【作者】 贡京京；

【导师】 苗丹民；

【作者基本信息】 第四军医大学 ， 应用心理学， 2008， 博士

【摘要】 面孔倒置效应的提出已经近四十年,但是对于它所揭示的意义到现在依然存在争论,有些研究者认为:倒置效应为面孔所独有,而且是面孔加工特异性的显著标志之一。虽然使用电生理、脑成像以及对脑损伤病人进行的神经心理学检查等手段,研究者已经找到选择性对面孔反应的脑区,然而,关于面孔的倒置效应是否其独有,且反应其特异性加工依然没有定论。还有研究者认为:面孔识别之所以特殊,是因为它要求在一个相对同质的物体范畴中,对亚范畴进行专家水平的分类,而人类是完成识别不同个体面孔的专家。因此,凡是对于具有专家技术的物类都可以出现倒置效应。因此,对于这个问题至今仍是争议的焦点。而作为世界上独一无二的汉字,许多心理学研究正围绕对它的视觉识别问题展开。与面孔认知类似,每一个有读写能力的成年人都可以说是识别汉字的“专家”,但作为一种古老的象形文字,汉字具有二维图形特征,因而具有独特的研究优势。已有研究证实在大脑存在着对语词特异加工的脑区部位——视觉语词形成区(Visual Word Form Area, VWFA),对语言有着高效加工功能。这表明了对于汉字的专家加工机制具有其解剖和功能基础。然而目前对于汉字的认知研究大多集中在语义、语音方面,而对于汉字本身的识别与认知,由于缺乏有效的实验方法和范式,研究相对不足,对于汉字字形本身的早期视觉加工研究相对缺乏。本研究试图通过对面孔和汉字倒置效应的行为学研究,深入探讨两者在认知加工机制上的异同,揭示倒置效应的本质;通过对面孔和汉字倒置效应的电生理学研究,进一步分析各个成分所代表的加工阶段和特征,并为深入研究其后续加工过程进行理论和实验的尝试与探索。本研究采用诱导激活的实验范式,比较正常被试在不同的诱导条件下对相同刺激反应的差异。本实验主要由两大部分构成:行为学研究和电生学研究(ERP)。利用“品”型汉字独特的结构与面孔相结合进行认知研究。在行为学研究中,四组被试分别在不同的诱导刺激下(面孔vs.字母、汉字vs.字母),在不同的认知条件下(倒置vs.旋转)对相同的比较刺激(品)的不同特征变化(结构特征vs.局部特征)进行认知加工,比较被试对比较刺激认知差异;第二部分为ERP实验,四组被试分别在不同的诱导刺激下(面孔vs.字母、汉字vs.字母),在不同的认知条件下(倒置vs.旋转)对相同的比较刺激(品)的方向进行判断,比较被试对相同刺激认知差异,并对其反应时、不同成分的潜伏期、波幅进行比较。通过本研究可以得到的主要结论如下:在本研究中,我们不仅在发现了面孔的倒置效应,而且在汉字组也发现了确切的倒置效应,同样的视觉刺激(比较刺激)在同样的任务中由于诱导条件不同,诱发了被试不同的认知加工策略,出现不同的反应模式。两种截然相对的倒置效应同时出现说明:倒置效应并非面孔效应所特有。虽然面孔的倒置效应是一个不依赖于实验范式的相对稳定的反应模式,但当被试认知任何具有共同结构和高度熟悉的刺激类型(如汉字),并经过长期的经验对其形成专家知识和技能时,都可以出现倒置效应,因此,面孔认知的倒置效应不能作为面孔认知特异加工的证据。同时,该结果充分支持了专家假说对倒置效应的解释。倒置面孔和汉字可能采取了一种不同于正立刺激的认知加工策略,在旋转120度左右出现认知策略的转换,反应时与角度的差异呈现非线性的关系。在本研究中,在面孔和汉字认知时,倒置效应在结构变化条件下显现和在局部变化条件下缺失支持了倒置时对两者结构性信息加工的破坏,从而说明了专家系统中结构性认知的作用。在Bruce-Young的模型中的直接视觉处理、结构编码、表情分析和面孔语言分析前还应有一个分类处理阶段,即在早期视觉加工阶段中的简单视觉分类阶段。面孔刺激的P1成分在面孔认知中并不是简单地反映低层次的视觉差异,而是反映了对面孔的整体认知,即将图形知觉为面孔。甚至部分地涉及了面孔结构加工。而汉字的早期知觉加工始自P1,以左侧脑区为主,在后续认知中双侧脑区共同作用。N170并非是面孔特异性成分,也并不是对面孔和非面孔刺激之间的低层次视觉变化的反映(Rossion, Gauthier, Tarr, Despland, Bruyer, Linotte, 2000)而是反映对多种物类具有的专家技术,即可以反映物类的专家式的加工。视觉刺激倒置可能主要影响面孔和汉字的潜伏期,而心理旋转则可能主要引起了波幅的改变。总之,面孔识别和汉字识别都是人们社会生活中的重要功能,本研究对两者的认知加工的异同的探索进行了有益的尝试。相信随着认知神经科学的发展和脑功能成像和电生理技术的不断进步,我们会对两者的神经基础及其结构功能获得更为系统和全面的认识。更多还原

【Abstract】 For about four decades, behavioral studies have revealed that picture-plane inversion dramatically impairs face recognition, which lead to the discussion of what its essence is. Some people think that the inversion effect is unique to faces, and this finding is regarded as evidence that the visual system uses special perceptual processing for faces different from the processing of other non-face objects. Though the selective face-specific areas in the brain have been found by the electrophysiological and neuroimaging studies, it’s still an inconclusive discussion. Other people deem that faces are special, because subjects are required to make visual subordinate-level categorization that concerns expertise in the comparatively homogenous category, and that human beings are expert at discriminating faces at individual levels. Therefore, every category of visual stimuli would show inversion effect, if only people acquire the experise of such visual stimuli. So far, this issue is still the focus of argument.The unique character, Chinese character, has become more and more popular in the visual cognition. Every literate Chinese adult is an expert at discriminating Chinese characters that possess two-dimension features and have advantages over other words as an old type of hieroglyph. It has been found that there is word-specific brain cortex that is named Visual Word Form Area (VWFA) and has effective functions of processing languages. Such findings indicate that there should be anatomic and functional elements of processing Chinese characters. However, more studies focus on the semantic and speech research, the grapheme research is comparatively insufficient, because the effective experimental methods and paradigms are difficult to find, and especially for the research about the early visual processing of Chinese characters themselves.The present research tries to study the behavioral results of inversion effect of faces and Chinese characters so as to compare these two cognitive mechanisms. What’s more, the electrophysiological research is also employed to explore the processing stages and features of the faces and Chinese characters further. In addition, such results can be used for reference to make more exploration and trials.The contextual priming paradigm was used in the present study, and the whole research consisted of two main parts: the behavioral study and ERP study. In the former study, four groups of subjects were induced to respond to the same comparison stimuli in two paradigms (faces vs. English letters, Chinese characters vs. English letters), two cognitive conditions (inversion vs. rotation), and two kinds of changes (configural changes vs. featural changes). The latter study, similar to the former, concerned ERP in which subjects were required to judge the orientation of the stimuli. At last, we compared the reaction time, ERP’s latencies and amplitudes.Some main conclusions were drawn as follows:We found inversion effect not only in recognition of faces, but also in that of Chinese characters, and the subjects responded to the same stimuli in different ways according to the different contextual priming conditions. The two inversion effects were opposite, which definitely demonstrated that inversion effect of faces was independent of experimental paradigm, however, other stimuli could also show this effect, as long as they acquired the common configuration and were very familiar to the human beings who were able to possess expertise in such stimuli by long-term experience. Therefore, the inversion effect couldn’t be regarded as the evidence that there was a unique processing stage for faces. By contrast, our results corroborated thoroughly the explanation of expertise hyhothesis for inversion effect.When we processed the up-side-down faces or Chinese characters, some certain strategies, different from that of processing upright stimuli, were employed, and the shift of the ways of cognitions might happen at an angle of 120 degrees. The regression results showed that there were nonlinear relationships between reaction times and angles. As a rule, the inversion effect only appeared in configural changes and was absent in featural changes, which indicated that inversion mainly impaired the processing of configural information, and such processing made senses in expertise system too.In the model of Bruce-Young, there should be a coarse visual categorization before direct visual processing, configural encoding, and analysis of expression and so on.P1 should be thought to represent an earlier face-specific processing stage as early as 100-120 ms following stimulus onset rather than simply reflect the low-level visual difference. What’s more, it might partly concern configural processing. As for the Chinese character, the early perception stage might originate from P1, and have left advantages over right, which might be followed by the bilateral processing.N170 should not be considered as the face-specific component, neither the reflection of low-level visual difference between faces and non-face stimuli. It should be considered as the reflection of expertise.The inversion of the visual stimuli was likely to delay the laencies of faces and Chinese characters; however, the mental rotation might probably influence the amplitudes.To sum up, the cognition of faces and Chinese characters are both crucial social functions in our life, and our research made a meaningful attempt to compare the two important visual stimuli. It’s believable that human beings will get more systematic understanding about their neurobiological bases and functions with the development of cognitve neuroscience, brain neuroimaging and electrophysiological technology.更多还原