摘要:
在测验中将错误题目的正确答案反馈给个体后，在随后的测验中，信心水平高的错误题目比信心水平低的错误题目更容易被改正，这一现象被称作“矫枉过正”。目前关于“矫枉过正”现象有三种理论解释。注意力理论认为相对于信心水平低时出错，人们发现自身信心水平高的题目竟然会犯错误，因此会唤起注意力，并且花费较多精力去学习正确答案；熟悉性理论指出，那些信心高时所犯错误的一般知识可能是人们一开始就知道的，学习一种已经存在于记忆中的信息远远比学习一个新颖的知识简单；最近学习区理论从学习的角度进行阐述“矫枉过正”可能的心理机制，最近学习区中的知识材料学起来不太困难，但是在没有进行再次学习(如反馈)的情况下也不会被直接提取出来，因此反馈本身很关键。“矫枉过正”的神经机制研究发现在错误矫正时，除了代表元记忆不匹配的共同脑区发生活动，背外侧前额叶只有在高信心回答错误的情况下发生活动，在低信心回答正确时没有出现活动，右背外侧前额叶也是如此，因此可以认为背外侧前额叶参与了错误矫正过程，是错误记忆的抑制机制。未来可以考虑在一些特殊人群等领域展开大量研究，如记忆障碍者、学习障碍者是否存在“矫枉过正”现象，这对于了解人类高级认知过程极具价值。
Abstract: Hypercorrection refers to the phenomenon that after the correct answers of questions are fed back to the individual in the test, in the following tests, comparing to the wrong questions of low confidence level, those of high confidence level are more likely to be corrected. There have been three kinds of theories to explain the phenomenon. Attention theory holds the view that people feel surprised when they have failed in questions which they are sure to get the right answers, so, it will arouse their attention and people will spend more time learning the correct answers. Familiarity theory points out that people have known the knowledge of the questions of high confidence level, and learning the information existing in the memory is much easier than learning new knowledge. Recent study area theory tries to explain this phenomenon from a teaching perspective that learning knowledge in recent study area is not difficult, but it will not be extracted directly without relearning (e.g. feedback), so the feedback itself is the key. Hypercorrection’s neural mechanism study found that the brain regions of metamemory which does not match are involved in error correction process; especially dorsolateral prefrontal is activated only in the high- confidence but wrong answers, which is the same as right dorsolateral prefrontal cortex. So dorsolateral prefrontal is the inhibition mechanism of false memories. In future research we should pay more attention to the areas of some special groups, such as memory disorder and learning disabilities; this is extremely valuable to know human advanced cognitive process.