The Effects And Mechanisms Of Reward On Episodic Memory Generalization And Mnemonic Discrimination

It is well established that items or information with high rewards will be remembered and recognized better than non-rewarded ones over time(Miendlarzewska et al.,2016;Shohamy & Adcock,2010).Previous research adopted the traditional old/new recognition paradigm to test episodic memory,in which subjects could make correct recognition with fewer memory traces(for example,the same dog presented in the test stage,subjects can easily recognize it as “old”).However,when organisms are confronted by ambiguous cues resembling previous signals of reward(a similar but not the same dog presented in the testing stage),could the reward experience promote the accuracy of the individual’s recognition of this type of stimulus? Do they make a memory generalization – recognize it as old,or make a memory discrimination – successfully recognize the difference between target and lure and recognize it as similar?The Mnemonic Similarity Task(MST)was designed to evaluate the individual’s memory discrimination ability by adding lures in the test phase(Stark et al.,2019).First,on the one hand,studies have confirmed that the similarity of lure and target will significantly affect whether individuals make generalization or discrimination(Hanert et al.,2017).On the other hand,behavioral research shows that reward promotes memory by improving the consolidation process after encoding,which shows that reward-related memory enhancement only occurs under the 24-hour delayed test condition.If reward-conditioned learning also improves memory generalization and discrimination by promoting memory consolidation.In that case,the reward-related enhancement of generalization also occurs under 24-hour delay conditions.Therefore,it is necessary to examine the critical variables that affect reward-conditioned learning to promote memory generalization and discrimination to verify the stability of this effect.Secondly,previous studies have confirmed that rewards refer to four sub-components:expected value,outcome,uncertainty,and reward prediction error(RPE)(Mason et al.,2017).Research on stimulus generalization shows that reward prediction error can reduce the generalization gradient(Kahnt et al.,2012),while uncertainty as a stress factor can improve generalization(Zweifel,2019).Therefore,whether different reward sub-components have different effects on memory generalization and memory discrimination needs further investigationTherefore,we first investigate whether reward conditioning promotes episodic memory generalization and discrimination and explore the boundary effects of time interval and stimulus similarity.Experiment 1 examined the impact of the time interval on memory generalization.Experiment 2 examines the effect of the similarity of stimuli between encoding and test on memory generalization.Secondly,we further investigated which reward component is the key to memory generalization and discrimination.Experiment 3 first controlled the expected reward to investigate reward uncertainty’s impact on episodic memory generalization.Experiment 4controlled the reward uncertainty to investigate the expected reward’s influence on episodic memory generalization.Finally,based on study 2 and study 3,we further explored the cognitive mechanism of reward sub-components on episodic memory generalization and discrimination.Experiments 5 and 6,respectively,compared the number of fixation counts on the picture display screen and the pupil dilation of the value feedback screen in the encoding stage.Based on the above research,the following main conclusions are drawn.(1)Reward-conditioned learning promotes episodic memory,which does not depend on consolidation.(2)Reward-conditioned learning promotes generalization,and the time interval does not affect the enhancement,which confirms that this effect is consolidation-independent.(3)Stimuli similarity is the boundary condition for reward enhancement of episodic memory generalization.(4)Reward uncertainty and expected reward promote episodic memory generalization by increasing the number of fixations on the picture,helping individuals judge more lure stimuli as seen.(5)Unsigned reward prediction error enhances mnemonic discrimination by enhancing arousal after encoding and improving the processing of item details.