationwide repeated cross-sectional and cohort studies

Author Correction: A Model of Exposure to Extreme Environmental Heat Uncovers the Human Transcriptome to Heat Stress

Integrative mendelian randomization approaches for therapeutic target prioritisation in immune-mediated diseases

Abstract Immune-mediated diseases (IMD) encompass a wide range of autoimmune and inflammatory disorders with aetiology related to immune system dysfunction, signifying a disease area with great potential for drug repurposing. In this study, we employed the genetically informed Mendelian Randomization (MR) method with two distinct exposure types: immune blood cell abundance and protein quantitative trait loci (pQTL) to validate and repurpose 834 drug targets which have been investigated for IMD treatment. Utilizing two-sample MR, we first established causal relationships between major peripheral immune cell types and 14 IMD. Robust associations, particularly with eosinophils, were confirmed across diseases such as asthma, eczema, sinusitis, and rheumatoid arthritis, revealing 59 high-confidence relationships. Intragenic variants associated with causal immune cell types were then extracted to create instruments for 371 existing IMD drug targets (“intermediate trait” MR). In parallel, we leveraged four large blood plasma protein QTL datasets to obtain complementary instruments for 361 targets (“pQTL” MR). In the intermediate trait MR analysis, we identified 811 gene-IMD associations (p-value < 0.05; 137 pairs below Bonferroni-adjusted p-value threshold), 169 of which were supported by strong colocalisation evidence (PP H4 ≥ 0.8). In the pQTL MR analysis, we similarly found 841 protein-IMD associations (p-value < 0.05; 90 pairs below Bonferroni-adjusted p-value threshold), 83 of which were confirmed with colocalization. Comparison with a list of approved drugs indicated low sensitivities across disease outcomes for both exposure types (intermediate trait MR: 0.49 ± 0.23 SD, pQTL MR: 0.28 ± 0.12 SD). Drug targets identified in the pQTL and intermediate trait MR analyses show limited overlap (13% at nominal p-value and 36% at Bonferroni-adjusted p-value threshold), presenting a comprehensive source of drug repurposing opportunities when the two approaches are combined.