Abstract: The memory of people are not always correct, when people sometimes declare to remember some events that in fact has never occurred, this is how false memories produced. Spreading activation theory tries to explain the mechanism of false memories. According to this theory, semantics related information is organized together in memory. Semantic activation of the learning items will spread to critical lures that have semantic association with the items through semantic network and thus activates the critical lures to generate false memory. Due to the fact that false memory has important influence on daily life, many researches focus on how to reduce false memory. Presenting a warning signal is the most common method in the laboratory to reduce false memory. DRM paradigm only shows one type of human false memory phenomenon. Studies show that mood also has great influence on human memory and this effect is mainly reflected in mood-congruent memory bias. Bower (1981) used associative network theory of emotion to explain this phenomenon. Based on spreading activation theory and associative network theory of emotion, we hypothesized that unconsciously spreading activated semantic and emotional information leads people to retrieve mood-congruent information. This information includes original input information and moods and semantics related non-original information, namely false memory information. False retrieval of mood-congruent non-original information generates mood-congruent false memory. Therefore, Experiment 1 examined whether the existence of mood-congruent false memory among normal participants. After warning cues were added, experiment 2 explored whether warning cue can reduce false memory. This study will explore the hypothesis: (1) error recognition rate of positive critical lures in the positive mood is significantly higher than that in the negative and neutral moods, and error recognition rate of negative critical lures in the negative mood is significantly higher than that in the positive and neutral moods; (2)warning cues cannot reduce mood-congruent false memory. Thirty-two and thirty participants were randomly selected in experiment 1 and 2 by campus BBS, respectively. All participants were right-handed with normal or corrected to normal visions, and they had no history of mood disorders. They were equally selected from science major and liberal arts major. All data were analyzed by SPSS 16.0. The hypothesis is supported by the result. The results showed that: (1) error recognition rate of positive critical lures in the positive mood was significantly higher than that in the negative and neutral moods and error recognition rate of negative critical lures in the negative mood was significantly higher than that in the positive and neutral moods; (2) warning cue cannot reduce false memory and mood-congruent false memory; (3) participants were inclined to make “remember” judgment to mood-congruent critical lures with or without warning. This study creatively combined spreading activation theory with associative network theory of emotion to explain mood-congruent false memory phenomenon. Theoretical significance of this study lies in the fact that it can further reveal the mechanism of the interaction between the left hemisphere of the brain which is responsible for word processing and the right hemisphere of the brain which is responsible for perceptual material processing. Practical value is mainly manifested in the fact that it can be used to explain how a pessimist becomes increasingly pessimistic and how an optimist becomes increasingly optimistic. It also has important application value for emotional management and emotional adjustment in psychological counseling.

Key words: DRM paradigm, warning cues, mood-congruent false memory