基于上转换纳米粒子的低损伤神经界面技术

Optogenetics is a neuromodulation technology that combines light control technology with genetic technology, thus allowing the selective activation and inhibition of the electrical activity in specific types of neurons with millisecond time resolution. Over the past several years, optogenetics has become a powerful tool for understanding the organization and functions of neural circuits, and it holds great promise to treat neurological disorders. To date, the excitation wavelengths of commonly employed opsins in optogenetics are located in the visible spectrum. This poses a serious limitation for neural activity regulation because the intense absorption and scattering of visible light by tissues lead to the loss of excitation light energy and also cause tissue heating. To regulate the activity of neurons in deep brain regions, it is necessary to implant optical fibers or optoelectronic devices into target brain areas, which however can induce severe tissue damage. Non- or minimally-invasive remote control technologies that can manipulate neural activity have been highly desirable in neuroscience research. Upconversion nanoparticles (UCNPs) can emit light with a short wavelength and high frequency upon excitation by light with a long wavelength and low frequency. Therefore, UCNPs can convert low-frequency near-infrared (NIR) light into high-frequency visible light for the activation of light-sensitive proteins, thus indirectly realizing the NIR optogenetic system. Because NIR light has a large tissue penetration depth, UCNP-mediated optogenetics has attracted significant interest for deep-tissue neuromodulation. However, in UCNP-mediated in vivo optogenetic experiments, as the up-conversion efficiency of UCNPs is low, it is generally necessary to apply high-power NIR light to obtain up-converted fluorescence with energy high enough to activate a photosensitive protein. High-power NIR light can cause thermal damage to tissues, which seriously restricts the applications of UCNPs in optogenetic technology. Therefore, the exploration of strategies to increase the up-conversion efficiency, fluorescence intensity, and biocompatibility of UCNPs is of great significance to their wide applications in optogenetic systems. This review summarizes recent developments and challenges in UCNP-mediated optogenetics for deep-brain neuromodulation. We firstly discuss the correspondence between the parameters of UCNPs and employed opsins in optogenetic experiments, which mainly include excitation wavelengths, emission wavelengths, and luminescent lifetimes. Thereafter, we introduce the methods to enhance the conversion efficiency of UCNPs, including optimizing the structure of UCNPs and modifying the organic dyes in UCNPs. In addition, we also discuss the future opportunities in combining UCNP-mediated optogenetics with flexible microelectrode technology for the long-term detection and regulation of neural activity in the case of minimal injury.     

面向脑机接口的犹他神经电极技术

A human brain is composed of a large number of interconnected neurons forming a neural network. To study the functional mechanism of the neural network, it is necessary to record the activity of individual neurons over a large area simultaneously. Brain-computer interface (BCI) refers to the connection established between the human/animal brain and computers/other electronic devices, which enables direct interaction between the brain and external devices. It plays an important role in understanding, protecting, and simulating the brain, especially in helping patients with neurological disorders to restore their impaired motor and sensory functions. Neural electrodes are electrophysiological devices that form the core of BCI, which convert neuronal electrical signals (carried by ions) into general electrical signals (carried by electrons). They can record or interfere with the state of neural activity. The Utah Electrode Array (UEA) designed by the University of Utah is a mainstream neural electrode fabricated by bulk micromachining. Its unique three-dimensional needle-like structure enables each electrode to obtain high spatiotemporal resolution and good insulation between each other. After implantation, the tip of each electrode affects only a small group of neurons around it even allowing to record the action potential of a single neuron. The availability of a large number of electrodes, high quality of signals, and long service life has made UEA the first choice for collecting neuronal signals. Moreover, UEA is the only implantable neural electrode that can record signals in the human cerebral cortex. This article mainly serves as an introduction to the construction, manufacturing process, and functioning of UEA, with a focus on the research progress in fabricating high-density electrode arrays, wireless neural interfaces, and optrode arrays using silicon, glass, and metal as that material of construction. We also discuss the surface modification techniques that can be used to reduce the electrode impedance, minimize the rejection by brain tissue, and improve the corrosion resistance of the electrode. In addition, we summarize the clinical applications where patients can control external devices and get sensory feedback by implanting UEA. Furthermore, we discuss the challenges faced by existing electrodes such as the difficulty in increasing electrode density, poor response of integrated wireless neural interface, and the problems of biocompatibility. To achieve stability and durability of the electrode, advancements in both material science and manufacturing technology are required. We hope that this review can broaden the scope of ideas for the development of UEA. The realization of a fully implantable neural microsystem can contribute to an improved understanding of the functional mechanisms of the neural network and treatment of neurological diseases.     

基于依存距离惩罚的泰汉神经机器翻译方法

在“一带一路”倡议的背景下,泰国与我国的沟通日益密切,泰汉机器翻译具有很高的应用需求。作为低资源语言机器翻译,泰汉之间缺乏大规模、高质量的平行语料,导致翻译效果不佳。融入依存句法知识可以使译文更符合句法约束,弥补没有大规模平行语料的缺陷,但泰语缺少成熟的依存解析工具和依存标注训练集。针对以上问题,基于无监督迁移获取泰语依存句法结构知识,并提出依存距离惩罚机制以减少依存噪声对翻译性能的干扰,并通过基于依存感知注意力机制的Transformer架构将所提方法融入翻译。实验结果表明,该方法能有效提升泰汉神经机器翻译的效果。     

高效液相色谱-串联质谱法同时检测血液中3种季铵盐类肌松剂

维库溴铵、罗库溴铵和泮库溴铵是一类广泛使用的非去极化肌松剂,使用过程中引起过敏反应甚至死亡的情况时有发生,快速检测血液中该类肌松剂的浓度,可为临床早期诊断提供有价值的信息.该类肌松剂为强极性的季铵盐类化合物,在反相色谱柱上难以保留,主要采用离子对色谱法进行分离.采用离子对色谱法时,加入的离子对试剂有离子抑制作用,可降低...     

56届匹茨堡分析化学暨应用谱学会议(Ⅱ)

应用更灵敏和更小型的分析技术探索脑秘密组的报告有:(59)M Wightman,应用电分析技术探测脑的化学;(60)A C Michael,通过来自脑部的伏安法和微渗析法的结果比较、测量组织对植入器件的反应;(61)A G Ewing,梳理出纳米级神经传递囊性能的分析方法;(62)N Allbritton,梳理出单细胞中信号转换通道的细节;(63)J V Sweedler,逐个细胞方式测定脑部的神经肽--大量质谱技术的平行微型分析.     

无声语音识别中的端点检测算法

无声语音识别常以表面肌电信号作为研究对象。表面肌电信号的端点检测是影响识别结果的一个重要因素。表面肌电信号与语音信号有类似之处。借助语音端点检测的方法对表面肌电信号进行分割是一种可行的思路。基于此,采用子带谱熵和梅尔倒谱距离作为信号端点检测的判决依据,通过粒子群算法优化支持向量机分类器给出端点检测结果。结果表明,在不同信噪比条件下,该算法有最高的检出率与最低的错误率。对于基于K最近邻（K-Nearest Neighbor,KNN）的无声语音识别任务,识别率达95.3%。     

基于人工神经网络的表面肌电信号分类器研究进展

表面肌电信号是一种重要的生理电信号.基于表面肌电信号建立人体康复动作识别系统操作方便,对身体无侵入性,对运动无干扰,有广阔的应用前景.肢体康复动作识别系统很大程度上依赖于信号特征提取和分类器的使用.文中对近几年基于人工神经网络建立的表面肌电信号分类器,包括LVQ分类器、ELM分类器、WNN分类器、ANFIS分类器、Al...     

广义神经传播方程的整体吸引子

使用Galerkin方法,结合Sobolev空间理论和不等式技巧,给出了广义神经传播方程解的存在唯一性定理,然后利用吸引子存在性定理,采用半群方法证明了方程整体吸引子的存在性.     

质谱法分析2型糖尿病大鼠神经视网膜差异表达蛋白

采用二维电泳(2DE)分离了正常SD大鼠和2型糖尿病模型大鼠神经视网膜组织总蛋白, 并用Image Master 5.0软件分析比较了正常组和糖尿病组2DE图像, 正常组检测到 3122±37(n=3)个蛋白质点; 糖尿病组检测到2702±21(n=3)个蛋白质点. 约150个蛋白质点的表达水平在两组之间存在明显差异(P<0.05). 在糖尿病组中表达上调的点68个, 下调的点82个. 选择20个差异表达蛋白质点进行肽质量指纹谱(PMF)或串联质谱鉴定, 其中7个蛋白已有报道与糖尿病视网膜病变(DR)相关, 10个蛋白尚未见有报道.     

兔肌肌酸激酶的分离纯化及部分性质的测定

本文提供了一个新开出的生物化学综合性教学大实验,完成本实验约需４天。实验中用到了多种重要的生化实验技术,对全面训练学生的实验能力有重要价值