Fine-Grained Video Captioning via Graph-based Multi-Granularity Interaction Learning.

Learning to generate continuous linguistic descriptions for multi-subject interactive videos in great details has particular applications in team sports auto-narrative. In contrast to traditional video caption, this task is more challenging as it requires simultaneous modeling of fine-grained individual actions, uncovering of spatio-temporal dependency structures of frequent group interactions, and then accurate mapping of these complex interaction details into long and detailed commentary. To explicitly address these challenges, we propose a novel framework Graph-based Learning for Multi-Granularity Interaction Representation (GLMGIR) for fine-grained team sports auto-narrative task. A multi-granular interaction modeling module is proposed to extract among-subjects' interactive actions in a progressive way for encoding both intra- and inter-team interactions. Based on the above multi-granular representations, a multi-granular attention module is developed to consider action/event descriptions of multiple spatio-temporal resolutions. Both modules are integrated seamlessly and work in a collaborative way to generate the final narrative. In the meantime, to facilitate reproducible research, we collect a new video dataset from YouTube.com called Sports Video Narrative dataset (SVN). It is a novel direction as it contains 6K team sports videos (i.e., NBA basketball games) with 10K ground-truth narratives(e.g., sentences). Furthermore, as previous metrics such as METEOR (i.e., used in coarse-grained video caption task) DO NOT cope with fine-grained sports narrative task well, we hence develop a novel evaluation metric named Fine-grained Captioning Evaluation (FCE), which measures how accurate the generated linguistic description reflects fine-grained action details as well as the overall spatio-temporal interactional structure. Extensive experiments on our SVN dataset have demonstrated the effectiveness of the proposed framework for fine-grained team sports video auto-narrative.

Processing of Affective Pictures: A Study Based on Functional Connectivity Network in the Cerebral Cortex.

Emotion plays an important role in people's life. However, the existing researches do not give a unified conclusion on the brain function network under different emotional states. In this study, pictures from the international affective picture system (IAPS) of different valences were presented to subjects with a fixed frequency blinking frequency to induce stable state visual evoked potential (SSVEP). With the source location method, the cerebral cortex source signal was reconstructed based on EEG signals, and then the difference in SSVEP amplitudes in key brain areas under different emotional states and the difference in brain function network connections among different brain areas were analysed in cortical space. The results of the study show that positive and negative emotions evoked greater activation intensities in the prefrontal, temporal, and parietal lobes compared with those of neutral emotion. The network connections with a significant difference between emotional states mainly appear in the alpha and gamma bands, and the network connections with significant differences between positive emotion and negative emotion mainly exist in the right middle temporal gyrus and the superior frontal gyrus on both sides. In addition, the long-range connections play an important role in the process of emotional processing, especially the connections among frontal gyrus, middle temporal gyrus, and middle occipital gyrus. The results of this study provide a reliable scientific basis for revealing and elucidating the neural mechanism of emotion processing and the selection of brain regions and frequency bands in emotion recognition based on EEG signals.

Neural Mechanism of Affective Perception: Evidence from Phase and Causality Analysis in the Cerebral Cortex.

Emotion plays an important role in people's lives. However, the neural mechanism of affective perception is still unclear. In this study, steady-state visual evoked potentials (SSVEPs) were used to explore information processing speed and interactions among cortical structures involved in affective perception. Pleasant, unpleasant, and neutral pictures selected from the International Affective Picture System were presented either in intact or phase-scrambled form at a fixed frequency, where the induced SSVEPs could be used as a frequency marker of brain activity with high temporal resolution and signal-to-noise ratio. Source estimation methods were used to reconstruct the cortical signals. The information processing of affective images was studied by phase and causal connection analysis in the cortical space to investigate the information processing speed of the local brain region and the dynamic interactions across brain regions. Experimental results showed that affective and semantic perception was accompanied by the acceleration of information processing speed in the ventral pathway. Unpleasant emotions had the fastest information processing speed in the ventral stream compared with pleasant and neutral emotions, including the middle occipital gyrus and the middle temporal gyrus, with a right hemisphere bias. In addition, unpleasant emotions were stronger than pleasant emotions in long-term causal connections in the bilateral middle temporal gyrus, and the direction was from the right hemisphere to the left hemisphere. These results provide unique insights into the cortical activities for affective perception and support the view that unpleasant emotions have priority in information perception in the middle temporal gyrus compared with pleasant and neutral emotions, with a right hemisphere bias.

Experimental study on bone defect repair by BMSCs combined with a light-sensitive material: g-C3N4/rGO.

Bone marrow mesenchymal stem cells (BMSCs), as seed cells, have played an important role in bone defect repair. However, efficiently amplifying and inducing BMSCs in vitro or vivo remains an urgent problem to be solved. Electrical stimulation has been beneficial to the proliferation and differentiation of BMSCs, but current electrical stimulation methods have a critical disadvantage in that they usually burn the skin. g-C3N4/rGO, a new photosensitive material, can produce photocurrent under natural light irradiation, thus reducing energy consumption. Our purpose was to explore whether this photocurrent can promote the proliferation and differentiation of BMSCs. g-C3N4/rGO synthesised under high temperature and pressure had negligible cytotoxicity as confirmed by methyl thiazolyl tetrazolium to BMSCs. Better osteogenesis was found in the blue light material group than in the light-shielding material group, exhibited by alizarin red staining, alkaline phosphatase activity, Western-Blot, and RT-qPCR. Animal experiments showed that the bone repair potential of the material group was significantly higher than that of the non-material group. Overall, we conclude that g-C3N4/rGO is a new non-toxic photosensitive material which can rapidly induce BMSCs into osteoblasts, accelerating bone regeneration and providing us with a feasible method of rapid bone repair.

High Gamma Band EEG Closely Related to Emotion: Evidence From Functional Network.

High-frequency electroencephalography (EEG) signals play an important role in research on human emotions. However, the different network patterns under different emotional states in the high gamma band (50-80 Hz) remain unclear. In this paper, we investigate different emotional states using functional network analysis on various frequency bands. We constructed multiple functional networks on different frequency bands and performed functional network analysis and time-frequency analysis on these frequency bands to determine the significant features that represent different emotional states. Furthermore, we verified the effectiveness of these features by using them in emotion recognition. Our experimental results revealed that the network connections in the high gamma band with significant differences among the positive, neutral, and negative emotional states were much denser than the network connections in the other frequency bands. The connections mainly occurred in the left prefrontal, left temporal, parietal, and occipital regions. Moreover, long-distance connections with significant differences among the emotional states were observed in the high frequency bands, particularly in the high gamma band. Additionally, high gamma band fusion features derived from the global efficiency, network connections, and differential entropies achieved the highest classification accuracies for both our dataset and the public dataset. These results are consistent with literature and provide further evidence that high gamma band EEG signals are more sensitive and effective than the EEG signals in other frequency bands in studying human affective perception.

Two-Level Domain Adaptation Neural Network for EEG-Based Emotion Recognition.

Emotion recognition plays an important part in human-computer interaction (HCI). Currently, the main challenge in electroencephalogram (EEG)-based emotion recognition is the non-stationarity of EEG signals, which causes performance of the trained model decreasing over time. In this paper, we propose a two-level domain adaptation neural network (TDANN) to construct a transfer model for EEG-based emotion recognition. Specifically, deep features from the topological graph, which preserve topological information from EEG signals, are extracted using a deep neural network. These features are then passed through TDANN for two-level domain confusion. The first level uses the maximum mean discrepancy (MMD) to reduce the distribution discrepancy of deep features between source domain and target domain, and the second uses the domain adversarial neural network (DANN) to force the deep features closer to their corresponding class centers. We evaluated the domain-transfer performance of the model on both our self-built data set and the public data set SEED. In the cross-day transfer experiment, the ability to accurately discriminate joy from other emotions was high: sadness (84%), anger (87.04%), and fear (85.32%) on the self-built data set. The accuracy reached 74.93% on the SEED data set. In the cross-subject transfer experiment, the ability to accurately discriminate joy from other emotions was equally high: sadness (83.79%), anger (84.13%), and fear (81.72%) on the self-built data set. The average accuracy reached 87.9% on the SEED data set, which was higher than WGAN-DA. The experimental results demonstrate that the proposed TDANN can effectively handle the domain transfer problem in EEG-based emotion recognition.

The Optimization Design Of Lactoferrin Loaded HupA Nanoemulsion For Targeted Drug Transport Via Intranasal Route.

BACKGROUND: Huperzine A (HupA) is a selective acetylcholinesterase inhibitor used to treat Alzheimer's disease. The existing dosage of HupA lacks brain selectivity and can cause serious side effects in the gastrointestinal and peripheral cholinergic systems. PURPOSE: The aim of this study was to develop and characterize a HupA nanoemulsion (NE) and a targeted HupA-NE modified with lactoferrin (Lf) for intranasal administration. METHODS: The NE was formulated using pseudo-ternary phase diagrams and optimized with response surface methodology. Particle size distribution and zeta potential were evaluated, and transmission electron microscopy was performed. We investigated the transport mechanisms of HupA-NEs into hCMEC/D3 cells, an in vitro model of the blood-brain barrier. HupA-NE, Lf-HupA-NE, and HupA solution were intranasally administered to rats to investigate the brain-targeting effects of these formulations. A drug targeting index (DTI) was calculated to determine brain-targeting efficiency. RESULTS: Optimized HupA-NE had a particle size of 15.24±0.67 nm, polydispersity index (PDI) of 0.128±0.025, and zeta potential of -4.48±0.97 mV. The composition of the optimized HupA-NE was 3.00% isopropyl myristate (IPM), 3.81% Capryol 90, and 40% Cremophor EL + Labrasol. NEs, particularly Lf-HupA-NE, were taken up into hCMEC/D3 cells to a greater extent than pure drug alone. Western blot analysis showed that hCMEC/D3 cells contained P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), and multidrug resistance associated protein 1 (MRP1) transporters. The likely mechanisms resulting in higher NE transport to the brain were uptake by specific transporters and transcytosis. In vivo, intranasal Lf-HupA-NE significantly enhanced drug delivery to the brain compared to HupA-NE, which was confirmed by differences in pharmacokinetic parameters. The DTI of Lf-HupA-NE (3.2±0.75) demonstrated brain targeting, and the area under the curve for Lf-HupA-NE was significantly higher than that for HupA-NE. CONCLUSION: Lf-HupA-NE is a promising nasal drug delivery carrier for facilitating delivery of HupA to the central nervous system.

Carboxymethyl Cellulose Binder Greatly Stabilizes Porous Hollow Carbon Submicrospheres in Capacitive K-Ion Storage.

On account of the large radius of K-ions, the electrodes can suffer huge deformation during K-ion insertion and extraction processes. In our work, we unveil the impact of using carboxymethyl cellulose (CMC) instead of poly(vinylidene fluoride) (PVDF) as binders for K-ion storage. Our porous hollow carbon submicrosphere anodes using the CMC binder exhibit a reversible capacity of 208 mA h g-1 after 50 cycles at 50 mA g-1, and even at a high current density of 1 A g-1, they achieve a reversible capacity of 111 mA h g-1 over 3000 cycles with almost no decay, demonstrating remarkably improved reversibility and cycling stability than those using PVDF (18 mA h g-1 after 3000 cycles at 1 A g-1). It is showed that the CMC binder can result in higher adhesion force and better mechanical performance than the PVDF binder, which can restrain the crack during a potassiation/depotassiation process. According to the test of adhesion force, the hollow carbon submicrospheres using the CMC binder show above three times of average adhesion force than that using the PVDF binder. Furthermore, based on the rational design, our hollow carbon submicrospheres also exhibit 62.3% specific capacity contribution below 0.5 V vs K/K+ region, which is helpful to design the full cell with high energy density. We believe that our work will highlight the binder effect to improve the K-ion storage performance.

Rational design of metal organic framework-derived FeS2 hollow nanocages@reduced graphene oxide for K-ion storage.

K-ion batteries (KIBs) have become one of the promising alternatives to lithium ion batteries. In this work, we are the first to utilize reduced graphene oxide (RGO) wrapped metal organic framework-derived FeS2 hollow nanocages (FeS2@RGO) as an anode for KIBs. Owing to the synergistic effect from FeS2 nanocages and RGO shells, our FeS2@RGO sample exhibited superior electrochemical performance. Such FeS2@RGO electrodes demonstrate a high capacity of 264 mA h g-1 after 50 cycles at 50 mA g-1 and 123 mA h g-1 after 420 cycles even at a large current density of 500 mA g-1. More importantly, we also explain the electrochemical reaction process about FeS2 and believe that these results would open the door for a novel class of long cycling performance anode materials in the KIB field.

Investigating Patterns for Self-Induced Emotion Recognition from EEG Signals.

Most current approaches to emotion recognition are based on neural signals elicited by affective materials such as images, sounds and videos. However, the application of neural patterns in the recognition of self-induced emotions remains uninvestigated. In this study we inferred the patterns and neural signatures of self-induced emotions from electroencephalogram (EEG) signals. The EEG signals of 30 participants were recorded while they watched 18 Chinese movie clips which were intended to elicit six discrete emotions, including joy, neutrality, sadness, disgust, anger and fear. After watching each movie clip the participants were asked to self-induce emotions by recalling a specific scene from each movie. We analyzed the important features, electrode distribution and average neural patterns of different self-induced emotions. Results demonstrated that features related to high-frequency rhythm of EEG signals from electrodes distributed in the bilateral temporal, prefrontal and occipital lobes have outstanding performance in the discrimination of emotions. Moreover, the six discrete categories of self-induced emotion exhibit specific neural patterns and brain topography distributions. We achieved an average accuracy of 87.36% in the discrimination of positive from negative self-induced emotions and 54.52% in the classification of emotions into six discrete categories. Our research will help promote the development of comprehensive endogenous emotion recognition methods.

Emotion Recognition from EEG Signals Using Multidimensional Information in EMD Domain.

This paper introduces a method for feature extraction and emotion recognition based on empirical mode decomposition (EMD). By using EMD, EEG signals are decomposed into Intrinsic Mode Functions (IMFs) automatically. Multidimensional information of IMF is utilized as features, the first difference of time series, the first difference of phase, and the normalized energy. The performance of the proposed method is verified on a publicly available emotional database. The results show that the three features are effective for emotion recognition. The role of each IMF is inquired and we find that high frequency component IMF1 has significant effect on different emotional states detection. The informative electrodes based on EMD strategy are analyzed. In addition, the classification accuracy of the proposed method is compared with several classical techniques, including fractal dimension (FD), sample entropy, differential entropy, and discrete wavelet transform (DWT). Experiment results on DEAP datasets demonstrate that our method can improve emotion recognition performance.

The impact of urban health insurance reform on hospital charges: a case study from two cities in China.

During the transition from a centrally planned to a market economy, China's urban health insurance system is being reformed. The control of the rapidly increasing hospital expenses will be a major determinant of the success of the reform. This study aims to examine the impact of the reform on hospital charges by comparing changes between two cities with different insurance systems and identifying determinants for those changes. Data was collected from six hospitals in two cities, one city implemented an urban health insurance reform, the other did not. Acute appendicitis and normal childbirth were used as tracers for calculating hospital charges. Methods included the retrospective review of medical records, interviews with health policy makers and hospital staff, focus group discussions, and the review of hospital and health insurance documents. The results showed that hospital charges per case of acute appendicitis and childbirth increased 101 and 94%, respectively, in the city without reform, and 41 and 34% in the city with reform, between 1995 and 1999. Health insurance arrangements and average LOS were the major determinants for hospital charges. Drugs and non-pharmacological treatments were the major service categories for charge containment. The combined measures of a single insurer, selective contracts, a new payment system, and use of an essential drug list, is regarded as the key features for an effective hospital charge control, and would appear to be successful measures for hospital expenditure containment within health insurance reform.