Search

Deep learning (DL) has substantially enhanced natural language processing (NLP) in healthcare research. However, the increasing complexity of DL-based NLP necessitates transparent model interpretability, or at least explainability, for reliable decision-making. This work presents a thorough scoping review of explainable and interpretable DL in healthcare NLP. The term “eXplainable and Interpretable Artificial Intelligence” (XIAI) is introduced to distinguish XAI from IAI. Different models are further categorized based on their functionality (model-, input-, output-based) and scope (local, global). Our analysis shows that attention mechanisms are the most prevalent emerging IAI technique. The use of IAI is growing, distinguishing it from XAI. The major challenges identified are that most XIAI does not explore “global” modelling processes, the lack of best practices, and the lack of systematic evaluation and benchmarks. One important opportunity is to use attention mechanisms to enhance multi-modal XIAI for personalized medicine. Additionally, combining DL with causal logic holds promise. Our discussion encourages the integration of XIAI in Large Language Models (LLMs) and domain-specific smaller models. In conclusion, XIAI adoption in healthcare requires dedicated in-house expertise. Collaboration with domain experts, end-users, and policymakers can lead to ready-to-use XIAI methods across NLP and medical tasks. While challenges exist, XIAI techniques offer a valuable foundation for interpretable NLP algorithms in healthcare.

Abstract Background Reducing production costs while producing high-quality livestock and poultry products is an ongoing concern in the livestock industry. The addition of oil to livestock and poultry diets can enhance feed palatability and improve growth performance. Emulsifiers can be used as potential feed supplements to improve dietary energy utilization and maintain the efficient productivity of broilers. Therefore, further investigation is warranted to evaluate whether dietary emulsifier supplementation can improve the efficiency of fat utilization in the diet of yellow-feathered broilers. In the present study, the effects of adding emulsifier to the diet on lipid metabolism and the performance of yellow-feathered broilers were tested. A total of 240 yellow-feasted broilers (21-day-old) were randomly divided into 4 groups (6 replicates per group, 10 broilers per replicate, half male and half female within each replicate). The groups were as follows: the control group (fed with basal diet), the group fed with basal diet supplemented with 500 mg/kg emulsifier, the group fed with a reduced oil diet (reduced by 1%) supplemented with 500 mg/kg emulsifier, and the group fed with a reduced oil diet supplemented with 500 mg/kg emulsifier. The trial lasted for 42 days, during which the average daily feed intake, average daily gain, and feed-to-gain ratio were measured. Additionally, the expression levels of lipid metabolism-related genes in the liver, abdominal fat and each intestinal segment were assessed. Results The results showed that compared with the basal diet group, (1) The average daily gain of the basal diet + 500 mg/kg emulsifier group significantly increased (P

This study explores the mechanism by which obstructive jaundice (OJ) induces liver damage through pyroptosis. We induced OJ in rats via bile duct ligation and assessed liver damage using serum biochemical markers and histological analysis of liver tissue. Pyroptosis was investigated through immunofluorescence, ELISA, Western blot, and quantitative RT-PCR techniques. Additionally, we examined intestinal function and fecal microbiota alterations in the rats using 16S rDNA sequencing. In vitro experiments involved co-culturing Kupffer cells and hepatocytes, which were then exposed to bile and lipopolysaccharide (LPS). Our findings indicated that OJ modified the gut microbiota, increasing LPS levels, which, in conjunction with bile, initiated a cycle of inflammation, fibrosis, and cell death in the liver. Mechanistically, OJ elevated necrotic markers such as ATP, which in turn activated pyroptotic pathways. Increased levels of pyroptosis-related molecules, including NLRP3, caspase-1, gasdermin D, and IL-18, were confirmed. In our co-cultured cell model, bile exposure resulted in cell death and ATP release, leading to the activation of the NLRP3 inflammasome and its downstream effectors, caspase-1 and IL-18. The combination of bile and LPS significantly intensified pyroptotic responses. This study is the first to demonstrate that LPS and bile synergistically exacerbate liver injury by promoting necrosis and pyroptosis, unveiling a novel mechanism of OJ-associated hepatic damage and suggesting avenues for potential preventive or therapeutic interventions.

To overcome the shortcoming wherein the accuracy of subsea cable detection can be affected by the determination of the bias vector, scale factors, and non-orthogonality corrections of the vector magnetometer, a real-time attitude-independent route tracking method for subsea power cables is investigated theoretically and experimentally by means of scalar magnetic field checking. The measurement of the magnetic field Bc produced by the current in a cable is made immune to the influence of the platform attitude by extracting the component of Bc along the geomagnetic field using a high-bandwidth self-oscillating optically pumped magnetometer. The self-oscillating frequency is proved to be independent of the attitude of the magnetometer with the theoretical model. Experiments are carried out to test the attitude-independent performance, and the effectiveness of route tracking is verified by the results of the sea experiment. The proposed method will effectively improve the ability to locate subsea cables under high sea conditions.

Introduction: More than 50 mutations in the MAPT gene result in heterogeneous forms of frontotemporal lobar dementia with tau inclusions (FTLD-Tau). However, early pathogenic events that lead to disease and the degree to which they are common across MAPT mutations remain poorly understood. The goal of this study is to determine whether there is a common molecular signature of FTLD-Tau.Methods: We analyzed genes differentially expressed in induced pluripotent stem cell–derived neurons (iPSC-neurons) that represent the three major categories of MAPT mutations: splicing (IVS10 + 16), exon 10 (p.P301L), and C-terminal (p.R406W) compared with isogenic controls. The genes that were commonly differentially expressed in MAPT IVS10 + 16, p.P301L, and p.R406W neurons were enriched in trans-synaptic signaling, neuronal processes, and lysosomal function. Many of these pathways are sensitive to disruptions in calcium homeostasis. One gene, CALB1, was significantly reduced across the three MAPT mutant iPSC-neurons and in a mouse model of tau accumulation. We observed a significant reduction in calcium levels in MAPT mutant neurons compared with isogenic controls, pointing to a functional consequence of this disrupted gene expression. Finally, a subset of genes commonly differentially expressed across MAPT mutations were also dysregulated in brains from MAPT mutation carriers and to a lesser extent in brains from sporadic Alzheimer disease and progressive supranuclear palsy, suggesting that molecular signatures relevant to genetic and sporadic forms of tauopathy are captured in a dish. The results from this study demonstrate that iPSC-neurons capture molecular processes that occur in human brains and can be used to pinpoint common molecular pathways involving synaptic and lysosomal function and neuronal development, which may be regulated by disruptions in calcium homeostasis.

Magnetic target localization using the magnetic gradient tensor (MGT) plays a significant role in underwater localization. However, this method inherently has a localization dead zone, which presents challenges for real-world applications. This paper delves into the root cause of this dead zone, identifying the non-invertibility of the MGT when the magnetic moment vector is orthogonal to the position vector from the target to the observation point. To tackle this issue, a method based on the eigenvector constraints is proposed. By constructing an objective function with eigenvector constraints and leveraging the property that its gradient at the observation point is zero, we derive an equivalent expression for the inverse of MGT that always holds and further develop a dead-zone-free localization method. To validate the robustness and efficacy of the proposed localization method, a comparative analysis with other methods is conducted. Simulation results in a 10 m × 10 m area under Gaussian noise demonstrate the proposed method’s capability to eliminate the dead zone and achieve an average localization error of 0.032 m. Experimental results further demonstrate that the proposed method eliminates the localization dead zone and exhibits greater robustness than the dominant method in the normal region. In summation, this paper provides an effective method for eliminating localization dead zone, offering a more stable and reliable method for magnetic target localization in practice.

Abstract Background Recent studies have highlighted the important role of long non-coding RNA SNHG16 in various human cancers. Here, we conducted a meta-analysis to investigate the effect of SNHG16 expression on clinicopathological features and prognosis in patients with different kinds of human cancers. Methods We performed a systematic search in electronic databases including PubMed, EMBASE, Cochrane Library and Web of Science, to investigate the potential association between SNHG16 expression and prognostic significance and clinical features in cancer patients. Odds ratios (ORs) or hazards ratios (HRs) with corresponding 95% confidence intervals (95% CIs) were pooled to estimate the prognosis value of SNHG16 by StataSE 15.0 software. Results A total of 16 eligible studies with 1299 patients were enrolled in our meta-analysis. The results revealed that increased expression level of SNHG16 was significantly associated with larger tumor size (OR: 3.357; 95% CI: 2.173–5.185; P

Pulse timing is an important topic in nuclear instrumentation, with far-reaching applications from high energy physics to radiation imaging. While high-speed analog-to-digital converters become more and more developed and accessible, their potential uses and merits in nuclear detector signal processing are still uncertain, partially due to associated timing algorithms which are not fully understood and utilized. In this paper, we propose a novel method based on deep learning for timing analysis of modularized detectors without explicit needs of labeling event data. By taking advantage of the intrinsic time correlations, a label-free loss function with a specially designed regularizer is formed to supervise the training of neural networks (NNs) towards a meaningful and accurate mapping function. We mathematically demonstrate the existence of the optimal function desired by the method, and give a systematic algorithm for training and calibration of the model. The proposed method is validated on two experimental datasets based on silicon photomultipliers as main transducers. In the toy experiment, the NN model achieves the single-channel time resolution of 8.8 ps and exhibits robustness against concept drift in the dataset. In the electromagnetic calorimeter experiment, several NN models (fully-connected, convolutional neural network and long short-term memory) are tested to show their conformance to the underlying physical constraint and to judge their performance against traditional methods. In total, the proposed method works well in either ideal or noisy experimental condition and recovers the time information from waveform samples successfully and precisely.

Tetrathiomolybdate (TM) and Cu-ATPases, e.g. Wilson (WLN) protein, affect the efficacy of common anticancer drug cisplatin. Here, the authors show that TM generates a protein dimer with a WLN domain by expelling copper and provide insight into the synergy of TM and cisplatin in cancer chemotherapy.

Abstract Background Multi-drug resistance (MDR) is one of the main obstacles for treatment of advanced/recurrent hepatocellular carcinoma (HCC). We have previously identified arsenic trioxide (ATO) as an effective metastasis/angiogenesis inhibitor in HCC. Here, we further found that MDR-HCC cells were more sensitive to ATO. Methods The MDR-HCC cells were used as experimental models. Biological functions were investigated using cell transfection, polymerase chain reaction, western blot, southwestern blot, immunostaining, immunoprecipitation plus atomic fluorescence spectrometry, and so on. Results The MDR-HCC cells underwent high oxidative stress condition, and employed adaptive mechanisms for them to survive; while ATO abolished such mechanisms via targeting the 14–3-3η/nuclear factor kappa B (NF-κB) feedback Loop. Briefly, in MDR cells, the increase of ROS activated NF-κB signaling, which transcriptionally activated 14–3-3η. Meanwhile, the activation of NF-κB can be constitutively maintained by 14–3-3η. As a NF-κB inhibitor, ATO transcriptionally inhibited the 14–3-3η mRNA level. Meanwhile, ATO was also validated to directly bind to 14–3-3η, enhancing the degradation of 14–3-3η protein in an ubiquitination-dependent manner. Knockdown of 14–3-3η reduced the ATO-induced reversal extents of drug resistance in MDR cells. Conclusion 14–3-3η/NF-κB feedback loop plays an important role in maintaining the MDR phenotype in HCC. Moreover, via targeting such feedback loop, ATO could be considered as a potential molecular targeted agent for the treatment of HCC.

Background: Cap-assisted endoscopic sclerotherapy is a new interventional therapy for internal hemorrhoids and rectal prolapse under colonoscopy. The proper length of the endoscopic injection needle is the core for performing cap-assisted endoscopic sclerotherapy well with more benefits and less complications. However, no data are currently available to guide endoscopists to consider the length of injection needle before cap-assisted endoscopic sclerotherapy. This study is designed to evaluate the efficacy and safety of cap-assisted endoscopic sclerotherapy with long or short injection needle in the treatment of internal hemorrhoids. Methods: This is a nationwide multi-center, prospective, single-blind and randomized controlled trial. Patients with grade I-II internal hemorrhoids who have failed to conservative treatments and grade III internal hemorrhoids who are not suitable for surgery or refuse surgery will be included. Participants will be randomized 1:1 into either long or short injection needle group. The primary outcome is the recurrence rate of internal hemorrhoids 24 weeks after cap-assisted endoscopic sclerotherapy. The secondary outcomes are as follows: (1) symptom severity score, (2) three-level EuroQoL five dimensions health scale scores, (3) occurrence of adverse events and severe adverse events, and (4) patients’ attitudes toward cap-assisted endoscopic sclerotherapy. Data collection will be conducted before and during operation, the 1st day, 1st week, 2nd week, and 24th week after cap-assisted endoscopic sclerotherapy. Discussion: The outcome of this study is expected to provide a practical clinical protocol of cap-assisted endoscopic sclerotherapy for patients with internal hemorrhoids and promote the use of this new endoscopic technique. Trial registration: ClinicalTrials.gov, NCT03917056. Registered on 12 April 2019.

Summary: Cannabinoids are reported to rescue cocaine-induced seizures (CISs), a severe complication in cocaine users. However, the molecular targets for cannabinoid therapy of CISs remain unclear. Here, we report that the systemic administration of cannabinoids alleviates CISs in a CB1/CB2-receptor-independent manner. In HEK293 cells and cortical neurons, cocaine-induced dysfunction of the glycine receptor (GlyR) is restored by cannabinoids. Such restoration is blocked by GlyRα1S296A mutation. Consistently, the therapeutic effects of cannabinoids on CISs are also eliminated in GlyRα1S296A mutant mice. Based on molecular dynamic simulation, the hydrogen-bonding interaction between cocaine and the GlyR is weakened by cannabinoid docking. Without altering cocaine distribution across the brain, cannabinoids significantly suppress cocaine-exaggerated neuronal excitability in the prefrontal cortex (PFC) and hippocampus by rehabilitating extra-synaptic GlyR function. Microinjection of cannabinoids into the PFC and hippocampus restores cocaine-puzzled neural activity and alleviates CISs. These findings suggest that using GlyR-hypersensitive cannabinoids may represent a potential therapeutic strategy for treating CISs. : Zou et al. identify the glycine receptor as a potential therapeutic target for cannabinoids in treating cocaine-induced seizures. The function of extra-synaptic glycine receptors in the prefrontal cortex and hippocampus is impaired by cocaine and rescued by cannabinoids, therefore leading to the therapeutic effects of cannabinoids in treating cocaine-induced seizures. Keywords: cocaine, seizure, cannabinoids, glycine receptor, extra-synaptic, prefrontal cortex, hippocampus

Abstract Background This study aimed to evaluate the feasibility, safety, and value of a quick technique for transendoscopic enteral tubing (TET) through mid-gut. Methods A prospective interventional study was performed in a single center. A TET tube was inserted into mid-gut through the nasal orifice and fixed on the pylorus wall by one tiny titanium endoscopic clip under anesthesia. The feasibility, safety, success rate, and satisfaction with TET placement were evaluated for enteral nutrition or fecal microbiota transplantation. Results A total of 86 patients underwent mid-gut TET. The success rate of the TET procedure was 98.8% (85/86). Mean tubing time of the TET procedure was 4.2 ± 1.9 min. 10 cases of procedure was enough for training of general endoscopist to shorten the procedure time (7.0 min vs 4.0 min, p

Abstract The ancient Chinese medical literature, as well as our prior clinical experience, suggests that fecal microbiota transplantation (FMT) could treat the inflammatory mass. We aimed to evaluate the efficacy and safety of multiple fresh FMTs for Crohn’s disease (CD) complicated with intraabdominal inflammatory mass. The "one-hour FMT protocol" was followed in all patients. Twenty-five patients were diagnosed with CD and related inflammatory mass by CT or MRI. All patients received the initial FMT followed by repeated FMTs every 3 months. The primary endpoint was clinical response (improvement and remission) and sustained clinical remission at 12 months. Secondary endpoints were improvement in size of phegmon/abscess based upon cross-sectional imaging and safety of FMT. 68.0% (17/25) and 52.0% (13/25) of patients achieved clinical response and clinical remission at 3 months post the initial FMT, respectively. The proportion of patients at 6 months, 12 months and 18 months achieving sustained clinical remission with sequential FMTs was 48.0% (12/25), 32.0% (8/25) and 22.7% (5/22), respectively. 9.5% (2/21) of patients achieved radiological healing and 71.4% (15/21) achieved radiological improvement. No severe adverse events related to FMT were observed. This pragmatic study suggested that sequential fresh FMTs might be a promising, safe and effective therapy to induce and maintain clinical remission in CD with intraabdominal inflammatory mass.

Detecting and classifying the modulation type of the intercepted noisy LPI (low probability of intercept) radar signals in real-time is a necessary survival technique in the electronic intelligence systems. Most radar signals have been designed to have LPI properties; therefore, the LPI radar waveform recognition technique (LWRT) has recently gained increasing attention. In this paper, we propose a multiple feature images joint decision (MFIJD) model with two different feature extraction structures that fully extract the pixel feature to obtain the pre-classification results of each feature image for the non-stationary characteristics of most LPI radar signals. The core technology of this model is combining the short-time autocorrelation feature image, double short-time autocorrelation feature image and the original signal time-frequency image (TFI) simultaneously input into the hybrid model classifier, which is suitable for non-stationary signals, and it has higher universality. We demonstrate the performance of MFIJD by simulating 11 types of the signals defined in this paper and generating training sets and test sets. The comparison with the literature shows that the proposed methods not only has a high universality for LPI radar signals, but also better adapts to LPI radar waveform recognition at low SNR (signal to noise ratio) environment. The overall recognition rate of the method reaches 87.7% when the SNR is −6 dB.

Background and study aims: Placement of a tube through the anus into the cecum has not yet been established as a method of administering whole-colonic treatment. The aim of this study was to evaluate the safety, feasibility, and value of transendoscopic enteral tubing (TET) for fecal microbiota transplantation (FMT) through the colon. Patients and methods: A prospective observational study was performed of FMT using a new colonic TET technique. Under endoscopic guidance, a TET tube was affixed to the cecum with clips. The safety, value, and satisfaction with the FMT by TET were evaluated. Results: A total of 54 patients underwent TET. The success rate of the TET procedure was 100 % (54/54). Duration of the TET procedures was 14.8 ± 5.8 min. During the TET tube retention period, 98.1 % (53/54) of patients were satisfied with TET. The retention time for whole-colon delivery of the fecal microbiota suspension was 12.4 ± 2.3 days. In 88.4 % (49/54) of cases, no discomfort was reported during injection through the TET tube of the microbiota suspension. No adverse events were see in patients who required tube extubation after FMT. Conclusions: Colonic TET is a novel, safe, convenient, and reliable procedure for FMT that results in a high degree of patient satisfaction.

[Objectives] To observe the therapeutic effect of intragastric administration of Jiangtang Shuxin recipe on diabetic heart failure (DHF) in rats and to explore its mechanism. [Methods] Fifty SD rats were randomly divided into five groups, with 10 rats in each group. DHF models were prepared in the low-dose group, high-dose group, Western medicine group, and model group except the control group. Rats in the low-dose and high-dose groups were given 1.0 and 1.5 g/(kg • d) Jiangtang Shuxin recipe suspension by gavage, respectively. Rats in the Western medicine group were given gliquidone and benazepril by gavage for 2 months, and were fed with high-fat diet. Rats in the control group were fed with ordinary diet. Fasting blood glucose (FBG), serum triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), C-reactive protein (CRP), interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), creatine kinase isoenzyme (CK-MB), lactate dehydrogenase (LDH), pathological morphology of myocardial tissue, NF-KB p65 protein and IKB(Y protein were compared among groups. [Results] Compared with the control group, the levels of FBG, serum TG, LDL-C, CRP, IL-6, TNF-α, CK-MB and LDH increased, while the level of serum HDL-C decreased. The myocardial tissue was seriously damaged, and the expression of NF-κB p65 protein increased, while the expression of IκBα protein decreased in the other four groups (all P <0.05). Compared with the mod¬el group, the levels of FBG, serum TG, LDL-C, CRP, IL-6, TNF-α, CK-MB and LDH decreased, while the serum HDL-C level increased. The myocardial tissue damage was alleviated, and the expression of NF-κB p65 protein decreased, while the expression of IκB(Y protein in¬creased in the low-dose group, high-dose group and Western medicine group (all P<0.05). Compared with the Western medicine group, the levels of FBG, serum TG, LDL-C, CRP, IL-6, TNF-α, CK-MB and LDH decreased, and the level of serum HDL-C increased in the high-dose group (all P<0.05). [Conclusions] Jiangtang Shuxin recipe has a therapeutic effect on DHF in rats, with the best effect in the high-dose group. It can not only alleviate high glucose and high fat state, but also reduce myocardial injury and inflammation, and improve the pathological morphology of myocardial cells. The mechanism may be related to its inhibition of NF-κB signaling pathway. [ABSTRACT FROM AUTHOR]