Abstract: Insight is one of the important ways to study creativity. A recent surge of interest into the neural underpinnings of insight has produced a banquet of research. Prototype heuristic theory is one of the cognitive mechanisms of insight proposed by china scholars recently. A lot of behavioral studies have shown that the theory can be divided into two parts: the prototype activation and heuristic information activation. These studies provide some evidences that the activation of the prototype came as a result of automatic processing and the activation of heuristic information in prototype maybe the result of controlled processing. Since there is no direct electrophysiological evidence to support the theory of two-stage view, therefore, this study explored the physiological mechanism of the prototype heuristics, using the ERP technique with characteristics of high temporal resolution, selected logogriph (a riddle about a Chinese Character) problem as the material. The experiments used the learning- test two-phase paradigm, the two phases of which are the learning phase and the test phase. Specifically, we used the called “five to five” experimental paradigm. 14 participants took part in the experiment. In the learning phase, the subjects learned five prototypical logogriphs in a random order. In the test phase, the subjects solved five target logogriphs paired with the prototypical logogriphs studied previously in a random order. The task for the subjects was to solve the target logogriphs based on the heuristic information obtained in the learning phase. In this study a Single-factor three-level experimental design was used. The three-level was: common component conditions (CCC), uncommon component conditions (UCC) and baseline conditions (BC). The results of the experiment showed that there is no significant difference among three conditions between 300 and 500 ms and CCC and UCC elicited a more positive ERP deflection (P800–1700) than did BC between 800 and 1700ms. We presumed that the insignificant difference between 300 and 500 ms maybe indicated that the activation of the prototypal logogriph from five was same in three conditions, which indicated that the activation of prototype logogriphs might be automatic. The P800-1700 component maybe reflected the application of the heuristic information to the test logogriph and getting the answer successfully. Subjects cannot solve the test logogriph directly from the heuristic information obtain from learning phase until access to the deep links between them. But in the baseline conditions, subjects can solve the logogriph in test phase based on heuristic information. These results might reflect the higher creative level in CCC and UCC conditions than that in BC condition. Additionally, the waveform elicited by CCC and UCC conditions was tend to consistent, which indicated that the same text may not be the key of prototype heuristics.

Key words: Chinese riddles, insight, prototype heuristics, Event-related potential

摘要： 以中国传统字谜为实验材料，采用多对多的“学习-测试”二阶段实验范式和ERP记录手段，研究了原型字谜和靶字谜之间是否含有共同字词对顿悟的原型启发的影响。结果发现，在靶字谜呈现后的700ms内，正确解决有共同字词、无共同字词和基线条件字谜所诱发的脑电活动基本一致。在800-1700ms时间窗口内，有共同字词条件和无共同字词条件比基线条件诱发了一个更正的P800-1700成分，可能反映了原型启发下原型的激活和新颖联系的形成。此外，有共同字词条件和无共同字词条件诱发的波形始终趋于一致，可能表明字面上相同的字词并没有对原型启发产生影响，而原型字谜和靶字谜之间内在的语义联系才是原型启发的关键。

关键词: 字谜□顿悟□原型启发□事件相关电位