具有离散性的连续性模型映射

为了实现在连续性空间中的离散分段问题,采用解决离散性问题的方法虚拟构造连续性集合,将连续性问题映射为分段的离散问题,根据集合中元素的离散特性实现连续性模型的分段求解.通过数据结构的设计与压缩路径的算法证实,模型的映射能够解决实际问题的分段求解.     

基于投影方法的约束独立成分分析

独立成分分析是解决盲源分离问题的一种有效工具，但ICA具有伸(dilation)与排序(permutation)的不确定性的本质特征。本利用一些约束条件，采用Lagrange乘子法并结合简单的投影方法，可以以特定的形式来进行独立成分的排序，并且可以在信号分离过程中规范化解混矩阵(demixing matrix)，能够系统地减轻ICA对于伸缩与排序的不确定性。仿真结果证实了算法的有效性。     

Application of Small-World Measures to Multichannel Event-Related Potential Activity during Generation of Global and Local Imagery

The Small world model has been successfully used to explore the abnormal pattern of brain information processing in some neuropsychiatric diseases, but not engaged in the study of cognitive functions. We apply the small-world measures： the clustering coefficient and average path length, to evaluate multi-channel event-related potential activity during the generation of global and local imagery. Results show that the brain functional networks of the global and local imagery generation are both small-world ones. In addition, the local imagery generation has a larger clustering coefficient, while the global imagery generation has a shorter average path length. These results support the global precedence in the global-local imagery generation, and reflect the different processing modes in which global imagery emphasizes particularly on global integration, while local imagery on local specialization. Our results indicate that small-world measures could be applied to quantify the difference of brain activities in different cognitive tasks, and further provide some explanations for cognitive behavior.     

基于肿瘤微环境响应的DNA纳米结构递药系统

DNA分子由于其独特的生物相容性和可编程性,在增强药物靶向性和降低药物毒性方面展现了独特的优势和巨大的潜力。随着人们对肿瘤微环境研究的深入和环境响应性的DNA触发器的研制,近些年已报道了许多基于肿瘤微环境响应的DNA纳米结构递药系统,这些DNA纳米结构递药系统结合了纳米运载工具良好的生物分布和药代动力学特性,以及小型药物载体的快速扩散和渗透特性。通过靶向广泛的肿瘤栖息地而不是肿瘤特异性受体,该策略有可能克服肿瘤异质性问题,并可用于设计诊断和治疗多种实体肿瘤的纳米颗粒。在体内能够稳定地转运,在肿瘤组织独特的微环境刺激下释放药物,能有效地控制药物释放部位和释放速度,极大地降低了肿瘤治疗的毒副作用。本文主要从pH响应型、GSH响应型、ATP响应型、酶响应型、抗原响应型五个方面,综述了基于肿瘤微环境响应的DNA纳米结构递药系统的最新研究进展,分类介绍了这些DNA纳米载体的设计策略和响应释放机制,此外,还重点介绍了该领域面临的前景和挑战。     

Modelling Human Cortical Network in Real Brain Space

Highly specific structural organization is of great significance in the topology of cortical networks. We introduce a human cortical network model, taking the specific cortical structure into account, in which nodes are brain sites placed in the actual positions of cerebral cortex and the establishment of edges depends on the spatial path length rather than the linear distance. The resulting network exhibits the essential features of cortical connectivity, properties of small-world networks and multiple clusters structure. Additionally, assortative mixing is also found in this model. All of these findings may be attributed to the specific cortical architecture.