Abstract: The current study investigated the cognitive and neural electrophysiological processes of true as well as the false memories in story contexts. The results from studies of human memory suggested that recognition memory performance reflected two distinct memory processes: familiarity and recollection. Familiarity was to assess the overall similarity between test items and all study items during 300-500ms.FN400 ERP was related to familiarity. Recollection involved the retrieval of special details of the study items during 400-800ms.LPC was related to recollection. Moreover, an additional old/new ERP (800-1100ms) tended to be most prominent over frontal scalp sites, which was considered to be related to post-retrieval processing, such as evaluation of the products of retrieval for source and specific item features. This study chose four story contexts as the material and adopted the learning and testing experimental paradigm. In the studying phase, participants were asked to remember story context which appeared on the computer screen. Then, the participants were required to do an old/new recognition test. There were three types of target words during the testing stage including old words, critical words and new words. Critical words were new and highly associated with the story contexts, in addition new words were unassociated with the contexts. Event-related potential（ERPs）and RT were measured when 22 healthy right-handed undergraduates did an old/new recognition test. None of them had history of mental illness or brain disease. After the experiment, we gave certain degree of compensation to subjects. Two subjects were rejected because of the lack of the number of artifact-free trials. The instrument for the statistical analysis used in the experiment is SPSS15.0. The results showed that（1）The critical items produced high rate of false memory. Moreover, false recognition demanded more RT than true recognition.（2）The FN400 which reflects familiarity-based process revealed no significant differences between true and false recognition, but there were more negative than Correct-Rejection recognition.（3）During the recollection-based process（400-800ms）,the amplitude of LPC elicited by false recognition and true recognition was not different, while there were more positive than Correct-Rejection recognition. （4）During the post-retrieval process（800-1110ms）,the difference among the ERPs amplitude elicited by recognition about three target words got remarkably significant difference. The amplitude of LSW elicited by false recognition was greater than true recognition and Correct-Rejection recognition. In conclusion, the results demonstrated that true and false memories had showed the same information processing models in story contextual cuing effect during recollection-based process and familiarity-based process. However the information processing models changed during the post-retrieval process: false memory needed much more information process than that of true memory for retrieval and comparing, so it took longer time to recognize. This study provided a new evidence for the Source – Monitoring Framework.

Key words: False Memory, Story Contexts, FN400, LPC, LSW

摘要： 以故事情境为实验材料，采用学习再认实验范式，通过操纵再认靶词与故事情境的关联性考察了情境故事对关联虚假记忆和真实再认神经活动的影响。结果发现：在记忆再认的熟悉性加工和回想性加工阶段，虚假记忆与真实记忆所诱发FN400和LPC成分走向一致；在额叶后提取阶段，虚假再认所诱发的晚慢成分波幅大于真实再认。结果说明：情境性关联虚假记忆与真实记忆表现出相似的熟悉性和回想性加工，但在后期需耗费额外的加工资源进行成功的提取。

关键词: 虚假记忆, 故事情境, FN400, LPC, LSW