Browse Articles
Discover research articles across all indexed journals
How language, culture, and geography shape online dialogue: Insights from Koo
Founded in India in 2020, the microblogging site ‘Koo’ launched as an alternative to mainstream social media platforms, with the explicit aim of catering to non-Western communities in their vernacular languages, and capitalising on a period of tension between the Indian government and Twitter which led many users to seek Twitter-alternatives. Drawing on a near-complete dataset totalling over 71M posts and 399M user interactions, we show how Koo attracted users from several countries including India, Nigeria and Brazil, but with variable levels of sustained user engagement. We highlight how Koo’s interaction network was shaped by multiple country-specific migrations displaying strong divides between linguistic and cultural communities, for instance, with English-speaking communities from India and Nigeria largely isolated from one another. Finally, we analyse the content shared by different linguistic communities and identify cultural patterns which, we speculate, promoted similar discourses across language groups. Our results show that for language groups of similar sizes, Indian languages fostered higher discourse diversity than non-Indian languages, possibly highlighting synergistic effects which boosted the uptake and retention of these groups. Despite this, Koo failed to capitalise on this synergy and ceased operations in July 2024. With this context, our study points to some of the possible reasons why the multilingual and politically diverse platform Koo struggled to remain sustainable, failing to stave off competition from its US-based competitors, despite its commitment to cultivating support for the different vernacular communities of Indian social media users.
Hearing people speak in different accents biases voice discrimination
Abstract Voice discrimination is a fundamentally different task when matching utterances than when matching identity across different words. Discriminating between speakers of different languages makes the task even harder because unfamiliar languages contain different phonemes that are less easily matched. Discriminating between people with different accents may also be difficult as even if the same words are uttered, the phonemes are different. To test this, we created a set of voices using voice cloning that have the same or different identity or accent (UK, Poland, and China) and speaking different phrases. We tested how accent, sentences, and identify affected bias to conflate different identities as the same person. Contrasting identity between different and same increased bias to judge people as the same by about 62%. Contrasting accent between different and same independently increased bias to judge people as the same by about 10%. Contrasting between different and same sentences, changed bias to label people the same more when the accents were different than when they were the same. Our results are consistent with the idea that we are biased to think people typically speak with one accent. Thus, accents affect voice discrimination independently of language familiarity.
The relative contributions of subjective and musical factors in music for sleep
Previous research into music for sleep has focused on describing the types of musical characteristics associated with such music. The current study aimed to increase understanding of sleep music by investigating subjective perceptions of listeners associated with music that is considered sleep inducing. A listening experiment asked participants to rate musical excerpts along subjective evaluation dimensions related to valence, arousal, and dissociation. Musical features of the stimuli presented were extracted to compare the relative contribution of subjective and objective aspects. Our results reveal important roles for valence and arousal, and highlight notions of comfort, liking, and dissociation that contribute to music that is perceived to be sleep inducing. The musical analysis largely conformed with previous research, with an emphasis on brightness. However, the subjective ratings overshadowed the musical features in predicting what music was perceived to be most sleep inducing. Our findings are relevant for music recommendation applications that rely on a features-based approach to selecting music to fulfil particular purposes. As applications increasingly emphasise personalisation, we have shown that an account of subjective appraisals is crucial to predict listeners’ experiences, providing recommendations for individually targeted therapeutic applications.
Validation and standardization of the content of three di-caffeoylquinic acid in Korean Ligularia fischeri by region of origin
The additive effect of the estimated glucose disposal rate and a body shape index on cardiovascular disease: A cross-sectional study
Background The glucose disposal rate (eGDR) and a body shape index (ABSI) are predictors strongly associated with cardiovascular disease (CVD) and outcomes. However, whether they have additive effects on CVD risk is unknown. This study aimed to investigate whether combined assessment of eGDR and ABSI could improve prediction of CVD risk. Methods The current study used data from NHANES from 1999 to 2018 and included 14,237 participants. Receiver operating characteristic (ROC) curve was used to evaluate the performance of each indicator in predicting CVD. Machine-learning algorithms were applied to screen variables to adjust the model. Finally, the ROC curve, net reclassification improvement (NRI), integrated discrimination improvement (IDI), calibration curve and decision curve analysis (DCA) were used to evaluate the predictive performance of the combination of eGDR and ABSI. Results The ROC curve showed that eGDR (C-statistics: 0.7255) and ABSI (0.7093) had the highest predictive performance. Among 14,237 participants, multivariate logistic regression showed that lower eGDR (≤6.448) and higher ABSI (≥0.086) significantly increased CVD risk (OR = 11.792, P < 0.05). The model adjusted by machine learning significantly improved CVD risk prediction (Model 3 vs. Model 1, C-statistics: 0.849 vs. 0.753). These findings were also consistent in the NRI (model 3 vs. model 1: 0.108), IDI (0.107), calibration curve, and DCA analyses. Subgroup analyses confirmed the robustness of these findings, with enhanced predictive performance particularly in younger populations. Conclusion The eGDR and ABSI have potential additive effects on predicting CVD risk, and have excellent predictive performance, which can evaluate cardiovascular risk more comprehensively.
Assessment of plant diversity index in degraded desert grassland using UAV hyperspectral multimodal data and Encoder-CNN
Harnessing subtractive genomics for drug target identification in Streptococcus agalactiae serotype v (atcc baa-611 / 2603 v/r) strain: An in-silico approach
Developing a therapeutic target for bacterial disease is challenging. In silico subtractive genomics methodology offer a promising alternative to traditional drug discovery methods. Streptococcus agalactiae infections depend on two crucial criteria: drug-resistance and the existence of virulence factors. It is essential to underline that S. agalactiae strains have emerged to be resistant to several drugs. Hence, there is a need for research on novel drugs and techniques that are potent, economical, productive, and dependable to combat S. agalactiae infections. In this study advanced computational techniques were exploited to examine potential druggable targets exclusive to this pathogen. Our study uncovered 200 non-homologous proteins in S. agalactiae serotype V (Strain ATCC BAA-611/ 2603 V/R) and identified 68 essential proteins indispensable for the bacterium’s survival. Therefore, these 68 proteins are potential targets for drug development. Subcellular localization analysis unveiled that the pathogen’s cytoplasmic membrane contained essential proteins among these vital non-homologous proteins. On the other hand, based on virulent protein predictions, six proteins were seen to be virulent. Among these, we prioritized two proteins (Sensor protein LytS and Galactosyl transferase CpsE which are exclusively found in S. agalactiae) as potential druggable targets and selected them for further structural investigation. The proteins chosen could serve as a foundation for the identification of a promising therapeutic compound that has the potential to neutralize these enzymatic proteins, thereby contributing to the reduction of risks linked to the drug-resistant S. agalactiae.
Exploring the molecular mechanism of budesonide enteric capsules in the treatment of IgA nephropathy based on bioinformatics
Evaluating the MT-CYB and MT-ATP6 variations in COVID-19 patients: A case-control study
The complications and lingering effects of COVID-19 have caused a global health crisis, prompting intense investigations into the mechanism by which SARS-CoV-2 causes the disease. The majority of symptoms associated with this infectious disease are linked to mitochondrial dysfunction, and recent studies indicate that SARS-CoV-2 can impair host mitochondrial function. This study aims to investigate mutations in the MT-CYB and MT-ATP6 genes of mtDNA in COVID-19 patients and their association with disease outcomes. Out of 110 individuals enrolled, 30 were diagnosed with COVID-19, while the remaining 80 were healthy. Following DNA isolation from the blood samples, the MT-CYB and MT-ATP6 genes were amplified through PCR, purified, and sequenced using Sanger sequencing. In the MT-CYB gene, 9 distinct mutations were found. Among these, novel mutation m.14942A > C was more prevalent in COVID-19-positive individuals than COVID-19 negative controls (p = 0.012 < 0.05) and exhibited a significant correlation with the disease as OR (95% CI) = 19.75 (2.264–172.246). According to in-silico analyses, this mutation is deemed deleterious and decreases the stability of the CYB protein. In case of the MT-ATP6 gene, among the identified 8 mutations, the m.8744T > G mutation was higher in COVID-19-positive individuals than healthy controls (p = 0.009 < 0.05) and showed a significant correlation with the disease: OR (95% CI) = 24.04 (2.81–205.62). According to in-silico analyses, this mutation is found to be pathogenic and reduces the stability of the ATP6 protein. In conclusion, one novel mtDNA mutation was identified in this study, and some of the mutations identified in the MT-CYB as well as MT-ATP6 genes, including the novel mutation, are relatively common in the COVID-19 patient group. Moreover, the findings demonstrated that these mutations may contribute to the pathogenesis of COVID-19.
Discovery of tumour indicating morphological changes in benign prostate biopsies through AI
Abstract Diagnostic needle biopsies that miss clinically significant prostate cancer (PCa) often sample benign tissue near hidden cancers. Such benign samples might still display subtle morphological signs of cancer elsewhere in the prostate. This study examined if artificial intelligence (AI) could detect these morphological clues in benign biopsies from men with elevated prostate-specific antigen (PSA) levels to predict subsequent diagnosis of clinically significant PCa within 30 months. We analysed biopsies from 232 men initially diagnosed as benign, matched for age, diagnosis year, and PSA levels-half were later diagnosed with PCa, while the rest remained cancer-free for at least eight years. The AI model accurately predicted future PCa diagnosis from initial benign biopsies (AUC = 0.82), highlighting patterns such as changes in stromal collagen and altered glandular epithelial cells. This demonstrates that AI analysis of routine haematoxylin-eosin biopsy sections can detect subtle signs indicating clinically significant PCa before it becomes histologically apparent. Such morphological patterns shed light on the broader tissue alterations induced by prostate cancer, even in benign tissue, potentially enhancing early detection and clinical decision-making.
The effects of landscape on visual preference and fatigue recovery among university students: Differences in gender, grade level and major
Exposure to natural landscapes has been shown to affect both physiological and psychological well-being, with the extent of these effects varying across different landscape types. However, the underlying mechanisms remain poorly understood. The association among stress reduction, environments characteristics and individual differences requires further investigation, particularly considering the complexity of landscape attributes and the variability of personal responses. In this study, 98 university students participated in a survey to evaluate the effects of different landscape types on visual preference and fatigue recovery. Physiological data (blood pressure, heart rate), psychological data (Perceived Restorative Scale), and visual preferences were analyzed before and after participants viewed the images of eight representative landscape space types: mountain, field, waterscape, lawn, desert, forest, artificial nature, plant. The results indicated that landscape type significantly influenced both physiological responses and emotional states, as well as participants’ perceived recovery from stress. Among the eight landscape spaces, water features and forests were reported to be the most restorative. Compared to freshmen, juniors exhibited greater improvements in physical and psychological recovery, alongside more positive evaluations of the environments. Notably, the desert landscape elicited varied responses depending on participants’ grade level and gender, suggesting that restoration effects may be modulated by individual characteristics. This may reflect an evolutionary predisposition to prefer natural features that enhance survival. These findings contribute to environmental psychology and provide valuable insights for educational practice and environmental design.
Body image and loneliness as mediators of the relationship between physical activity and exercise self-efficacy in college students
EchoMamba: A new Mamba model for fast and efficient hyperspectral image classification
The classification of hyperspectral images (HSI) is an important foundation in the field of remote sensing. Mamba architectures based on state space model (SSM) have shown great potential in the field of HSI processing due to their powerful long-range sequence modeling capabilities and the efficiency advantages of linear computing. Based on this theoretical basis, We propose a novel deep learning framework: long-sequence Mamba (EchoMamba), which combines the powerful long sequence processing capabilities of Long Short-Term Memory(LSTM) and Mamba to further explore the spectral dimension of HSI, and carry out more in-depth mining and learning of the spectral dimension of HSI. Compared with the previous HSI classification model, the experimental results show that EchoMamba can significantly reduce the training time cost of HSI and effectively improve the performance of the classification task.This study not only advances the current state of HSI classification but also provides a robust foundation for future research in spectral-spatial feature extraction and large-scale remote sensing applications.
High order Interaction and Wavelet Convolution Network for visible infrared person reidentification
Abstract Visible-infrared person re-identification (VI-ReID) remains a challenging task due to significant cross-modal discrepancies and poor image quality. While existing methods predominantly employ deep and complex neural networks to extract shared cross-modal features, these approaches inevitably discard critical primitive features during high-level feature abstraction. To address this limitation, we propose the High-order Interaction and Wavelet Convolution Network (HIW-Net) that systematically integrates primitive features at multiple feature interaction stages, thereby compensating for information loss in High-order representations. Furthermore, our framework uses wavelet convolution to mine more diverse features and solve the problem of insufficient feature extraction. We create the RegDB_shape datasets with the help of the Segment Anything Model(SAM) tool to supplement the training set. Extensive experiments on the SYSU-MM01 and RegDB datasets show the superiority of the proposed HIW-Net over several other state-of-the-art methods, proves the effectiveness of this method.
Global burden of disease due to opioid, amphetamine, cocaine, and cannabis use disorders, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021
Drug use disorders (DUDs) represent a major global health challenge, leading to substantial morbidity and mortality, while also being compounded by social and structural barriers. In this study, we examined global epidemiological trends in DUDs over the past three decades to inform clinical and public health responses. We extracted data on the incidence, deaths, and disability-adjusted life years (DALYs) attributable to DUDs from the 2021 Global Burden of Diseases, Injuries, and Risk Factors Study between 1990 and 2021. Age-standardized incidence (ASIR), mortality (ASMR), and DALY rates per 100,000 population were calculated. The analysis focused on four DUDs—opioid, amphetamine, cocaine, and cannabis use disorders—and further stratified rates by sex, Socio-demographic Index (SDI), countries, and world regions. In 2021, there were 13.6 (95% UI, 11.6–15.7) million new cases, 137,278 (95% UI, 129,269–146,181) deaths, and 15.6 (95% UI, 12.8–18.1) million DALYs attributed to DUDs. Between 1990 and 2021, the ASIR decreased by 8.1%, while the ASMR and DALY rates rose by 30.8% and 14.8%, respectively. Opioid use disorder accounted for the highest ASIR (169.4 [95% UI, 145.1–195.0] per 100,000), ASMR (1.7 [95% UI, 1.6–1.8] per 100,000), and age-standardized DALY rate (191.0 [95% UI, 156.1–222.8] per 100,000) in 2021. Sex and geographical variations were notable, with males and world regions like high-income North America, Australasia, and Eastern/Western Europe showing disproportionately higher rates. Overall, these findings highlight rising mortality and morbidity rates despite a modest decline in incidence, underscoring the need for tailored public health interventions, advancing harm reduction programs, and expanding access to treatment.
Author Correction: High-sensitivity acceleration sensor detecting micro-mechanomyogram and deep learning approach for parkinson’s disease classification
Bioinformatics analysis and experimental validation of the potential relationship between bacterial lipopolysaccharide and oral squamous cell carcinoma
Background Advances in science and medicine have led to the identification of bacterial virulence factors (including lipopolysaccharide, LPS) and their key role in the occurrence and outcome of tumors. However, the effect of LPS on oral squamous cell carcinoma (OSCC) has yet to be fully understood. Objective Hence, based on host genes related to bacterial LPS, the study investigated the potential role and mechanism of oral bacteria in OSCC via bioinformatics analysis and experimental validation. Methods The sequencing datasets of OSCC were screened using the GEO database and the bacterial LPS-related genes were searched in the GeneCards database to identify the LPS-related differentially expressed genes (LR-DEGs) in OSCC. The molecular mechanism of bacteria affecting OSCC was explored through GO and KEGG enrichment analysis, as well as protein-protein interaction (PPI) network and module analysis. Subsequently, seven algorithms were integrated to identify the LPS-related hub genes (LRHGs), and their diagnostic specificities were explored by receiver operating characteristic (ROC) and transcription levels were verified by qRT-PCR. Immune infiltration was then analyzed. Results We found a total of 345 LR-DEGs. GO and KEGG enrichment analysis demonstrated that the LR-DEGs were mainly enriched in inflammation-related pathways including cytokine-cytokine receptor interaction and IL-17 signaling, suggesting that bacteria may promote the development of OSCC through LPS-related gene-mediated inflammatory response. PPI and module analysis results revealed the presence of a complex regulatory network involving LR-DEGs. Totally, five LRHGs (including Cxcl8, Cxcl10, Il-1β, Il-6 and Mmp9) were screened out. Based on ROC analysis, the five LRHGs represented potential diagnostic biomarkers for OSCC (AUC > 0.7). The results of qRT-PCR, WB, ELISA and IF indicated that all LRHGs were upregulated in OSCC (P < 0.05). Immune infiltration analysis showed that LRHGs were closely related to the immunocyte infiltration level, suggesting a potential target for OSCC immunotherapy. In this study, 345 LR-DEGs and 5 LRHGs were identified in bacterial LPS-regulated OSCC progression. Importantly, the 5 LRHGs may mediate the OSCC progression in the host through inflammation-related pathways. These findings suggest that bacterial LPS plays a vital role in OSCC. Conclusion Our study provides novel insights into the pathogenesis and development of oral bacteria in OSCC. The LRHGs identified in this study are crucial for the diagnosis of OSCC, and also provide new insights into the molecular mechanisms and targeted therapies of OSCC.
Design of a multi-Epitope mRNA vaccine against Brucella type IV secretion system using reverse vaccinology and immunogenicity approaches
MSC and HUVEC co-cultured fillers overcome intractable fistula in a new mouse model
Anastomotic leakage can lead to intractable fistulae after gastrointestinal surgery in patients with severe comorbidities. In this study, we aimed to devise new intractable fistula mouse models and evaluate the utility of the fillers containing human mesenchymal stem cells (MSCs) and human umbilical vein endothelial cells (HUVECs). After determining the optimal ratio of MSCs to HUVECs as fillers, we created new intractable fistula mouse models and verified the usefulness of the above-mentioned fillers for these fistulas. As the filler containing a 1:1 ratio of MSC: HUVEC showed the highest expression of FGF2 and VEGF among the organization-forming fillers, we determined that this was the optimal ratio. When this filler was transplanted into irradiated and steroid-treated mice with excisional wounds, the skin defects healed significantly faster in the filler-transplanted group than in the non-transplanted group (P < 0.05). Furthermore, we established a new mouse model of a gastrointestinal fistula by securing the cecum to the abdominal wall and puncturing the skin, abdominal wall, and intestinal wall with an indwelling needle. The fistula remained patent for at least seven days and was intractable. Unlike the adhesive group (group 1) (0/5) and the group implanted with fillers containing MSCs (group 2) (1/5), all fistulas were closed in the group implanted with fillers containing MSCs and HUVECs (group 3) (5/5). This study demonstrated that a treatment strategy using HUVEC is advantageous for treating intractable fistulae connected to the gastrointestinal tract. HUVEC should be included when fillers are used to close fistulas.