【作者】 郑成艳； 严璘璘； 王哲；

【Author】 Zheng Chengyan;Yan Linlin;Wang Zhe;Department of Psychology,School of Science,Zhejiang Sci-tech University;

【机构】 浙江理工大学理学院心理学系；

【摘要】 面孔整体加工是指知觉系统有选择性地快速、准确地整合面孔各部分信息,从而形成一种整体表征。类别-个体化理论认为,对于外群体成员,人们往往关注类别信息,倾向于将成员均归为一类,而对于内群体成员,更关注个体信息,倾向于将每个成员形成单独的整体表征。早期支持该理论的研究表明,高加索人对本族面孔的组合面孔效应大于他族面孔。但近年来,亚洲人或高加索人在知觉本族和他族面孔时行为上均表现出相同程度的组合面孔效应(整体加工的指标)。这意味着人们在面孔知觉早期是否存在类别信息和个体信息的选择性注意及其对后续加工的影响等问题并不清楚。为了考察面孔知觉早期是否存在信息选择及其对后续加工模式的影响,本研究采用事件相关电位(event-related potential,ERP)来探测个体在面孔知觉最初几百毫秒的电生理信号变化(如P100、N170、P200和N200)及其与后续加工模式之间的关系。实验选取19名(9男)中国被试,测量其对三类面孔(亚洲面孔、高加索面孔以及亚洲-高加索两歧面孔)的组合面孔效应及其脑电成分幅度和潜伏期。行为结果表明,中国被试对于亚洲面孔和高加索面孔表现出相同的组合面孔效应(ps <0.01),但对两歧面孔未表现出类似的组合面孔效应(p=0.52)。ERP结果表明,他族面孔诱发P100和N170幅度大于本族和两歧面孔(ps <0.001),而本族和两歧面孔之间没有明显差异(p> 0.05)。他族面孔诱发P200成分的幅度大于两歧面孔(p <0.001),且两歧面孔诱发P200的幅度大于本族面孔(p=0.05)。两歧面孔诱发N200的幅度小于本族面孔(p=0.02),与他族面孔没有明显差异(p=0.43)。ERP与行为数据的相关分析进一步显示,对亚洲面孔而言,P200的幅度越小,组合面孔效应越大(r=-0.45,p=0.05);对高加索面孔和两歧面孔而言,P200的幅度和组合面孔效应没有相关(高加索面孔r=-0.13,p=0.59;两歧面孔r=-0.09,p=0.70)。上述结果提示,在面孔知觉早期,他族信息诱发更强的注意,而本族信息能促使个体形成整体加工,支持了类别-个体化理论。更多还原

【Abstract】 Holistic Face processing refers to the selectively rapid and accurate integration of face information by the perceptual system, so as to form a whole representation. Categorization-Individuation Model proposes that attention to category-diagnostic facial characteristics of out-group members can cause targets to be perceived and remember as more homogeneous. While selective attention to the identity-diagnostic facial characteristics of in-group members tend to form a whole representation of each member. Early work supported this hypothesis by showing stronger composite effect size in own-race faces than in other-race faces. While more recent studies showed that same composite-face effect(an index for holistic face processing) sizes were observed with composite face paradigm for own-race faces and other-race faces by Asians and Caucasians. It was not clear whether selective attention to the category-or identity-diagnostic facial characteristics be at the early stage of face perception and how to influence the subsequent face processing. In order to investigate whether characteristic selection be at the early stage and how to influence the subsequent face processing, an event-related potential(ERP) was used to detect the electrophysiological components(e.g., P100, N170, P200, and N200) during the first few hundred milliseconds of face perception and the relationships with subsequent face processing. Nineteen(9 males) Chinese participants were recruited. We measured the composite face effect size of three face types(Asian faces, Caucasian faces, and Asian-Caucasian ambiguous faces) and the amplitude of the ERP component. The behavioral results showed that same composite face effect sizes were observed for Asian faces and Caucasian faces(ps < 0.01), but no similar effect for the ambiguous faces(p =0.52). The ERP results showed that P100 and N170 amplitudes were significantly larger for other-race faces than own-race and ambiguous faces(ps < 0.001), but no difference between own-race and ambiguous faces(p > 0.05).P200 amplitude was significantly larger to other-race faces than ambiguous faces(p < 0.001), and it was marginally larger for ambiguous faces than own-race faces(p = 0.05). N200 amplitude was markedly smaller for ambiguous faces than own-race faces(p = 0.02), but there was no difference between other-race and ambiguous faces(p = 0.43).The correlation analysis between neural responses and behavioral data showed that the smaller P200 amplitude, the larger composite face effect for Asian faces(r =-0.45, p = 0.05). However, there was no correlation between the amplitude of P200 and the composite face effect for Caucasian(r =-0.13,p = 0.59) and ambiguous faces(r =-0.09,p = 0.70). The results suggest stronger attention to category-diagnostic characteristics were elicited by other-race faces at the early stage of perception. And attention to identity-diagnostic characteristics for own-race faces played a role in holistic face processing which support the Categorization-Individuation Model.更多还原