【作者】 郑金龙；

【导师】 舒斯云；

【作者基本信息】 南方医科大学 ， 神经生物学， 2008， 博士

【摘要】 研究背景和目的人类大脑皮质在高级认知功能中的作用似乎不可替代,但临床皮质下结构损伤的病例同样出现相关高级认知功能障碍,提示皮质下结构也可能是一些高级认知功能的重要神经基础。大脑皮质下结构主要包括以纹状体为主的基底核及其附近的大脑内部灰质团块。皮质下结构功能的研究既是一个老课题又是一个崭新的课题。由于人类大脑皮质的高度发达,纹状体等皮质下结构已退居皮质下中枢的地位,其主要功能是对运动及其技巧的协调。但不断涌现的证据表明皮质下结构也参与了一些高级认知功能。被认为是学习记忆新区的纹状体边缘区(marginal division of the striatum,MrD,简称边缘区)参与了联合型学习、陈述性记忆和数字工作记忆。学习记忆是大脑重要的高级认知功能之一,它一直是神经科学研究领域的热点和难点。记忆的神经基础的研究普遍集中于大脑皮质,而对皮质下结构是否参与记忆认知功能的关注却较少,即使关注也主要着眼于与运动学习和习惯形成有关的非陈述性记忆(Non-declarative memory)。Baddeley在短时记忆的基础上提出并完善了工作记忆(Working memory,WM)的概念。研究表明许多大脑皮质在工作记忆过程中发挥着重要作用,尤其是前额叶皮质表现得尤为突出。长期以来一直认为陈述性记忆(Declarative memory)的神经基础是颞叶海马记忆系统,纹状体似乎只与非陈述性记忆有关,而且有关皮质下结构在工作记忆中的作用知之甚少。语义记忆是一种重要的陈述性记忆,是关于概念和事实的一般知识的记忆。几乎所有的语义记忆的神经基础研究都聚焦于大脑皮质和语义记忆的提取阶段,鲜有涉及皮质下结构和语义记忆的编码阶段;同时,语义记忆的功能脑区的研究均与语义记忆各类型的特殊性有关,而不是语义记忆的普遍性。空间知觉信息指的是物体的位置、朝向、远近、运动方向等外部属性信息。大量研究表明空间记忆信息的处理几乎责任于大脑皮质,而且不同皮质脑区对空间记忆信息处理的不同作用尚存在争议。然而,许多病位不在大脑皮质的疾病如帕金森氏病(Parkinson’s disease,PD)和亨廷顿病(Huntington’s disease,HD)也同样出现视空间障碍,说明人脑处理空间信息不仅与大脑皮质有关,而且还可能与皮质下结构有关。学习记忆是复杂的高级认知功能,研究手段大多具有创伤性。利用创伤性技术很难观察到活体脑的工作过程,特别是各种行为状态时脑内的变化。随着现代科学技术的发展,神经影像学技术在神经科学的科研和临床中得到广泛运用,尤其是功能影像学的运用,为研究学习记忆的神经基础提供了新的有力的技术手段,对探求其确切的机制赋予了希望。功能磁共振成像(Functional magnetic resonance imaging,fMRI)的基本原理是结合了血氧水平依赖性(Blood oxygen level dependent,BOLD)原理和回波平面成像技术(Eecho planar imaging,EPI),无创性fMRI融合了MRI(Magnetic resonance imaging,MRI)和EPI技术中的高时间、空间分辨率和高信噪比的成分,可以同时获得形态和功能图像,已在认知神经科学研究领域中得到关注和应用。fMRI技术不仅能定位学习记忆的神经心理活动的脑功能区,而且可以同时观察多个脑区的活动,从而可以探讨各个功能区域之间的相互关系,为进一步探明学习记忆的机制等提供相关答案。因此,通过fMRI技术,可以帮助我们进一步研究大脑皮质下结构是否参与各类型的学习记忆以及皮质与皮下结构在学习记忆认知功能中的相互关系等提供广阔的前景。我们运用fMRI技术结合神经心理学方法,研究语义工作记忆(Semanticworking memory)及空间工作记忆(Spatial working memory,SWM)在人脑中相应功能区的激活情况,旨在明确人脑皮质下结构是否参与了语义及空间工作记忆以及皮质和皮质下结构在语义和空间学习记忆认知过程中的相互作用;揭示语义工作记忆编码和提取不同阶段的神经基础,探讨语义工作记忆的一般共同神经机制;进一步明确空间工作记忆的各功能脑区的作用。对皮质下结构,特别是纹状体的记忆认知功能的进一步研究将有助于我们对皮质下、老年性痴呆疾病,特别是HD、PD等疾病所致的认知功能障碍的发病机制和防治的研究提供新的线索和科学理论依据。材料和方法我们选取语义分类、配对词语和空间位置作为学习记忆材料,对12和16名分别参加两种不同的语义工作记忆任务即语义分类工作记忆任务和词语联想学习记忆任务以及10名参加空间工作记忆任务的右利手健康成人志愿者(年龄范围均为20-23岁)在执行任务作业的同时,分别进行脑功能磁共振扫描。实验采用组块设计(Block design)模式即记忆任务-对照任务,记忆任务与基线对照任务交替进行,每个相关任务均包含四个循环。fMRI数据分析:采用统计参数图(Statistics parameter mapping 99,SPM99,welcome Dept of Cognitive Neurology;http://www.fil.ion.ucl.ac.uk/spm)软件对纳入的被试者的fMRI数据进行分析和脑功能区定位。单个被试者的数据分析:SPM软件对fMRI图像数据预处理后,通过信号反应曲线的相关分析来检验不同记忆任务下激活部位的差异显著性,即将循环中相同记忆任务的数据叠加减去相对应的基线对照叠加数据,得到与相应记忆任务直接相关的脑区激活数据。将语义工作记忆任务和空间工作记忆任务统计阈值概率分别设为P≤0.005和P≤0.001,激活范围的阈值设定为10或5个以上象素,即连续激活象素达到10或5个以上的区域为激活区,得出激活范围和最大激活强度。组分析:通过SPM99软件将纳入的被试者同一任务的数据一起分析,统计阈值概率设定与单个被试者数据分析一致,将获得平均激活图叠加于Texas大学的Talairach标准三维脑模板上,对脑的激活区进行定位,得到感兴趣区(Regions of interest,ROI)的激活强度、像素值和中心坐标等。结果一、语义分类工作记忆任务编码时激活的脑区:额叶(Brodmann area,BA,布鲁德曼分区,下同。BA6/9/46/47)、颞枕叶(BA19/37)、顶叶(BA40/7)和岛叶(BA13);大脑皮质下结构有尾状核、屏状核和丘脑;另外,小脑有不同程度的激活;其中左侧额叶中回BA6/9/46被激活的强度和像素范围最显著。任务提取时激活的脑区:额叶(BA6/9/46/47)、扣带回(BA32)、顶叶(7/40/39)、枕叶(BA17/18/19)和岛叶(BA13);激活的皮下结构包括纹状体及其边缘区、屏状核和丘脑;另外,小脑和皮质下白质也有不同程度的激活;其中左侧额上回BA6被激活的强度和像素范围最显著。两个阶段所激活的皮质脑区都有极其明显的左半球优势。二、词语联想学习记忆任务编码时激活的脑区:枕叶(BA18/19)、额叶(BA6/9/46/47)、顶叶(BA7/39/40)和颞叶(BA20/21/37);大脑皮下结构有纹状体及其边缘区和丘脑;另外,小脑和皮质下白质有不同程度的激活;其中左侧枕叶BA18/19被激活的强度和像素范围最显著。任务提取时激活的脑区:顶叶(BA7/40)、额叶(BA6/8/9/10/46/47)和枕叶(BA19);激活的皮下结构包括纹状体及其边缘区、红核和黑质;另外,小脑和白质也有不同程度的激活;其中左侧顶上小叶BA7被激活的强度和像素范围最显著。两个阶段所激活的皮质脑区都有极其明显的左半球优势。三、两种不同的语义工作记忆任务在编码和提取的不同阶段所激活的脑区表现出相似的一些变化:1.编码阶段被显著激活的皮质脑区集中于左侧皮质的腹外侧和背外侧中下部,而提取阶段被显著激活的脑区却转移到左侧皮质的背外侧部,2.激活的皮质脑区的强度和范围,编码阶段比提取阶段显著;3.激活的皮下结构的强度和范围,提取阶段比编码阶段显著。四、比较两种不同的语义工作记忆任务所激活的共同脑区有:额叶中下回(BA6/9/46/47),额上回(BA6),顶上小叶(BA7),缘上回(BA40),颞枕交界区(BA37/19),纹状体、丘脑和小脑。所激活的皮质共同脑区有明显的左半球优势,且在外侧裂语言带周围形成弧形的激活带。五、空间工作记忆任务激活脑区:双侧顶叶的楔前叶、顶上小叶、缘上回(BA7/40),双侧额上中下回(BA6/9/47),双侧枕叶和颞叶梭状回(BA19/37),右侧海马旁回(BA30)和右侧扣带回(BA25);被激活的大脑皮质下结构有右侧屏状核、右侧尾状核、左侧丘脑和左侧黑质;同时,双侧小脑均被明显激活。所激活的强度和范围以双侧顶叶最为显著,左侧顶叶与双侧额叶以及右侧顶叶与左侧额叶之间的激活强度差异均有显著性意义(P＜0.05)。结论一、语义工作记忆任务是大脑皮质和皮质下结构等一起共同参与协作完成的,皮质下结构参与了陈述性记忆和工作记忆。二、纹状体边缘区参与了语义工作记忆,进一步证实了纹状体边缘区是学习记忆的脑区。三、语义工作记忆编码和提取阶段所依赖的脑神经基础不同,编码时依赖于左半球皮质腹外侧和背外侧中下部,而提取阶段则依赖于左半球皮质背外侧和皮质下结构,皮质下结构在语义工作记忆的提取阶段起重要作用,是语义工作记忆神经环路中的重要组成部分。四、语义工作记忆有其共同的神经基础:主要包括大脑左侧皮质腹外侧和背外侧中下部的颞枕叶的BA37/19、额中下回BA46/9/6/47和左侧皮质背外侧的额上回BA6、顶叶的BA7/40,纹状体、丘脑和小脑,皮质的共同脑区恰好位于传统优势半球外侧裂语言区外围。五、语义工作记忆可能存在着其特有的神经环路:左侧皮质腹外侧和背外侧中下部接受神经冲动,分析并编码语义信息,皮质下结构输出语义信息到左侧皮质背外侧,从而完成语义信息的提取。六、空间工作记忆任务也是大脑皮质和皮质下结构等一起共同参与,协作完成的,人脑顶叶是空间材料工作记忆的关键脑区,而不是前额叶。更多还原

【Abstract】 Background and objectivesThe cortex of human brain seems to play exclusive role in high cognitive functions. However, patients with subcortical structures lesions appeared to be impairment in high cognitive functions, suggesting that subcortical structures may play an important role in some high cognitive functions. The subcortical structures are mainly consisted of the basal ganglia and other subcortical gray matter. Studies on the roles of the subcortical structures were always an interesting and hot topic. The human striatum, as main parts of the basal ganglia, developed to be a subcortical central region on motorial effects due to the high development of the cerebral cortex. However, there is growing evidence suggesting that the subcortical structures, such as striatum, are also related to the high cognitive function in human brain. A newly discovered learning and memory area in the brain, termed marginal division (MrD) of the neostriatum, had been confirmed to be associated with the declarative and digital working memory. As one of the important high cognitive function, learning and memory had always been foci in neuroscience research area. Researches on the neural basis of memory mainly focused on cerebral cortex; comparatively there wss only limited studies investigating the roles of these regions in memory. The majority of these limited studies generally focused on non-declarative memory. Baddeley first introduced the concept of working memory based on short-term memory. A large number of studies indicated that the cerebral cortex, especially the prefrontal lobe, played an important role in manipulating the information of working memory. The neural basis of declarative memory had been examined to rely on medial temporal system for a long time. The striatum seemed to be related only to non-declarative memory rather than declarative memory, and little was known about the role of subcortical structures associated with working memory. Semantic memory, an important declarative memory, refers to our general knowledge of concepts and facts, which are quite different from the specific memory of personal experiences. Nearly all studies of semantic memory were focused on the semantic retrieval phase of the brain but not on the encoding phase of the brain, and observations of semantic memory were described in relation to the functional specialization but not to the functional integration of the brain. Spatial information refers to the external information of the objects including the site, the direction, the distance, and so on. Nearly all reports of spatial memory suggested that the cerebral cortex were responsible for processing the information of spatial memory, and the role of the different cortical regions contributed to spatial memory is still a matter of debate. However, the subcortical structures diseases, such as Parkinson’s disease and Huntington’s disease, may develop the visual spatial disorder during the early stage of disease, suggesting the subcortical structures probably subserve to spatial memory. Learning and memory is complicated high cognitive function; most techniques of research on it were invasive. It was always very difficult to examine the working process in vivo brain by invasive techniques. With the development of modern science and researches, neuroimaging techniques, especially functional neuroimaging techniques, have been applied extensively in research and clinic of the neuroscience area. Functional magnetic resonance imaging (fMRI) is one of the non-invasive functional neuroimaging techniques, which has been applied in neuroscience research area since its invention in the early 1990s. Its basic principles are related to the physiology of the blood oxygenation level-dependent (BOLD) contrast mechanism and to the acquisition of functional time-series with echo planar imaging (EPI). fMRI has rapidly assumed a leading role among the techniques used to localize brain activity. The spatial and temporal resolution provided by state-of-the-art MR technology and its non-invasive character, which allows multiple studies of the same subject, are some of the main advantages of fMRI over the other functional neuroimaging modalities that are based on changes in blood flow and cortical metabolism. fMRI can offer a good prospect to investigate the fundamental mechanism of learning and memory.The present study attempted to investigate the collaborative activities of the cortical and subcortical structures during the semantic and spatial working memory processing by fMRI study, to examine the common neural network in the brain underlying two different semantic working memory tasks (the semantic category working memory task and the paired-word associative learning and memory task), to compare the neural basis for different phases of semantic working memory, to make sure about the role of different brain regions involved in spatial working memory processing. Study on the role of the subcortical structures in human high cognitive function helps us further understanding of fundamental mechanism related to cognitive impairment and dementia, specifically concerned with the dementia in basal ganglia disorders such as Parkinson’s disease or Huntington’s disease. Materials and MethodsWe selected semantic category, paired word, and spatial site as study materials. Twelve and sixteen right-handed healthy volunteers (20-23 years old) were recruited to participate respectively in the semantic category working memory task and the paired-word associated learning and memory task, while ten right-handed healthy volunteers (20-23 years old) participated the spatial working memory task. Functional imaging was carried out by a 1.5 Tesla Magnetom scanner while healthy volunteers performed the memory tasks. A control task was performed for the block-design in each test. Memory task and baseline were arranged alternatively and each task included four cycles. SPM99 (Welcome Department of Cognitive Neurology; http://www.fil.ion.ucl.ac.uk/spm) was used to analyze the fMRI experimental data and to identify the activated brain regions. Images from the first 6 seconds of acquisition of each session were removed from further functional data processing to minimize the transit effects of hemodynamic responses. The preprocessing of fMRI data, including the slice time correction, the realignment, the coregistration, the normalization and the smoothing, were preformed. Activation maps were generated by using a cross-correlation method, where the activity of each pixel was correlated to a boxcar function that was convolved with the canonical hemodynamic response function. Subject-specific linear contrasts, including the encoding condition versus baseline and the retrieval condition versus baseline for each of the effects of interest, were assessed. These contrasts were entered into a standard SPM second-level analysis treating subjects as a random effect, using a one-sample t-test. The expected mean difference value for the t-test was set to zero. A voxelwise intensity threshold (P≤0.005 or P≤0.001) and a spatial extent (10 or 5 voxels as the minimum cluster size) were set for multiple comparisons. All coordinates reported were in Talairach space.Results From the above design of experiments, the results were showed as the following: 1. Brain regions of activation in both cerebral cortex and subcortical structures were observed in the encoding and retrieval phases of the semantic category working memory task. During the encoding phase of this memory task, the cerebral cortex including the middle gyrus (BA46/9/6), the superior gyrus (BA6) and the inferior gyrus (BA47) of the frontal lobe, the superior parietal lobule (BA7) and the supramarginal gyrus (BA40) of the parietal lobe, the fusiform gyrus (BA37) of the temporal lobe, the insula (BA13) and the occipital lobe (BA19) were activated with left hemisphere predominance. While the subcortial structures including caudate, claustrum, and thalamus were activated. Other brain regions such as cerebellum were also activated. The most striking activation region was the left middle frontal gyrus (BA46/9/6). During the retrieval phase of this memory task, the cerebral cortex including the frontal lobe (BA46/9/6/47), the cingluate gyrus (BA32), the parietal lobe (BA7/40/39), the occipital lobe (BA17/18/19) and the insula (BA13) were activated with left hemisphere predominance. While the subcortical structures including striatum and its MrD, claustrum, and thalamus were activated. The activations of the cerebellum and some sub-gyral white matter were also observed. The most striking activation region was the left superior frontal gyrus (BA6). 2. Brain regions of activation in both cerebral cortex and subcortical structures were observed in the encoding and retrieval phases of the paired-word associative learning and memory task. During the encoding phase of this memory task, the cerebral cortex including the occipital lobe (BA18/19), the frontal lobe (BA46/9/6/47), the parietal lobe (BA7/40/39), and the temporal lobe (BA20/21/37) were activated with left hemisphere predominance. While the subcortical structures including striatum and its MrD, and thalamus were activated. The activations of the cerebellum and some sub-gyral white matter were also observed. The most striking activation region was the left the occipital lobe (BA18/19). During the retrieval phase of this memory task, the cerebral cortex including the frontal gyrus (BA46/9/6/8/10/47), the parietal lobe (BA7/40) and the occipital lobe (BA18/19) were activated with left hemisphere predominance. While the subcortical structures including striatum and its MrD, the substantia nigra and the red nucleus were activated. The activations of the cerebellum and some sub-gyral white matter were also observed. The most striking activation region was the left parietal lobe (BA7). 3. The activated brain areas displayed some similar alterations from encoding phase to retrieval one of the two memory tasks. The stronger and major foci activated areas in cerebral cortex from the left ventrolateral and mid-dorsolateral regions during encoding phase transferred to the left dorsolateral regions during retrieval phase. The intensity and extent of activated regions in the cerebral cortex during encoding phase were stronger than those during retrieval phase, whereas the opposite activation pattern was found for the subcortical structures during the two phases. 4. The major significant activated brain regions for both memory tasks were overlapped, including the middle and superior frontal gyrus(BA6/9/46), the inferior frontal gyrus(BA47), the superior parietal lobule(BA7), the supramarginal gyrus(BA40) and the occipitotemporal(BA37/19) region, the striatum, the thalamus and the cerebellum. The cerebral cortex was activated with a strong left lateralization during the two tasks, and the common activated areas of the two different semantic working memory tasks in the cerebral cortex formed an arcuate area surrounding the perisylvian language cortex zone.5. Both cortex and subcortical structures were activated in the spatial working memory task. The brain cortex areas including the bilateral precuneus lobules and superior lobules as well as supramarginal gyrus of the parietal lobe (BA7/40), the bilateral superior and middle as well inferior gyrus of the prefrontal lobe (BA6/9/47), the bilateral occipitotemporal lobes (BA19/37), the right parahippocampal gyrus (BA30) and the right cingulated gyrus (BA25) were activated during the task. The subcortical structures including the right caudate of the neostriatum, the right claustrum, the left thalamus and the left substantia nigra in midbrain were activated during the task. Meanwhile, the bilateral cerebellums were also prominently activated. The most striking activation region was the parietal lobe (BA7/40). The differences of the intensity between the left parietal lobe and the bilateral frontal lobes, the right parietal lobe and the left frontal lobe were significant (P<0.05) .ConclusionsFrom the above results of this study, we can draw the conclusion as the following: our study demonstrate that the information of the semantic working memory in human brain is manipulated by the collaborative activities within cortex, and between cerebral cortex and subcortical structures, and the subcortical structures are involved in declarative memory and working memory. MrD is further confirmed to be a new area associated with learning and memory by this study because it is activated during the semantic working memory tasks. The encoding and retrieval phases of the semantic working memory depend on different neural foundations in human brain. The executions of the semantic memory mostly rely on the left ventrolateral and mid-dorsolateral cortical regions during the encoding phase, whereas it primarily depend on the left dorsolateral cortical regions and the subcortical structures during the retrieval phase. Subcortical structures may play an important role in the retrieval phase of the semantic working memory task, and also is a key component of the semantic working memory neural circuit. The results indicate that there is a general neural network contributing to the semantic working memory. The network include the left superior, middle, inferior frontal gyri (BA46/9/6/47), left superior parietal lobule (BA7), left supramarginal gyrus (BA40), left occipitotemporal region (BA19/37), striatum, thalamus and cerebellum. The common activated cortex areas form an arc surrounding the perisylvian language cortex zone with strong left lateralization. There is a neural circuit of semantic working memory as the following: the ventrolateral and mid-dorsolateral cortical regions receive the impulses, analyze and encode the information; the subcortical structures, mainly the basal ganglia, deliver the information to the dorsolateral cortical regions to accomplish semantic information retrieval. The subcortical structures as well as the cerebral cortex contribute to the spatial working memory, and the human parietal lobes are crucial brain regions in manipulating information of the spatial working memory rather than the frontal lobes.更多还原