Enhanced Adjacency Matrix-Based Lightweight Graph Convolution Network for Action Recognition.

Graph convolutional networks (GCNs), which extend convolutional neural networks (CNNs) to non-Euclidean structures, have been utilized to promote skeleton-based human action recognition research and have made substantial progress in doing so. However, there are still some challenges in the construction of recognition models based on GCNs. In this paper, we propose an enhanced adjacency matrix-based graph convolutional network with a combinatorial attention mechanism (CA-EAMGCN) for skeleton-based action recognition. Firstly, an enhanced adjacency matrix is constructed to expand the model's perceptive field of global node features. Secondly, a feature selection fusion module (FSFM) is designed to provide an optimal fusion ratio for multiple input features of the model. Finally, a combinatorial attention mechanism is devised. Specifically, our spatial-temporal (ST) attention module and limb attention module (LAM) are integrated into a multi-input branch and a mainstream network of the proposed model, respectively. Extensive experiments on three large-scale datasets, namely the NTU RGB+D 60, NTU RGB+D 120 and UAV-Human datasets, show that the proposed model takes into account both requirements of light weight and recognition accuracy. This demonstrates the effectiveness of our method.

TFC-GCN: Lightweight Temporal Feature Cross-Extraction Graph Convolutional Network for Skeleton-Based Action Recognition.

For skeleton-based action recognition, graph convolutional networks (GCN) have absolute advantages. Existing state-of-the-art (SOTA) methods tended to focus on extracting and identifying features from all bones and joints. However, they ignored many new input features which could be discovered. Moreover, many GCN-based action recognition models did not pay sufficient attention to the extraction of temporal features. In addition, most models had swollen structures due to too many parameters. In order to solve the problems mentioned above, a temporal feature cross-extraction graph convolutional network (TFC-GCN) is proposed, which has a small number of parameters. Firstly, we propose the feature extraction strategy of the relative displacements of joints, which is fitted for the relative displacement between its previous and subsequent frames. Then, TFC-GCN uses a temporal feature cross-extraction block with gated information filtering to excavate high-level representations for human actions. Finally, we propose a stitching spatial-temporal attention (SST-Att) block for different joints to be given different weights so as to obtain favorable results for classification. FLOPs and the number of parameters of TFC-GCN reach 1.90 G and 0.18 M, respectively. The superiority has been verified on three large-scale public datasets, namely NTU RGB + D60, NTU RGB + D120 and UAV-Human.

CPMF-Net: Multi-Feature Network Based on Collaborative Patches for Retinal Vessel Segmentation.

As an important basis of clinical diagnosis, the morphology of retinal vessels is very useful for the early diagnosis of some eye diseases. In recent years, with the rapid development of deep learning technology, automatic segmentation methods based on it have made considerable progresses in the field of retinal blood vessel segmentation. However, due to the complexity of vessel structure and the poor quality of some images, retinal vessel segmentation, especially the segmentation of Capillaries, is still a challenging task. In this work, we propose a new retinal blood vessel segmentation method, called multi-feature segmentation, based on collaborative patches. First, we design a new collaborative patch training method which effectively compensates for the pixel information loss in the patch extraction through information transmission between collaborative patches. Additionally, the collaborative patch training strategy can simultaneously have the characteristics of low occupancy, easy structure and high accuracy. Then, we design a multi-feature network to gather a variety of information features. The hierarchical network structure, together with the integration of the adaptive coordinate attention module and the gated self-attention module, enables these rich information features to be used for segmentation. Finally, we evaluate the proposed method on two public datasets, namely DRIVE and STARE, and compare the results of our method with those of other nine advanced methods. The results show that our method outperforms other existing methods.

Epigallocatechin-3-gallate induced primary cultures of rat hippocampal neurons death linked to calcium overload and oxidative stress.

Epigallocatechin-3-gallate (EGCG), a catechin polyphenols component, is the main ingredient of green tea extract. It has been reported that EGCG is a potent antioxidant and beneficial in oxidative stress-related diseases, but others and our previous study showed that EGCG has pro-oxidant effects at high concentration. Thus, in this study, we tried to examine the possible pathway of EGCG-induced cell death in cultures of rat hippocampal neurons. Our results showed that EGCG caused a rapid elevation of intracellular free calcium levels ([Ca(2+)](i)) in a dose-dependent way. Exposure to EGCG dose- and time-dependently increased the production of reactive oxygen species (ROS) and reduced mitochondrial membrane potential (Deltapsi(m)) as well as the Bcl-2/Bax expression ratio. Importantly, acetoxymethyl ester of 5,5'-dimethyl-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid, ethylene glycol-bis-(2-aminoethyl)-N,N,N',N'-tetraacetic acid, and vitamin E could attenuate EGCG-induced apoptotic responses, including ROS generation, mitochondrial dysfunction, and finally partially prevented EGCG-induced cell death. Furthermore, treatment of hippocampal neurons with EGCG resulted in an elevation of caspase-3 and caspase-9 activities with no significant accompaniment of lactate dehydrogenase release, which provided further evidence that apoptosis was the dominant mode of EGCG-induced cell death in cultures of hippocampal neurons. Taken together, these findings indicated that EGCG induced hippocampal neuron death through the mitochondrion-dependent pathway.

Effects of EGCG on voltage-gated sodium channels in primary cultures of rat hippocampal CA1 neurons.

(-)-Epigallocatechin-3-gallate (EGCG), the main active component of green tea, is commonly known for its beneficial properties at low doses. On the other hand, little is known about the adverse effects of EGCG. Voltage-gated sodium channel (VGSC) is responsible for both initiation and propagation of action potentials of the neurons in the hippocampus and throughout the central nervous system (CNS). In this study, the effects of EGCG on voltage-gated sodium channel currents (I(Na)) were investigated in rat primary cultures of hippocampal CA1 neurons via the conventional whole-cell patch-clamp technique. We found that I(Na) was not affected by EGCG at the concentration of 0.1microM, but was completely blocked by EGCG at the concentration of 400microM and higher, and EGCG reduced the amplitudes of I(Na) in a concentration-dependent manner in the range of 0.1-400microM. Furthermore, our results also showed that at the concentration of 100microM, EGCG was known to have the following performances: (1) it decreased the activation threshold and the voltage at which the maximum I(Na) current was evoked, caused negative shifts of I(Na) steady-state activation curve. (2) It enlarged I(Na) tail-currents. (3) It induced a left shift of the steady-state inactivation. (4) It reduced fraction of available sodium channels. (5) It delayed the activation of I(Na) in a voltage-dependent manner. (6) It prolonged the time course of the fast inactivation of sodium channels. (7) It accelerated the activity-dependent attenuation of I(Na). On the basis of these findings, we propose that EGCG could impair certain physiological functions of VGSCs, which may contribute, directly or indirectly, to EGCG's effects in CNS.

Unmodified CdSe quantum dots induce elevation of cytoplasmic calcium levels and impairment of functional properties of sodium channels in rat primary cultured hippocampal neurons.

BACKGROUND: The growing applications of nanotechnologic products, such as quantum dots (QDs), increase the likelihood of exposure. Furthermore, their accumulation in the bioenvironment and retention in cells and tissues are arousing increasing worries about the potentially harmful side effects of these nanotechnologic products. Previous studies concerning QD cytotoxicity focused on the reactive oxygen species produced by QDs. Cellular calcium homeostasis dysregulation caused by QDs may be also responsible for QD cytotoxicity. Meanwhile the interference of QDs with voltage-gated sodium channel (VGSC) current (I(Na)) may lead to changes in electrical activity and worsen neurotoxicologic damage. OBJECTIVE: We aimed to investigate the potential for neurotoxicity of cadmium selenium QDs in a hippocampal neuronal culture model, focusing on cytoplasmic calcium levels and VGSCs function. METHODS: We used confocal laser scanning and standard whole-cell patch clamp techniques. RESULTS: We found that a) QDs induced neuron death dose dependently; b) cytoplasmic calcium levels were elevated for an extended period by QD treatment, which was due to both extracellular calcium influx and internal calcium release from endoplasmic reticulum; and c) QD treatment enhanced activation and inactivation of I(Na), prolonged the time course of activation, slowed I(Na) recovery, and reduced the fraction of available VGSCs. CONCLUSION: Results in this study provide new insights into QD toxicology and reveal potential risks of their future applications in biology and medicine.

Involvement of cyclin D1/CDK4 and pRb mediated by PI3K/AKT pathway activation in Pb2+ -induced neuronal death in cultured hippocampal neurons.

Lead (Pb) is widely recognized as a neurotoxicant. One of the suggested mechanisms of lead neurotoxicity is apoptotic cell death. And the mechanism by which Pb(2+) causes neuronal death is not well understood. The present study sought to examine the obligate nature of cyclin D1/cyclin-dependent kinase 4 (CDK4), phosphorylation of its substrate retinoblastoma protein (pRb) and its select upstream signal phosphoinositide 3-kinase (PI3K)/AKT pathway in the death of primary cultured rat hippocampal neurons evoked by Pb(2+). Our data showed that lead treatment of primary hippocampal cultures results in dose-dependent cell death. Inhibition of CDK4 prevented Pb(2+)-induced neuronal death significantly but was incomplete. In addition, we demonstrated that the levels of cyclin D1 and pRb/p107 were increased during Pb(2+) treatment. These elevated expression persisted up to 48 h, returning to control levels after 72 h. We also presented pharmacological and morphological evidences that cyclin D1/CDK4 and pRb/p107 were required for such kind of neuronal death. Addition of the PI3K inhibitor LY294002 (30 microM) or wortmannin (100 nM) significantly rescued the cultured hippocampal neurons from death caused by Pb(2+). And that Pb(2+)-elicited phospho-AKT (Ser473) participated in the induction of cyclin D1 and partial pRb/p107 expression. These results provide evidences that cell cycle elements play a required role in the death of neurons evoked by Pb(2+) and suggest that certain signaling elements upstream of cyclin D1/CDK4 are modified and/or required for this form of neuronal death.