Frame Invariance and Scalability of Neural Operators for Partial Differential Equations

Partial differential equations (PDEs) play a dominant role in the mathematical modeling of many complex dynamical processes. Solving these PDEs often requires prohibitively high computational costs, especially when multiple evaluations must be made for different parameters or conditions. After training, neural operators can provide PDEs solutions significantly faster than traditional PDE solvers. In this work, invariance properties and computational complexity of two neural operators are examined for transport PDE of a scalar quantity. Neural operator based on graph kernel network (GKN) operates on graph-structured data to incorporate nonlocal dependencies. Here we propose a modified formulation of GKN to achieve frame invariance. Vector cloud neural network (VCNN) is an alternate neural operator with embedded frame invariance which operates on point cloud data. GKN-based neural operator demonstrates slightly better predictive performance compared to VCNN. However, GKN requires an excessively high computational cost that increases quadratically with the increasing number of discretized objects as compared to a linear increase for VCNN.     

痛情绪相关神经递质及神经环路研究进展＊

痛情绪是指因疼痛引发的情绪和情感体验，是疾病过程中最常见的一种情绪。痛情绪相关神经机制非常复杂，但主要与单胺类神经递质、神经肽和某些神经环路有关，笔者将结合目前研究现状分别从以上两方面展开，就痛情绪相关单胺类神经递质和神经肽在受体分类、脑区通路、共疾病以及各神经递质之间的联系和痛情绪相关神经环路中各个蛋白的作用机制等方面进行探讨。     

中西医联合治疗糖尿病周围神经病变效果观察

目的研究中西医联合治疗糖尿病周围神经病变患者的临床效果。方法按照随机数字表法将2017年7月-2018年6月期间本院收治的72例糖尿病周围神经病变患者分为两组,对照组36例采用常规西医治疗,观察组36例采用中西医联合治疗。将两组的神经传导速度、临床疗效、用药安全性进行比对。结果观察组糖尿病周围神经病变患者治疗后正中神经、尺神经的MNCV、SNCV均快于对照组,临床总有效率高于对照组,P<0.05;两组的不良反应发生率进行比较差异无统计学意义,P>0.05。结论中西医联合治疗糖尿病周围神经病变患者具有较高的有效性和安全性。     

基于神经网络和支持向量机的中医体质辨识模型研究＊

目的 提出基于舌象和形体特征的中医体质辨识模型，探讨中医以舌辨质客观化、规范化方法。方法 提取客观化的舌象特征（舌质与舌苔的色度和饱和度、舌苔纹理的平均亮度、平滑度以及齿痕特征）和形体特征，建立基于神经网络和支持向量机的客观化舌象和形体特征的辅助中医体质识别模型。结果 对平和质、气虚质、阴虚质和气淤质四种体质进行模型的训练和测试，结果表明融合舌象特征和形体特征的中医体质辨识模型能有效地辅助中医体质识别，且支持向量机对四种体质辨识效果总体上优于神经网络。结论 基于客观化的舌象特征和形体特征辅助中医体质辨识有利于提高中医体质辨识的客观化水平，合理选择机器学习算法可以提高中医体质辨识的准确性。     

基于粒子群优化的BP神经网络对血脂异常气郁痰阻型中医证候的疗效预测＊

目的 选用两种药物分析对血脂异常、中医证候的疗效，以中医证候积分为基础建立血脂异常气郁痰阻型的疗效表征模型，进而提出基于粒子群优化的BP神经网络预测两种方法的疗效。方法 收集50名符合血脂异常气郁痰阻型患者的基本资料，将患者随机分为解郁祛痰化浊方组（JQHP）、血脂康胶囊组（XZKC），每组25例，观察疗程12周治疗前后的血脂水平，治疗前及治疗期间每4周记录一次中医证候积分，共4次。定性疗效分析参考血脂防治指南和中医疗效积分，采用t检验、非参数检验及X²分析的方法；神经网络模型采用粒子群优化的BP神经网络，定量构建血脂异常气郁痰阻证的预测模型，设置为三层网络，输入层节点数为15，隐含层节点数为6，输出层节点数为1。针对JQHP组和XZKC组的数据，分别选择80%的数据进行训练，20%的数据进行验证。结果 两组治疗前、治疗4周、治疗8周、治疗12周的中医证候积分随治疗周期的增加而下降（P < 0.05）。JQHP组除了“形体肥胖”外，其余14项症状均有显著的改善（P < 0.05），XZKC组治疗前后比较除“形体肥胖”、“头晕”、“胸闷”外，其余12项均改善（P < 0.05）。以中医证候积分建立的基于粒子群优化BP神经网络的预测精度高、稳定性佳，可以定量预测两种药物对未知患者的疗效。结论 JQHP和XZKC可以改善血脂异常患者的血脂水平，改善气郁痰阻型患者的症状，此为定性分析。利用中医证候积分建立的基于粒子群优化BP神经网络两种治疗方法的预测模型，可定量预测未知患者的疗效。以此思路将有助于找到一种特定治疗方法的适应人群，从而提高该治疗方法的针对性，实现中医药的智能化、精准化、定量化。     

A case of sudden cardiac death due to mitochondrial disease

A 25-year-old, emaciated man without medical treatment was found to have died suddenly at home by his mother. At autopsy, there were no injuries to his body, but significant circulatory insufficiency was observed. Electron microscopy revealed abnormal mitochondria in cells of the cardiac conduction system. The conduction system was filled with mitochondrial size abnormalities and mitochondrial cristae abnormalities. No notable abnormal findings were observed in other organs. Genetic examination of the blood revealed the mitochondrial pathogenetic variant m.3243A>G. Epileptic seizures, diabetic ketoacidosis, and hyperosmolar hyperglycemic state were unlikely to be the cause of sudden death. The cause of death was diagnosed as arrhythmia possibly induced by the failure of the cardiac conduction system due to mitochondrial disease. This is a rare case of sudden death caused by an accumulation of abnormal mitochondria in the cardiac conduction system.     

Roles for Event Representations in Sensorimotor Experience,Memory Formation,and Language Processing

The unifying theme for the papers in this volume is that event‐predictive representations play an important role in cognitive processes, and they are a particularly fruitful object of study for cognitive scientists. In this paper, we present our own model of event representations that draws together several disparate strands of research into event representations, relating to event perception, event encodings in working memory and their role in prediction, and events as they are reported in language. Our main claim is that the notion of deictic routines originally developed by Ballard et al. (1997) is of great use in explaining how these different types of event representation interface with one another.     

电针对脑缺血大鼠侧脑室室管膜下区巢蛋白表达的影响

目的观察电针对脑缺血大鼠侧脑室室管膜下区巢蛋白(Nestin)表达的影响,探讨不同针刺处方防治脑缺血作用机制。方法大鼠随机分为假手术组、模型组、电针1组、电针2组。电针1组电针双侧“丰隆”(ST40)(2/100 Hz,1~3 mA),每天1次,每次30分钟,连续治疗7天;第8天,电针1、2组和模型组用50%FeCl 3滤纸贴敷大脑中动脉20分钟造成脑血栓模型,术后电针1、2组电针ST40、“百会”(GV20),每天1次,连续7天。术后1天利用神经功能缺损评分法(neurological deficit score,NDS)进行行为学检测,苏木素伊红染色观察缺血半暗带组织形态,免疫组化法检测缺血侧侧脑室室管膜下区Nestin的表达。结果模型组大鼠NDS升高(P<0.05),缺血半暗带神经细胞水肿明显,神经元核固缩,术后侧脑室室管膜下区Nestin面密度升高,持续至术后7天(P<0.01)。电针治疗后神经功能缺损评分降低(P<0.05),缺血半暗带神经细胞水肿减轻,细胞形态趋于正常,电针1、2组侧脑室背外侧角Nestin表达均增高,从术后1天(电针1组P<0.01,电针2组P<0.05)持续至术后7天(P<0.01),电针组侧脑室外侧壁Nestin表达在术后1天升高(P<0.01),术后7天升高差异没有统计学意义(P>0.05)。电针1、2组之间差异无统计学意义(P>0.05),但趋势上电针1组高于电针2组。结论电针可改善脑缺血后神经功能,减轻缺血性脑损伤,促进侧脑室室管膜下区神经干细胞增殖,促进脑损伤后神经修复,在脑缺血前进行电针治疗效果更佳。