PSYCH OpenIR  > 认知与发展心理学研究室
生物运动识别的神经机制
其他题名Neural mechanisms for the recognition of biological motion
张波
学位类型博士
导师张弢
2017-10
学位授予单位中国科学院研究生院
学位授予地点北京
学位专业认知神经科学
关键词生物运动 Mt/mst 猕猴 Neoanalysis 电生理数据分析工具包
摘要

     生物体在空间中的整体性移动行为被称为生物运动。识别生物运动是神经系统一种与生俱有、毕生稳定且跨物种的基本能力。它不仅关乎自然界中个体的生存,对包括人类在内的灵长类生物的社会交互也有重要的作用。对生物运动的行为学研究已经比较深入。研究发现,PLD(point-light displays)类型的生物运动刺激尽管只由若干置于主要关节处的运动光点组成,被试却可以从这种极为简化的刺激形式中识别出行走方向、性别、情绪状态、身份特征等丰富的信息。和面孔识别类似,生物运动识别也存在倒置效应,甚至刚出生两天的婴儿就表现出对正立生物运动刺激的偏好。尽管近年来神经影像学和脑损伤案例ᨀ供了若干可能参与生物运动信息加工的候选脑区,尤其是位于颞上沟(STS,superior temporal sulcus)的中侧颞叶区(MT,middle temporal area)和中上颞叶区(MST,medial superior temporal area)很可能在生物运动信息加工中起着重要作用,但目前对于生物运动信息加工的神经编码机制却几乎一无所知。我们采用清醒猴神经电生理胞外记录技术,在猕猴观看生物运动刺激的同时,用微电极记录大脑皮层颞上回的 MT 和 MST 功能区神经元的电活动,以研究视觉系统背侧通路中级皮层 MT 区和高级皮层 MST 区加工生物运动信息的神经机制。结果发现,MT 和 MST 脑区的神经元都能被 PLD 类型的生物运动刺激所激活,表现为刺激呈现期间,神经元的发放强度增强。但 MT 并不能像 MST 那样能有效区分生物运动的不同特征信息(form:intact vs scramble;inversion:upright vs inverted) 。因此,MST才是特异性参与生物运动信息加工的关键脑区,而 MT 不是。此外,我们还对 MST 脑区神经元对生物运动刺激不同特征信息的区分能力与其对不同光流模式的区分能力进行了相关性分析。结果发现,MST脑区神经元对生物运动刺激form特征信息(intact vs scramble)的区分与其对 radiation 光流(contraction/expansion)的区分达中等程度相关(r=0.46,p<0.001),对生物运动刺激 inversion 特征信息(upright vs inverted)的区分与其对 rotation 光流(clockwise/counter-clockwise rotation)的区分能力达中等程度相关(r=0.46,p<0.001)。这表明,MST 的神经元之所以能够区分生物运动的不同特征信息很可能源于该脑区神经元对不同光流模式(contraction/expansion,clockwise/counter-clockwise rotation)的选择性发放。
    此外,在电生理实验数据处理过程中,我们开发了一款基于 Python 的电生理数据分析平台,NeoAnalysis,用于电生理数据的快速处理和分析。该平台能够导入不同设备记录的数据,并对数据格式进行标准化。它ᨀ供了交互性好的图形用户界面用于线下动作电位自动/手动检测、自动/手动分类和模拟信号滤波。其分析功能涵盖spike train analysis、local field potentials analysis和行为反应分析(比如saccade 的检测和ᨀ取),可进行数据绘图和统计计算,并且分析过程中能够自动按实验参数设置对数据进行分组,不需要使用者ᨀ前进行复杂的预处理工作。最后,NeoAnalysis 还支持群体水平的数据分析。该数据分析平台覆盖了电生理实验数据分析处理的整个流程,功能强大,易于使用,值得在行业内推广。

其他摘要

   The detection and interpretation of biological motion are essential for the survival of individual and particularly for the social interaction of primates. In the past few decades, a large number of behavior studies have been done to investigate biological motion. Traditionally, the biological motion that consists of only a few light-spots was used as stimulus. Even so, a lot of information can be extracted from such stimulus including walking direction, gender, mood state, and identity. It has also been shown that biological motion recognition exists inversion effect, meaning subjects can distinguish upright biological motion from its inverted state. In recent years, neuroimaging studies and brain injury cases have pointed out several candidate brain regions (e.g. MT and MST) that might participate in the processing of biological motion. However, the neuronal mechanisms for the encoding of biological motion are barely known so far. Here, we use extracellular recording techniques to study how biological motion is processed in areas MT and MST of the dorsal visual pathway in awake monkeys. Specifically, neuronal activities in MT and MST were recorded using microelectrodes when the monkeys were watching biological motion stimulus. Our results showed that neurons in both MT and MST could be activated by biological motion stimuli that consist of light-spots. However, only neurons in MST could effectively distinguish the different characteristic features of biological motion including form (intact vs. scramble) and inversion (upright vs. inverted). Therefore, we propose that MST, but not MT, is the key region participating in the processing of biological motion. Furthermore, the correlation analysis showed that the discrimination of MST neurons in the form (intact vs. scramble) of biological motion was related to the discrimination in radiating (contraction/expansion)  optical flow (r=0.46, p<0.001), and the discrimination in the inversion (upright vs inverted) of biological motion was related to the discrimination in rotating (clockwise/counter-clockwise) optic flow (r=0.46, p<0.001). This suggests that neurons in MST are capable of responding to different optic flow patterns (contraction / expansion, clockwise / counter-clockwise rotation) might be the underlying reason why MST can distinguish different biological motion features.
    Besides, we developed a Python-based open-source toolbox, referred to as NeoAnalysis, for electrophysiological data processing and analysis. This toolbox supports importing raw data recorded by various data acquiring systems and converts them to a standardized format with a well-defined data structure. The functions provided by NeoAnalysis cover the entire workflow of electrophysiological data processing and analysis, which include spike detection, offline spike sorting, spike train and local field potentials analysis, and behavioral response (e.g. saccade) detection and extraction. Furthermore, NeoAnalysis provides user-friendly graphic user interfaces for spike sorting and signal filtering, and all its analysis functions take into account the automatical sorting of experimental conditions, which frees users from the burden of data preprocessing. In addition, NeoAnalysis also supports analysis at the population level. NeoAnalysis is a powerful toolbox for users doing electrophysiological experiments and is worth distributing in the field.

语种中文
文献类型学位论文
条目标识符http://ir.psych.ac.cn/handle/311026/22238
专题认知与发展心理学研究室
作者单位中国科学院心理研究所
推荐引用方式
GB/T 7714
张波. 生物运动识别的神经机制[D]. 北京. 中国科学院研究生院,2017.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
张波-博士学位论文.pdf(12272KB)学位论文 开放获取CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[张波]的文章
百度学术
百度学术中相似的文章
[张波]的文章
必应学术
必应学术中相似的文章
[张波]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。