Attempts have been made to use cell transplantation and biomaterials to promote cell proliferation,differentiation,migration,and survival,as well as angiogenesis,in the context of brain injury.However,whether bioactiv...Attempts have been made to use cell transplantation and biomaterials to promote cell proliferation,differentiation,migration,and survival,as well as angiogenesis,in the context of brain injury.However,whether bioactive materials can repair the damage caused by ischemic stroke by activating endogenous neurogenesis and angiogenesis is still unknown.In this study,we applied chitosan gel loaded with basic fibroblast growth factor to the stroke cavity 7 days after ischemic stroke in rats.The gel slowly released basic fibroblast growth factor,which improved the local microenvironment,activated endogenous neural stem/progenitor cells,and recruited these cells to migrate toward the penumbra and stroke cavity and subsequently differentiate into neurons,while enhancing angiogenesis in the penumbra and stroke cavity and ultimately leading to partial functional recovery.This study revealed the mechanism by which bioactive materials repair ischemic strokes,thus providing a new strategy for the clinical application of bioactive materials in the treatment of ischemic stroke.展开更多
随着城市轨道交通的快速发展,客流量的准确预测对于改善运营服务至关重要。为了解决当前地铁客流预测存在的时空特性挖掘不充分等问题,进一步提高预测的精度与效率,研究了基于动态图神经常微分方程模型(multivariate time series with d...随着城市轨道交通的快速发展,客流量的准确预测对于改善运营服务至关重要。为了解决当前地铁客流预测存在的时空特性挖掘不充分等问题,进一步提高预测的精度与效率,研究了基于动态图神经常微分方程模型(multivariate time series with dynamic graph neural ordinary differential equations,MTGODE)的地铁短时客流预测方法。该方法通彭颢1贺玉过学习地铁站点间的动态关联强度构建动态拓扑图结构,基于学习得到的动态图进行连续图传播以传递时空信息、挖掘客流的依赖关系,并采用残差卷积提取多时间尺度下的周期性模式,实现了对站点间时空动态的连续表征,克服了传统图卷积网络模型难以刻画动态空间依赖的局限性。此外,为了充分挖掘不同站点间客流分布的时空规律,综合利用北京地铁自动售检票系统(auto fare collection,AFC)刷卡数据、天气数据、空气质量数据以及车站周边用地属性数据构建多源融合的客流预测模型。通过选取地铁北京站和积水潭站-东直门站的历史数据开展进站客流和OD客流预测实验,结果表明:与多个基准模型相比,该模型在平均绝对误差、均方根误差和平均百分比误差这3个指标中均取得了更优的预测效果,相较最优基准模型扩散卷积循环神经网络(diffusion convolutional recurrent neural network,DCRNN)分别降低了9.93%,12.30%,9.23%,对地铁客流时空分布的拟合程度更好,模型具有更好的预测精度、稳定性和拟合能力。展开更多
基金supported by the National Natural Science Foundation of China,Nos.81941011(to XL),31771053(to HD),31730030(to XL),31971279(to ZY),31900749(to PH),31650001(to XL),31320103903(to XL),31670988(to ZY)the Natural Science Foundation of Beijing,Nos.7222004(to HD)+1 种基金a grant from Ministry of Science and Technology of China,Nos.2017YFC1104002(to ZY),2017YFC1104001(to XL)a grant from Beihang University,No.JKF-YG-22-B001(to FH)。
文摘Attempts have been made to use cell transplantation and biomaterials to promote cell proliferation,differentiation,migration,and survival,as well as angiogenesis,in the context of brain injury.However,whether bioactive materials can repair the damage caused by ischemic stroke by activating endogenous neurogenesis and angiogenesis is still unknown.In this study,we applied chitosan gel loaded with basic fibroblast growth factor to the stroke cavity 7 days after ischemic stroke in rats.The gel slowly released basic fibroblast growth factor,which improved the local microenvironment,activated endogenous neural stem/progenitor cells,and recruited these cells to migrate toward the penumbra and stroke cavity and subsequently differentiate into neurons,while enhancing angiogenesis in the penumbra and stroke cavity and ultimately leading to partial functional recovery.This study revealed the mechanism by which bioactive materials repair ischemic strokes,thus providing a new strategy for the clinical application of bioactive materials in the treatment of ischemic stroke.
文摘随着城市轨道交通的快速发展,客流量的准确预测对于改善运营服务至关重要。为了解决当前地铁客流预测存在的时空特性挖掘不充分等问题,进一步提高预测的精度与效率,研究了基于动态图神经常微分方程模型(multivariate time series with dynamic graph neural ordinary differential equations,MTGODE)的地铁短时客流预测方法。该方法通彭颢1贺玉过学习地铁站点间的动态关联强度构建动态拓扑图结构,基于学习得到的动态图进行连续图传播以传递时空信息、挖掘客流的依赖关系,并采用残差卷积提取多时间尺度下的周期性模式,实现了对站点间时空动态的连续表征,克服了传统图卷积网络模型难以刻画动态空间依赖的局限性。此外,为了充分挖掘不同站点间客流分布的时空规律,综合利用北京地铁自动售检票系统(auto fare collection,AFC)刷卡数据、天气数据、空气质量数据以及车站周边用地属性数据构建多源融合的客流预测模型。通过选取地铁北京站和积水潭站-东直门站的历史数据开展进站客流和OD客流预测实验,结果表明:与多个基准模型相比,该模型在平均绝对误差、均方根误差和平均百分比误差这3个指标中均取得了更优的预测效果,相较最优基准模型扩散卷积循环神经网络(diffusion convolutional recurrent neural network,DCRNN)分别降低了9.93%,12.30%,9.23%,对地铁客流时空分布的拟合程度更好,模型具有更好的预测精度、稳定性和拟合能力。
