A system hierarchy for brain-inspired computing.

Neuromorphic computing draws inspiration from the brain to provide computing technology and architecture with the potential to drive the next wave of computer engineering1-13. Such brain-inspired computing also provides a promising platform for the development of artificial general intelligence14,15. However, unlike conventional computing systems, which have a well established computer hierarchy built around the concept of Turing completeness and the von Neumann architecture16-18, there is currently no generalized system hierarchy or understanding of completeness for brain-inspired computing. This affects the compatibility between software and hardware, impairing the programming flexibility and development productivity of brain-inspired computing. Here we propose 'neuromorphic completeness', which relaxes the requirement for hardware completeness, and a corresponding system hierarchy, which consists of a Turing-complete software-abstraction model and a versatile abstract neuromorphic architecture. Using this hierarchy, various programs can be described as uniform representations and transformed into the equivalent executable on any neuromorphic complete hardware-that is, it ensures programming-language portability, hardware completeness and compilation feasibility. We implement toolchain software to support the execution of different types of program on various typical hardware platforms, demonstrating the advantage of our system hierarchy, including a new system-design dimension introduced by the neuromorphic completeness. We expect that our study will enable efficient and compatible progress in all aspects of brain-inspired computing systems, facilitating the development of various applications, including artificial general intelligence.

Towards artificial general intelligence with hybrid Tianjic chip architecture.

There are two general approaches to developing artificial general intelligence (AGI)1: computer-science-oriented and neuroscience-oriented. Because of the fundamental differences in their formulations and coding schemes, these two approaches rely on distinct and incompatible platforms2-8, retarding the development of AGI. A general platform that could support the prevailing computer-science-based artificial neural networks as well as neuroscience-inspired models and algorithms is highly desirable. Here we present the Tianjic chip, which integrates the two approaches to provide a hybrid, synergistic platform. The Tianjic chip adopts a many-core architecture, reconfigurable building blocks and a streamlined dataflow with hybrid coding schemes, and can not only accommodate computer-science-based machine-learning algorithms, but also easily implement brain-inspired circuits and several coding schemes. Using just one chip, we demonstrate the simultaneous processing of versatile algorithms and models in an unmanned bicycle system, realizing real-time object detection, tracking, voice control, obstacle avoidance and balance control. Our study is expected to stimulate AGI development by paving the way to more generalized hardware platforms.

Early matrix softening contributes to vascular smooth muscle cell phenotype switching and aortic dissection through down-regulation of microRNA-143/145.

Thoracic aortic dissection (TAD) is characterized by extracellular matrix (ECM) dysregulation. Aberrations in the ECM stiffness can lead to changes in cellular functions. However, the mechanism by which ECM softening regulates vascular smooth muscle cell (VSMCs) phenotype switching remains unclear. To understand this mechanism, we cultured VSMCs in a soft extracellular matrix and discovered that the expression of microRNA (miR)-143/145, mediated by activation of the AKT signalling pathway, decreased significantly. Furthermore, overexpression of miR-143/145 reduced BAPN-induced aortic softening, switching the VSMC synthetic phenotype and the incidence of TAD in mice. Additionally, high-throughput sequencing of immunoprecipitated RNA indicated that the TEA domain transcription factor 1 (TEAD1) is a common target gene of miR-143/145, which was subsequently verified using a luciferase reporter assay. TEAD1 is upregulated in soft ECM hydrogels in vitro, whereas the switch to a synthetic phenotype in VSMCs decreases after TEAD1 knockdown. Finally, we verified that miR-143/145 levels are associated with disease severity and prognosis in patients with thoracic aortic dissection. ECM softening, as a result of promoting the VSMCs switch to a synthetic phenotype by downregulating miR-143/145, is an early trigger of TAD and provides a therapeutic target for this fatal disease. miR-143/145 plays a role in the early detection of aortic dissection and its severity and prognosis, which can offer information for future risk stratification of patients with dissection.

Impact of Lysine Succinylation on the Biology of Fungi.

Post-translational modifications (PTMs) play a crucial role in protein functionality and the control of various cellular processes and secondary metabolites (SMs) in fungi. Lysine succinylation (Ksuc) is an emerging protein PTM characterized by the addition of a succinyl group to a lysine residue, which induces substantial alteration in the chemical and structural properties of the affected protein. This chemical alteration is reversible, dynamic in nature, and evolutionarily conserved. Recent investigations of numerous proteins that undergo significant succinylation have underscored the potential significance of Ksuc in various biological processes, encompassing normal physiological functions and the development of certain pathological processes and metabolites. This review aims to elucidate the molecular mechanisms underlying Ksuc and its diverse functions in fungi. Both conventional investigation techniques and predictive tools for identifying Ksuc sites were also considered. A more profound comprehension of Ksuc and its impact on the biology of fungi have the potential to unveil new insights into post-translational modification and may pave the way for innovative approaches that can be applied across various clinical contexts in the management of mycotoxins.

Computational drug repositioning based on the relationships between substructure-indication.

At present, computational methods for drug repositioning are mainly based on the whole structures of drugs, which limits the discovery of new functions due to the similarities between local structures of drugs. In this article, we, for the first time, integrated the features of chemical-genomics (substructure-domain) and pharmaco-genomics (domain-indication) based on the assumption that drug-target interactions are mediated by the substructures of drugs and the domains of proteins to identify the relationships between substructure-indication and establish a drug-substructure-indication network for predicting all therapeutic effects of tested drugs through only information on the substructures of drugs. In total, 83 205 drug-indication relationships with different correlation scores were obtained. We used three different verification methods to indicate the accuracy of the method and the reliability of the scoring system. We predicted all indications of olaparib using our method, including the known antitumor effect and unknown antiviral effect verified by literature, and we also discovered the inhibitory mechanism of olaparib toward DNA repair through its specific sub494 (o = C-C: C), as it participates in the low synthesis of the poly subfunction of the apoptosis pathway (hsa04210) by inhibiting the Inositol 1,4,5-trisphosphate receptor(s) (ITPRs) and hydrolyzing poly (ADP ribose) polymerases. ElectroCardioGrams of four drugs (quinidine, amiodarone, milrinone and fosinopril) demonstrated the effect of anti-arrhythmia. Unlike previous studies focusing on the overall structures of drugs, our research has great potential in the search for more therapeutic effects of drugs and in predicting all potential effects and mechanisms of a drug from the local structural similarity.

Quantification of Early Diffuse Myocardial Fibrosis Through 7.0 T Cardiac Magnetic Resonance T1 Mapping in a Type 1 Diabetic Mellitus Mouse Model.

BACKGROUND: Diffuse myocardial interstitial fibrosis (DMIF) is a key factor for heart failure (HF) in diabetic cardiomyopathy. MRI T1-mapping technique can quantitatively evaluate DMIF. PURPOSE: To evaluate of early DMIF in a type 1 diabetes mellitus (T1DM) mouse model through 7.0 T MRI T1 mapping. STUDY TYPE: Prospective. ANIMAL MODEL: A total of 50 8-week-old C57Bl/6J male mice were divided into control (n = 20) and T1DM (n = 30) groups. FIELD STRENGTH/SEQUENCE: A 7.0 T small animal MRI; gradient echo Look-Locker inversion recovery T1-mapping sequence; cine MRI. Scans were acquired in control and T1DM mice every 4 weeks until 24 weeks. ASSESSMENT: End-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (EF), left ventricle (LV) mass, fractional shortening (FS), and E/A ratio. They were evaluated through echocardiography and cine MRI. The extracellular volume fraction (ECV) was calculated. Sirius Red staining was performed and calculated collagen volume fraction (CVF). STATISTICAL TESTS: Differences in ECV and CVF between two groups were analyzed using one-way analysis of variance. The correlation between ECV and CVF was assessed using Pearson's correlations. RESULTS: Compared with the control group, a progressive decrease in FS, EF, and E/A ratio was observed in the T1DM group. Both ECV and CVF values gradually increased during diabetes progression. A significant increase in ECV and CVF values was observed at 12 weeks (ECV: 32.5% ± 1.6% vs. 28.1% ± 1.8%; CVF: 6.9% ± 1.8% vs. 3.3% ± 1.1%). ECV showed a strong correlation with CVF (r = 0.856). DATA CONCLUSION: ECV is an accurate and feasible imaging marker that can be used to quantitatively assess DMIF changes over time in T1DM mice. ECV has potential to accurately detect DMIF in the early stage and may be a useful imaging tool to assess the need for early intervention in T1DM mice. EVIDENCE LEVEL: 1 TECHNICAL EFFICACY: Stage 3.

Information evolution in complex networks.

Many biological phenomena or social events critically depend on how information evolves in complex networks. However, a general theory to characterize information evolution is yet absent. Consequently, numerous unknowns remain about the mechanisms underlying information evolution. Among these unknowns, a fundamental problem, being a seeming paradox, lies in the coexistence of local randomness, manifested as the stochastic distortion of information content during individual-individual diffusion, and global regularity, illustrated by specific non-random patterns of information content on the network scale. Here, we attempt to formalize information evolution and explain the coexistence of randomness and regularity in complex networks. Applying network dynamics and information theory, we discover that a certain amount of information, determined by the selectivity of networks to the input information, frequently survives from random distortion. Other information will inevitably experience distortion or dissipation, whose speeds are shaped by the diversity of information selectivity in networks. The discovered laws exist irrespective of noise, but noise accounts for disturbing them. We further demonstrate the ubiquity of our discovered laws by analyzing the emergence of neural tuning properties in the primary visual and medial temporal cortices of animal brains and the emergence of extreme opinions in social networks.

An oriented-collagen scaffold including Wnt5a promotes osteochondral regeneration and cartilage interface integration in a rabbit model.

Articular cartilage regeneration remains a major challenge in orthopedics. Noncanonical Wnt5a is a particularly attractive growth factor in this context; Wnt5a inhibits chondrocyte hypertrophy but maintains chondrogenesis. We designed a novel, vertically oriented-collagen scaffold. The effect of Wnt5a on MSCs and chondrocytes and the therapeutic effects of the Wnt5a/oriented-collagen scaffold in terms of osteochondral repair and cartilage integration were evaluated. In vitro, the proliferation, migration, and differentiation of MSCs and chondrocytes treated with Wnt5a, and the mechanisms thereof, were assessed. mRNA microarray analysis was performed to compare the expression profiles of MSCs before and after Wnt5a treatment. In vivo, full-thickness cylindrical osteochondral defects (4 mm in diameter, 3 mm in depth) were created in the patellar grooves of 24 New Zealand white rabbits and implanted with oriented-collagen scaffolds (n = 8), Wnt5a/oriented-collagen scaffolds (n = 8), or nothing (n = 8). After 6 and 12 weeks, integration and tissue responses were evaluated. The proliferation, migration, chondrogenic differentiation, and extracellular matrix formation of/by MSCs and chondrocytes improved greatly after treatment with Wnt5a. Western blotting showed that the PI3K/AKT/JNK signaling pathway was activated. Microarray analysis revealed that the Wnt5a group exhibited a significant upregulation of the PI3K pathway. Reactome GSEA pathway interaction analysis revealed that such upregulation was associated with collagen and extracellular matrix formation. In vivo, the Wnt5a/oriented-collagen scaffold group exhibited optimal interface integration, cartilage regeneration, and collagenous fiber arrangement, accompanied by significantly increased glycosaminoglycan and collagen accumulations in the zones of regeneration and integration, compared to the other groups. Gene expression analysis showed that the levels of mRNAs encoding genes involved in cartilage formation were significantly increased in the Wnt5a/oriented, collagen scaffold group (all P < .05). Wnt5a promoted the proliferation, migration, and chondrogenic differentiation of MSCs and chondrocytes via the activation of the PI3K/AKT/JNK signaling pathway. The Wnt5a/oriented-collagen constructs enhanced the structure-specific regeneration of hyaline cartilage in a rabbit model and may be a promising treatment for the repair of human cartilage defects.

Reducing exposure to COVID-19 by improving access to fever clinics: an empirical research of the Shenzhen area of China.

BACKGROUND: The current 2019 coronavirus disease (COVID-19) pandemic is hitting citizen's life and health like never before, with its significant loss to human life and a huge economic toll. In this case, the fever clinics (FCs) were still preserved as one of the most effective control measures in China, but this work is based on experience and lacks scientific and effective guidance. Here, we use travel time to link facilities and populations at risk of COVID-19 and identify the dynamic allocation of patients' medical needs, and then propose the optimized allocation scheme of FCs. METHODS: We selected Shenzhen, China, to collect geospatial resources of epidemic communities (ECs) and FCs to determine the ECs' cumulative opportunities of visiting FCs, as well as evaluate the rationality of medical resources in current ECs. Also, we use the Location Set Covering Problem (LSCP) model to optimize the allocation of FCs and evaluate efficiency. RESULTS: Firstly, we divide the current ECs into 3 groups based on travel time and cumulative opportunities of visiting FCs within 30 min: Low-need communities (22.06%), medium-need communities (59.8%), and high-need communities (18.14%) with 0,1-2 and no less than 3 opportunities of visiting FCs. Besides, our work proposes two allocation schemes of fever clinics through the LSCP model. Among which, selecting secondary and above hospitals as an alternative in Scheme 1, will increase the coverage rate of hospitals in medium-need and high-need communities from 59.8% to 80.88%. In Scheme 2, selecting primary and above hospitals as an alternative will increase the coverage rate of hospitals in medium-need and high-need communities to 85.29%, with the average travel time reducing from 22.42 min to 17.94 min. CONCLUSIONS: The optimized allocation scheme can achieve two objectives: a. equal access to medical services for different types of communities has improved while reducing the overutilization of high-quality medical resources. b. the travel time for medical treatment in the community has reduced, thus improving medical accessibility. On this basis, during the early screening in prevention and control of the outbreak, the specific suggestions for implementation in developing and less developed countries are made.

A role for differential variable gene pairing in creating T cell receptors specific for unique major histocompatibility ligands.

A limited set of T cell receptor (TCR) variable (V) gene segments are used to create a repertoire of TCRs that recognize all major histocompatibility complex (MHC) ligands within a species. How individual αßTCRs are constructed to specifically recognize a limited set of MHC ligands is unclear. Here we have identified a role for the differential pairing of particular V gene segments in creating TCRs that recognized MHC class II ligands exclusively, or cross-reacted with classical and nonclassical MHC class I ligands. Biophysical and structural experiments indicated that TCR specificity for MHC ligands is not driven by germline-encoded pairwise interactions.Rather, identical TCRß chains can have altered peptide-MHC (pMHC) binding modes when paired with different TCRα chains. The ability of TCR chain pairing to modify how V region residues interact with pMHC helps to explain how the same V genes are used to create TCRs specific for unique MHC ligands.

Bias in social interactions and emergence of extremism in complex social networks.

Emergence of extremism in social networks is among the most appealing topics of opinion dynamics in computational sociophysics in recent decades. Most of the existing studies presume that the initial existence of certain groups of opinion extremities and the intrinsic stubbornness in individuals' characteristics are the key factors allowing the tenacity or even prevalence of such extreme opinions. We propose a modification to the consensus making in bounded-confidence models where two interacting individuals holding not so different opinions tend to reach a consensus by adopting an intermediate opinion of their previous ones. We show that if individuals make biased compromises, extremism may still arise without a need of an explicit classification of extremists and their associated characteristics. With such biased consensus making, several clusters of diversified opinions are gradually formed up in a general trend of shifting toward the extreme opinions close to the two ends of the opinion range, which may allow extremism communities to emerge and moderate views to be dwindled. Furthermore, we assume stronger compromise bias near opinion extremes. It is found that such a case allows moderate opinions a greater chance to survive compared to that of the case where the bias extent is universal across the opinion space. As to the extreme opinion holders' lower tolerances toward different opinions, which arguably may exist in many real-life social systems, they significantly decrease the size of extreme opinion communities rather than helping them to prevail. Brief discussions are presented on the significance and implications of these observations in real-life social systems.

Mass spectrometric determination of disulfide bonds and free cysteine in grass carp IgM isoforms.

Disulfide bonds are fundamental in establishing Ig structure and maintaining Ig biological function. Here, we analysed disulfide bonds and free cysteine in three grass carp IgM isoforms (monomeric, dimeric/trimeric, and tetrameric IgM) by liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS). The results revealed that Cys574 residue status at the C-terminal tail differed substantially in monomeric IgM in comparison with polymeric IgM, Cys574 was found as free thiol in monomeric IgM, while it formed disulfide linkages in dimeric/trimeric and tetrameric IgM. Five intra-chain disulfide bonds in the CH1~CH4 and CL1 domains, as well as one H-H and one H-L inter-chain disulfide linkages, were also observed and shown identical connectivity in monomeric, dimeric/trimeric, and tetrameric IgM. These findings represent the first experimental assignments of disulfide linkages of grass carp IgM and reveal that grass carp IgM isoform formation is due to alternative disulfide bonds connecting the Cys574 residue at the C-terminal tail.

Highly Compact Artificial Memristive Neuron with Low Energy Consumption.

Neuromorphic systems aim to implement large-scale artificial neural network on hardware to ultimately realize human-level intelligence. The recent development of nonsilicon nanodevices has opened the huge potential of full memristive neural networks (FMNN), consisting of memristive neurons and synapses, for neuromorphic applications. Unlike the widely reported memristive synapses, the development of artificial neurons on memristive devices has less progress. Sophisticated neural dynamics is the major obstacle behind the lagging. Here a rich dynamics-driven artificial neuron is demonstrated, which successfully emulates partial essential neural features of neural processing, including leaky integration, automatic threshold-driven fire, and self-recovery, in a unified manner. The realization of bioplausible artificial neurons on a single device with ultralow power consumption paves the way for constructing energy-efficient large-scale FMNN and may boost the development of neuromorphic systems with high density, low power, and fast speed.

Effect of CDR3 sequences and distal V gene residues in regulating TCR-MHC contacts and ligand specificity.

The mature T cell repertoire has the ability to orchestrate immunity to a wide range of potential pathogen challenges. This ability stems from thymic development producing individual T cell clonotypes that express TCRs with unique patterns of Ag reactivity. The Ag specificity of TCRs is created from the combinatorial pairing of one of a set of germline encoded TCR Vα and Vß gene segments with randomly created CDR3 sequences. How the amalgamation of germline encoded and randomly created TCR sequences results in Ag receptors with unique patterns of ligand specificity is not fully understood. Using cellular, biophysical, and structural analyses, we show that CDR3α residues can modulate the geometry in which TCRs bind peptide-MHC (pMHC), governing whether and how germline encoded TCR Vα and Vß residues interact with MHC. In addition, a CDR1α residue that is positioned distal to the TCR-pMHC binding interface is shown to contribute to the peptide specificity of T cells. These findings demonstrate that the specificity of individual T cell clonotypes arises not only from TCR residues that create direct contacts with the pMHC, but also from a collection of indirect effects that modulate how TCR residues are used to bind pMHC.

Transport of glial cell line-derived neurotrophic factor into liposomes across the blood-brain barrier: in vitro and in vivo studies.

Glial cell line-derived neurotrophic factor (GDNF) was encapsulated into liposomes in order to protect it from enzyme degradation in vivo and promote its permeability across the blood-brain barrier (BBB). In this study, GDNF conventional liposomes (GDNF-L) and GDNF target sterically stabilized liposomes (GDNF-SSL-T) were prepared. The average size of liposomes was below 90 nm. A primary model of BBB was established and evaluated by transendothelial electrical resistance (TEER) and permeability. This BBB model was employed to study the permeability of GDNF liposomes in vitro. The results indicated that the liposomes could enhance transport of GDNF across the BBB and GDNF-SSL-T had achieved the best transport efficacy. The distribution of GDNF liposomes was studied in vivo. Free GDNF and GDNF-L were eliminated rapidly in the circulation. GDNF-SSL-T has a prolonged circulation time in the blood and favorable brain delivery. The values of the area under the curve (AUC(0-1 h)) in the brain of GDNF-SSL-T was 8.1 times and 6.8 times more than that of free GDNF and GDNF-L, respectively. These results showed that GDNF-SSL-T realized the aim of targeted delivery of therapeutic proteins to central nervous system.

Antinociceptive effect of intrathecal microencapsulated human pheochromocytoma cell in a rat model of bone cancer pain.

Human pheochromocytoma cells, which are demonstrated to contain and release met-enkephalin and norepinephrine, may be a promising resource for cell therapy in cancer-induced intractable pain. Intrathecal injection of alginate-poly (l) lysine-alginate (APA) microencapsulated human pheochromocytoma cells leads to antinociceptive effect in a rat model of bone cancer pain, and this effect was blocked by opioid antagonist naloxone and alpha 2-adrenergic antagonist rauwolscine. Neurochemical changes of cerebrospinal fluid are in accordance with the analgesic responses. Taken together, these data support that human pheochromocytoma cell implant-induced antinociception was mediated by met-enkephalin and norepinephrine secreted from the cell implants and acting at spinal receptors. Spinal implantation of microencapsulated human pheochromocytoma cells may provide an alternative approach for the therapy of chronic intractable pain.

[Prediction models for the 15 years risk of new-onset hypertension in Chinese people aged from 35 to 64 years old].

OBJECTIVE: To set up prediction models for the risk of new-onset hypertension in Chinese people and explore the risk scores to facilitate the clinical application. METHODS: A cohort set up since 1992 with participants aged 35-64 years old from 11 provinces and cities of China was prospectively studied. Logistic regression was used to analyze the risk factors for the incidence of hypertension within 15 years and the prediction models and risk scores were developed with the regression coefficient. The performance of the prediction models were tested and compared with the Framingham model. RESULTS: A total of 3 899 participants free from hypertension at baseline with 15 years follow-up were enrolled in the study. Within 15 years, 1 776 cases of incident hypertension were ascertained with a incidence rate of 45.6%. Two prediction models were set up with age, systolic blood pressure, diastolic blood pressure, BMI and the history of parental hypertension in the Model 1, while TG and HDL-C added on the basis of Model 1 in the Model 2. Good performance of discrimination and calibration was established in both models with significant difference in C statistics and no significant difference in net reclassification improvement (NRI) index. CONCLUSION: The hypertension risk prediction models can be used to estimate an individual's absolute risk for hypertension and could facilitate the management of potential hypertension patients.

Chitin Deacetylase Homologous Gene cda Contributes to Development and Aflatoxin Synthesis in Aspergillus flavus.

The fungal cell wall serves as the primary interface between fungi and their external environment, providing protection and facilitating interactions with the surroundings. Chitin is a vital structural element in fungal cell wall. Chitin deacetylase (CDA) can transform chitin into chitosan through deacetylation, providing various biological functions across fungal species. Although this modification is widespread in fungi, the biological functions of CDA enzymes in Aspergillus flavus remain largely unexplored. In this study, we aimed to investigate the biofunctions of the CDA family in A. flavus. The A. flavus genome contains six annotated putative chitin deacetylases. We constructed knockout strains targeting each member of the CDA family, including Δcda1, Δcda2, Δcda3, Δcda4, Δcda5, and Δcda6. Functional analyses revealed that the deletion of CDA family members neither significantly affects the chitin content nor exhibits the expected chitin deacetylation function in A. flavus. However, the Δcda6 strain displayed distinct phenotypic characteristics compared to the wild-type (WT), including an increased conidia count, decreased mycelium production, heightened aflatoxin production, and impaired seed colonization. Subcellular localization experiments indicated the cellular localization of CDA6 protein within the cell wall of A. flavus filaments. Moreover, our findings highlight the significance of the CBD1 and CBD2 structural domains in mediating the functional role of the CDA6 protein. Overall, we analyzed the gene functions of CDA family in A. flavus, which contribute to a deeper understanding of the mechanisms underlying aflatoxin contamination and lay the groundwork for potential biocontrol strategies targeting A. flavus.

Development of a novel lncRNA-derived immune gene score using machine learning-based ensembles for predicting the survival of HCC.

BACKGROUND: Long noncoding RNAs (lncRNAs) are implicated in the tumor immunology of hepatocellular carcinoma (HCC). METHODS: HCC mRNA and lncRNA expression profiles were used to extract immune-related genes with the ImmPort database, and immune-related lncRNAs with the ImmLnc algorithm. The MOVICS package was used to cluster immune-related mRNA, immune-related lncRNA, gene mutation and methylation data on HCC from the TCGA. GEO and ICGC datasets were used to validate the model. Data from single-cell sequencing was used to determine the expression of genes from the model in various immune cell types. RESULTS: With this model, the area under the curve (AUC) for 1-, 3- and 5-year survival of HCC patients was 0.862, 0.869 and 0.912, respectively. Single-cell sequencing showed EREG was significantly expressed in a variety of immune cell types. Knockdown of the EREG target gene resulted in significant anti-apoptosis, pro-proliferation and pro-migration effects in HepG2 and HUH7 cells. Moreover, serum and liver tissue EREG levels in HCC patients were significantly higher than those of healthy control patients. CONCLUSION: We built a prognostic model with good accuracy for predicting HCC patient survival. EREG is a potential immunotherapeutic target and a promising prognostic biomarker.

Multimodal fall detection for solitary individuals based on audio-video decision fusion processing.

Falls often pose significant safety risks to solitary individuals, especially the elderly. Implementing a fast and efficient fall detection system is an effective strategy to address this hidden danger. We propose a multimodal method based on audio and video. On the basis of using non-intrusive equipment, it reduces to a certain extent the false negative situation that the most commonly used video-based methods may face due to insufficient lighting conditions, exceeding the monitoring range, etc. Therefore, in the foreseeable future, methods based on audio and video fusion are expected to become the best solution for fall detection. Specifically, this article outlines the following methodology: the video-based model utilizes YOLOv7-Pose to extract key skeleton joints, which are then fed into a two stream Spatial Temporal Graph Convolutional Network (ST-GCN) for classification. Meanwhile, the audio-based model employs log-scaled mel spectrograms to capture different features, which are processed through the MobileNetV2 architecture for detection. The final decision fusion of the two results is achieved through linear weighting and Dempster-Shafer (D-S) theory. After evaluation, our multimodal fall detection method significantly outperforms the single modality method, especially the evaluation metric sensitivity increased from 81.67% in single video modality to 96.67% (linear weighting) and 97.50% (D-S theory), which emphasizing the effectiveness of integrating video and audio data to achieve more powerful and reliable fall detection in complex and diverse daily life environments.