Effect of electroacupuncture at "Neiguan" (PC 6) and "Zusanli" (ST 36) on gastrointestinal hormone in the antral tissue of rats with functional dyspepsia. / 电针"å† å ³""足三里"å¯¹åŠŸèƒ½æ€§æ¶ˆåŒ–ä¸è‰¯å¤§é¼ èƒƒçª¦ç»„ç»‡èƒƒè‚ æ¿€ç´ çš„å½±å“.

OBJECTIVES: To observe the effects of electroacupuncture (EA) at "Neiguan" (PC 6) and "Zusanli"(ST 36) on the gastric emptying rate, the level of serotonin (5-HT) and the protein expression of motilin (MTL), ghrelin, substance P (SP) and vasoactive intestinal peptide (VIP) in the antral tissue of the rats with functional dyspepsia (FD) and explore the effect mechanism of EA in treatment of FD. METHODS: A total of 21 SPF male SD rat pups were randomly divided into a normal group, a model group and an EA group, with 7 rats in each group. In the model group and the EA group, FD model was prepared by the gavage with 0.1% sucrose iodoacetamide solution combined with the modified small platform method. After the successful modeling, EA was applied to "Neiguan" (PC 6) and "Zusanli"(ST 36) in the rats of the EA group, with disperse-dense wave, 20 Hz/100 Hz in frequency, stimulated for 30 min, once daily, for 7 days consecutively. Before and after intervention, the general condition of the rats was observed in each group. After the completion of intervention, the gastric emptying rate was measured, the morphological changes of gastric antral tissue were observed using HE staining, the level of 5-HT was detected with ELISA method, and the protein expression of MTL, ghrelin, SP, and VIP was determined with Western blot method in the antral tissue of rats. RESULTS: In the normal group, the rats were in a good mental state, with lustrous fur, flexible movement and the increase of food intake and body mass. In the model group, the rats were poor in mental state, lack of lustre in fur, preference for the body curled up, reduced activity and response; and a part of rats had loose stool, obviously enlarged gastric body and gastric food retention. In the EA group, the general condition of rats, e.g. the mental state, food intake and activity, were improved, the gastric body got smaller obviously and the gastric food retention was reduced when compared with the model group. The antral structure was intact, the glands were rich and no injury of the gastric mucosa was found, e.g. inflammatory reaction and edema in the rats of each group. Compared with the normal group, the gastric emptying rate was decreased (P<0.01), 5-HT level was increased (P<0.01), the protein expression of MTL and ghrelin was reduced (P<0.01) and that of VIP was elevated (P<0.01) in the rats of the model group. The gastric emptying rate was increased (P<0.01), 5-HT level was decreased (P<0.01), and the protein expression of MTL and ghrelin was elevated (P<0.05, P<0.01) in the rats of the EA group when compared with those in the model group. CONCLUSIONS: Electroacupuncture at "Neiguan" (PC 6) and "Zusanli"(ST 36) may effectively relieve gastric dysfunction, strengthen gastric motility and promote gastric emptying so as to alleviate the symptoms of dyspepsia in FD rats, and its mechanism may be related to the regulation of gastrointestinal hormones in the antral tissue.

Heterogeneous Collaborative Learning for Personalized Healthcare Analytics via Messenger Distillation.

The Healthcare Internet-of-Things (IoT) framework aims to provide personalized medical services with edge devices. Due to the inevitable data sparsity on an individual device, cross-device collaboration is introduced to enhance the power of distributed artificial intelligence. Conventional collaborative learning protocols (e.g., sharing model parameters or gradients) strictly require the homogeneity of all participant models. However, real-life end devices have various hardware configurations (e.g., compute resources), leading to heterogeneous on-device models with different architectures. Moreover, clients (i.e., end devices) may participate in the collaborative learning process at different times. In this paper, we propose a Similarity-Quality-based Messenger Distillation (SQMD) framework for heterogeneous asynchronous on-device healthcare analytics. By introducing a preloaded reference dataset, SQMD enables all participant devices to distill knowledge from peers via messengers (i.e., the soft labels of the reference dataset generated by clients) without assuming the same model architecture. Furthermore, the messengers also carry important auxiliary information to calculate the similarity between clients and evaluate the quality of each client model, based on which the central server creates and maintains a dynamic collaboration graph (communication graph) to improve the personalization and reliability of SQMD under asynchronous conditions. Extensive experiments on three real-life datasets show that SQMD achieves superior performance.

Quaternion Factorization Machines: A Lightweight Solution to Intricate Feature Interaction Modeling.

Due to the sparsity of available features in web-scale predictive analytics, combinatorial features become a crucial means for deriving accurate predictions. As a well-established approach, a factorization machine (FM) is capable of automatically learning high-order interactions among features to make predictions without the need for manual feature engineering. With the prominent development of deep neural networks (DNNs), there is a recent and ongoing trend of enhancing the expressiveness of FM-based models with DNNs. However, though better results are obtained with DNN-based FM variants, such performance gain is paid off by an enormous amount (usually millions) of excessive model parameters on top of the plain FM. Consequently, the heavy parameterization impedes the real-life practicality of those deep models, especially efficient deployment on resource-constrained Internet of Things (IoT) and edge devices. In this article, we move beyond the traditional real space where most deep FM-based models are defined and seek solutions from quaternion representations within the hypercomplex space. Specifically, we propose the quaternion factorization machine (QFM) and quaternion neural factorization machine (QNFM), which are two novel lightweight and memory-efficient quaternion-valued models for sparse predictive analytics. By introducing a brand new take on FM-based models with the notion of quaternion algebra, our models not only enable expressive inter-component feature interactions but also significantly reduce the parameter size due to lower degrees of freedom in the hypercomplex Hamilton product compared with real-valued matrix multiplication. Extensive experimental results on three large-scale datasets demonstrate that QFM achieves 4.36% performance improvement over the plain FM without introducing any extra parameters, while QNFM outperforms all baselines with up to two magnitudes' parameter size reduction in comparison to state-of-the-art peer methods.

Time Interval-Enhanced Graph Neural Network for Shared-Account Cross-Domain Sequential Recommendation.

Shared-account cross-domain sequential recommendation (SCSR) task aims to recommend the next item via leveraging the mixed user behaviors in multiple domains. It is gaining immense research attention as more and more users tend to sign up on different platforms and share accounts with others to access domain-specific services. Existing works on SCSR mainly rely on mining sequential patterns via recurrent neural network (RNN)-based models, which suffer from the following limitations: 1) RNN-based methods overwhelmingly target discovering sequential dependencies in single-user behaviors and they are not expressive enough to capture the relationships among multiple entities in SCSR; 2) all existing methods bridge two domains via knowledge transfer in the latent space and ignore the explicit cross-domain graph structure; and 3) none existing studies consider the time interval information among items, which is essential in the sequential recommendation for characterizing different items and learning discriminative representations for them. In this work, we propose a new graph-based solution, namely, time interval-enhanced domain-aware graph convolutional network (TiDA-GCN), to address the above challenges. Specifically, we first link users and items in each domain as a graph. Then, we devise a domain-aware graph convolution network to learn user-specific node representations. To fully account for users' domain-specific preferences on items, two effective attention mechanisms are further developed to selectively guide the message-passing process. Moreover, to further enhance item-and account-level representation learning, we incorporate the time interval into the message passing and design an account-aware self-attention module for learning items' interactive characteristics. Experiments demonstrate the superiority of our proposed method from various aspects.

Stunting among kindergarten children in China in the context of COVID-19: A cross-sectional study.

Background: The impact of COVID-19 has most likely increased the prevalence of stunting. The study aimed to determine the prevalence of stunting among kindergarten children in the context of coronavirus disease 2019 (COVID-19) in Longgang District, Shenzhen, China, and its risk factors. Methods: A cross-sectional study was conducted to identify children from 11 sub districts of 481 kindergartens in the Longgang District of Shenzhen City from May to July 2021. In the context of COVID-19, an online survey was conducted to gather demographic information, height, birth information, and lifestyle. The prevalence of stunting was calculated, and the risk factors were analyzed using binary logistic regression with three stepwise models. Results: A total of 118,404 subjects were included from May to July 2021, with a response and questionnaire effective rates of 85.75% and 95.03%, respectively. The prevalence of stunting and severe stunting were 3.3% and 0.8%, respectively. Model 3 showed that risk factors for stunting were male sex [odds ratio (OR) = 1.07], low birth weight (OR = 2.02), insufficient sleep time (OR = 1.08), less food intake than their peers (OR = 1.66), slower eating than their peers (OR = 1.16), accompanied by grandparents alone or non-lineal relatives (reference: parents accompanying) (OR = 1.23, 1.51), and children induced to eat (OR = 1.17). Protective factors included only-child status (OR = 0.66), reported high activity (OR = 0.37, 0.26, 0.23), parents with high education levels (father: OR = 0.87, 0.69; mother: OR = 0.69, 0.58), high monthly income per capita of the family (OR = 0.88, 0.74, 0.68), and allowing children to make food choices (OR = 0.82). Conclusion: The stunting rate of children in kindergartens in Longgang District is 3.3%, close to the level of developed countries but higher than the average level of developed cities in China. The relatively high stunting rate in children under 3 years old in 2021 may be associated with the influence of COVID-19. Appropriate policies should be formulated for individuals and families with children to help children establish good living habits and reduce stunting.

[Comparison of stimulation-induced local changes of expression of local Meckel cell-related proteins and neuropeptides and number of mast cells among acupuncture and moxibustion alone or in combination in chronic atrophic gastritis rats].

OBJECTIVE: To compare the effects of acupuncture and moxibustion alone or in combination on the number of mast cells and expression levels of cytoketatin 18 (CK18) and CK19 (marker of Meckel cells), and calcitonin gene-related peptide (CGRP), neuropeptide-Y (NPY) and bradykinin (BK) in the local acupoint area of rats with chronic atrophic gastritis (CAG). METHODS: Fifty male SD rats were randomly divided into normal, CAG model, moxibustion, acupuncture and acupuncture+moxi-bustion groups (10 rats in each group). The CAG model was established by gavage of 1-methyl-3-nitro-1-nitrosoguanidine (170 µg/mL,1 mL/100 g, once a week) and 40% ethanol solution (twice a week) for 12 consecutive weeks. After successful establishment of CAG model, moxibustion, manual acupuncture or acupuncture+moxibustion was applied to bilateral "Zusanli" (ST36) and "Zhongwan"(CV12) for 15 min, once daily for 14 consecutive days. At the end of the experiment, the gastric mucosal tissues were collected for observing histopathological changes of gastric mucosa after H.E. staining, and the tissues of the stimulated ST36 region collected for detecting the expression levels of CK18, CK19, CGRP, NPY and BK and the number of mast cells in the local ST36 region by immunohistochemistry. RESULTS: Compared with the normal group, the number of mast cells, the expression levels of CK19, NPY and BK in the ST36 area were significantly increased (P<0.05), and the expression level of CGRP was apparently decreased (P<0.05) in the model group. In comparison with the model group, the number of mast cells and the expression levels of CGRP and NPY in the moxibustion group, the expression of CGRP in the acupuncture group, and the number of mast cells, as well as the expression levels of CK18, CK19 and CGRP in the acupuncture+moxibustion group were significantly up-re-gulated (P<0.05). The effect of acupuncture combined with moxibustion was obviously superior to that of moxibustion or acupuncture in up-regulating the expression of CK18 and CK19 (P<0.05) and superior to that of moxibustion in down-regulating BK expressio level (P<0.05). No significant changes were found in the expression of CK18 after modeling (vs the normal group), in the expression levels of CK18, CK19 and BK after moxibustion and acupuncture (vs the model group), in the number of mast cells and expression of NPY after acupuncture (vs the model group), and in the expression levels of NPY and BK after acupuncture+moxibustion (vs the model group) (P>0.05). H.E. staining showed infiltration of many lymphocytes in the gastric mucosa and submucosal layers, atrophy and necrosis of lots of main cells with vacuole-like changes, and disordered arrangement of the atrophic glands in the model group, which was milder particularly in the acupuncture + moxibustion group. CONCLUSION: Acupuncture combined with moxibustion of ST36 can up-regulate the levels of local CK18, CK19 and CGRP proteins and number of mast cells, moxibustion may up-regulate the levels of CGRP and NPY and number of mast cells, while acupuncture may up-regulate the expression of CGRP in the local stimulated area in CAG rats.

An Integrated First Principal and Deep Learning Approach for Modeling Nitrous Oxide Emissions from Wastewater Treatment Plants.

Mathematical modeling plays a critical role toward the mitigation of nitrous oxide (N2O) emissions from wastewater treatment plants (WWTPs). In this work, we proposed a novel hybrid modeling approach by integrating the first principal model with deep learning techniques to predict N2O emissions. The hybrid model was successfully implemented and validated with the N2O emission data from a full-scale WWTP. This hybrid model is demonstrated to have higher accuracy for N2O emission modeling in the WWTP than the mechanistic model or pure deep learning model. Equally important, the hybrid model is more applicable than the pure deep learning model due to the lower requirement of data and the pure mechanistic model due to the less calibration requirement. This superior performance was due to the hybrid nature of the proposed model. It integrated the essential wastewater treatment knowledge as the first principal component and the less understood N2O production processes by the data-driven deep learning approach. The developed hybrid model was also successfully implemented under different circumstances for the prediction of N2O flux, which showed the generalizability of the model. The hybrid model also showed great potential to be applied for the N2O mitigation work. Nevertheless, the capability of the hybrid model in evaluating N2O mitigation strategies still requires validation with experiments. Going beyond N2O modeling in WWTP, the novel hybridization modeling concept can potentially be applied to other environmental systems.

Personalized On-Device E-Health Analytics With Decentralized Block Coordinate Descent.

Actuated by the growing attention to personal healthcare and the pandemic, the popularity of E-health is proliferating. Nowadays, enhancement on medical diagnosis via machine learning models has been highly effective in many aspects of e-health analytics. Nevertheless, in the classic cloud-based/centralized e-health paradigms, all the data will be centrally stored on the server to facilitate model training, which inevitably incurs privacy concerns and high time delay. Distributed solutions like Decentralized Stochastic Gradient Descent (D-SGD) are proposed to provide safe and timely diagnostic results based on personal devices. However, methods like D-SGD are subject to the gradient vanishing issue and usually proceed slowly at the early training stage, thereby impeding the effectiveness and efficiency of training. In addition, existing methods are prone to learning models that are biased towards users with dense data, compromising the fairness when providing E-health analytics for minority groups. In this paper, we propose a Decentralized Block Coordinate Descent (D-BCD) learning framework that can better optimize deep neural network-based models distributed on decentralized devices for E-health analytics. As a gradient-free optimization method, Block Coordinate Descent (BCD) mitigates the gradient vanishing issue and converges faster at the early stage compared with the conventional gradient-based optimization. To overcome the potential data scarcity issues for users' local data, we propose similarity-based model aggregation that allows each on-device model to leverage knowledge from similar neighbor models, so as to achieve both personalization and high accuracy for the learned models. Benchmarking experiments on three real-world datasets illustrate the effectiveness and practicality of our proposed D-BCD, where additional simulation study showcases the strong applicability of D-BCD in real-life E-health scenarios.

FENet: A Frequency Extraction Network for Obstructive Sleep Apnea Detection.

Obstructive Sleep Apnea (OSA) is a highly prevalent but inconspicuous disease that seriously jeopardizes the health of human beings. Polysomnography (PSG), the gold standard of detecting OSA, requires multiple specialized sensors for signal collection, hence patients have to physically visit hospitals and bear the costly treatment for a single detection. Recently, many single-sensor alternatives have been proposed to improve the cost efficiency and convenience. Among these methods, solutions based on RR-interval (i.e., the interval between two consecutive pulses) signals reach a satisfactory balance among comfort, portability and detection accuracy. In this paper, we advance RR-interval based OSA detection by considering its real-world practicality from energy perspectives. As photoplethysmogram (PPG) pulse sensors are commonly equipped on smart wrist-worn wearable devices (e.g., smart watches and wristbands), the energy efficiency of the detection model is crucial to fully support an overnight observation on patients. This creates challenges as the PPG sensors are unable to keep collecting continuous signals due to the limited battery capacity on smart wrist-worn devices. Therefore, we propose a novel Frequency Extraction Network (FENet), which can extract features from different frequency bands of the input RR-interval signals and generate continuous detection results with downsampled, discontinuous RR-interval signals. With the help of the one-to-multiple structure, FENet requires only one-third of the operation time of the PPG sensor, thus sharply cutting down the energy consumption and enabling overnight diagnosis. Experimental results on real OSA datasets reveal the state-of-the-art performance of FENet.

Disease Prediction via Graph Neural Networks.

With the increasingly available electronic medical records (EMRs), disease prediction has recently gained immense research attention, where an accurate classifier needs to be trained to map the input prediction signals (e.g., symptoms, patient demographics, etc.) to the estimated diseases for each patient. However, existing machine learning-based solutions heavily rely on abundant manually labeled EMR training data to ensure satisfactory prediction results, impeding their performance in the existence of rare diseases that are subject to severe data scarcity. For each rare disease, the limited EMR data can hardly offer sufficient information for a model to correctly distinguish its identity from other diseases with similar clinical symptoms. Furthermore, most existing disease prediction approaches are based on the sequential EMRs collected for every patient and are unable to handle new patients without historical EMRs, reducing their real-life practicality. In this paper, we introduce an innovative model based on Graph Neural Networks (GNNs) for disease prediction, which utilizes external knowledge bases to augment the insufficient EMR data, and learns highly representative node embeddings for patients, diseases and symptoms from the medical concept graph and patient record graph respectively constructed from the medical knowledge base and EMRs. By aggregating information from directly connected neighbor nodes, the proposed neural graph encoder can effectively generate embeddings that capture knowledge from both data sources, and is able to inductively infer the embeddings for a new patient based on the symptoms reported in her/his EMRs to allow for accurate prediction on both general diseases and rare diseases. Extensive experiments on a real-world EMR dataset have demonstrated the state-of-the-art performance of our proposed model.

Few-Shot Deep Adversarial Learning for Video-based Person Re-identification.

Video-based person re-identification (re-ID) refers to matching people across camera views from arbitrary unaligned video footages. Existing methods rely on supervision signals to optimise a projected space under which the distances between inter/intra-videos are maximised/minimised. However, this demands exhaustively labelling people across camera views, rendering them unable to be scaled in large networked cameras. Also, it is noticed that learning effective video representations with view invariance is not explicitly addressed for which features exhibit different distributions otherwise. Thus, matching videos for person re-ID demands flexible models to capture the dynamics in time-series observations and learn view-invariant representations with access to limited labeled training samples. In this paper, we propose a novel few-shot deep learning approach to videobased person re-ID, to learn comparable representations that are discriminative and view-invariant. The proposed method is developed on the variational recurrent neural networks (VRNNs) and trained adversarially to produce latent variables with temporal dependencies that are highly discriminative yet view-invariant in matching persons. Through extensive experiments conducted on three benchmark datasets, we empirically show the capability of our method in creating view-invariant temporal features and state-of-the-art performance achieved by our method.

MCP: A multi-component learning machine to predict protein secondary structure.

The Gene or DNA sequence in every cell does not control genetic properties on its own; Rather, this is done through the translation of DNA into protein and subsequent formation of a certain 3D structure. The biological function of a protein is tightly connected to its specific 3D structure. Prediction of the protein secondary structure is a crucial intermediate step towards elucidating its 3D structure and function. Traditional experimental methods for prediction of protein structure are expensive and time-consuming. Nevertheless, the average accuracy of the suggested solutions has hardly reached beyond 80%. The possible underlying reasons are the ambiguous sequence-structure relation, noise in input protein data, class imbalance, and the high dimensionality of the encoding schemes. Furthermore, we utilize a compound string dissimilarity measure to directly interpret protein sequence content and avoid information loss. In order to improve accuracy, we employ two different classifiers including support vector machine and fuzzy nearest neighbor and collectively aggregate the classification outcomes to infer the final protein structures. We conduct comprehensive experiments to compare our model with the current state-of-the-art approaches. The experimental results demonstrate that given a set of input sequences, our multi-component framework can accurately predict the protein structure. Nevertheless, the effectiveness of our unified model can be further enhanced through framework configuration.

[Impact of ecological protection construction on schistosomiasis transmission of Qionghai Lake wetland in Xichang City].

OBJECTIVE: To understand the impact of Qionghai Lake wetland ecological protection construction on the prevalence of schistosomiasis, so as to provide the evidence for formulating the strategies for schistosomiasis control and prevention. METHODS: A retrospective survey of the construction of Qionghai Lake wetland was performed, and eleven villages around the wetland were surveyed for schistosomiasis endemic situation. The influence of the wetland project on the schistosomiasis prevalence and Oncomelania hupensis snail status were investigated. RESULTS: Before the construction of Qionghai Lake wetland, the snail elimination and extended chemotherapy for residents was performed. After the project was finished, the roads and ditches were hardened. From 2009 to 2014, the schistosome infection rate of residents declined from 0.37% to 0. No schistosome infected snails were found and in recent 2 years, no snails were found. No mice were infected in the sentinel tests. CONCLUSIONS: The construction of Qionghai Lake wetland effectively eliminates snails, and interrupts the transmission of schistosomiasis. However, the environment of the wetland is more suitable for snail breeding, and therefore, the surveillance still should be strengthened.

[Water exposure modes and times of different populations in mountainous schistosomiasis endemic areas near Qionghai lake].

OBJECTIVE: To investigate water exposure modes and times of different populations in mountainous schistosomiasis endemic areas and to inform about the control strategies. METHODS: All 1054 residents from populations around Qionghai Lake were randomly sampled according to occupation for a retrospective questionnaire survey in November 2001. Each individual was interviewed for his/her mode, frequency, and duration of water exposure occurring between April and October 2001. RESULTS: The average exposure times and intensity were higher in farmers (median: 16 - 18 min/day and 2.41 - 2.5, respectively) who grow rice, tobacco, and vegetables than others (median: 3.74 - 7.39 min/day and 0.81 - 1.52, respectively); exposure frequency was found highest in farmers (median: 2.04 times/day) in all occupations; schoolchildren had low exposure frequency and times, but very high exposure intensity (median 2.34). Between April and June it is an agriculturally busy season, that is also a peak season of water exposure of adults. Schoolchildren's water exposure peaks on July and August, mainly due to playing water and swimming. Exposure times and intensities were higher in females than in males. CONCLUSION: Water exposure modes, times, and intensities of different populations were different in mountainous schistosomiasis endemic areas of Xichang. Between April and June should be the peak infection season of adults who are engaging in agricultural activities, while July to August should be the peak infection season for schoolchildren with non-agricultural activities.