Journal of Psychological Science ›› 2023, Vol. 46 ›› Issue (4): 817-824.DOI: 10.16719/j.cnki.1671-6981.20230407

• General Psychology, Experimental Psychology & Ergonomics • Previous Articles     Next Articles

Effect of Unsigned Reward Prediction Error on Mnemonic Discrimination: Evidence from Behavioral and Eye-Tracking

Yue Yang, Jiang Yingjie, Long Yiting, Wang Kaiyu   

  1. School of Psychology, Northeast Normal University, Changchun, 130024
  • Online:2023-07-20 Published:2023-08-14

奖赏预测误差对记忆辨别的影响:行为及眼动的证据*

岳阳, 姜英杰**, 龙翼婷, 王凯玉   

  1. 东北师范大学心理学院,长春,130024
  • 通讯作者: ** 姜英杰,E-mail: jiangyj993@nenu.edu.cn
  • 基金资助:
    * 本研究得到吉林省自然科学基金面上项目(20230101149JC)和国家自然科学基金面上项目(32271095)的资助

Abstract: Mnemonic discrimination was defined as a behavioral index of pattern separation to reflect the ability to discriminate memory representations of similar features. The purpose of this study was to examine the effects of reward conditioning on mnemonic discrimination. In this research, participants underwent Pavlovian reward conditioning, in which objects from one category were paired with a reward, and objects from a different category were unpaired. Participants were given a subsequent surprise memory test 24 hours later, in which they saw another series of stimuli, some of which were seen once before in the conditioning phase (targets), some were the same stimuli but displayed in different positions and angles (lures), and some were new (foils). They were given instructions that they should identify each item as “old”, “similar”, or “new”. We calculated the lure discrimination index, which operationalized as people correctly identify similar lures as similar relative to misidentifying new foils as similar.
The pre-experiment showed that, from the category level, items highly similar to those from the object category previously paired with reward were recognized as similar for lure items more often than items from the reward-unpaired category (p < .05). The result supported reward conditioning learning enhances discrimination. To further test the effects of signed and unsigned prediction errors on discrimination, we changed the reward feedback rate from 50% in the high reward category to 80%. The result supported that unsigned reward prediction error, rather than signed reward prediction error, promotes mnemonic discrimination (p < .01).
In Experiment 2, we collected pupil dilation, which can reflect the brain arousal state caused by reward prediction error, to explore further the impact of unsigned reward prediction error on discrimination. The behavioral result repeated the result in Exp 1, which showed that unsigned reward prediction error enhances discrimination (p < .05). Eye movement results showed that the greater the unsigned reward prediction error, the greater the pupil dilated (p < .001). Besides, for the lure hit items, the pupil dilation of high URPE was significantly higher than low URPE (p < .001). However, the pupil dilation difference between high and low URPE is insignificant for the lure false alarm items. Pupil dilation is the positive predictor of mnemonic discrimination (β = 4.95, SE = 2.14, z = 2.31, p < .05). The results suggested that unsigned reward prediction error positively influences mnemonic discrimination, probably by enhancing the salience or arousal during post-encoding consolidation.
To sum up, we conclude that reward conditioning can promote episodic memory discrimination. Further, the unsigned reward prediction error could promote memory discrimination by enhancing the brain arousal state, helping the memory of the verbatim traces of items, and supporting the lure correct rejection. Future work could consider using a broader continuous range of unsigned prediction errors and collecting information that can more directly reflect the activity of locus coeruleus, such as α- salivary amylase, to test the effect of the unsigned reward prediction error on mnemonic discrimination and investigate its cognitive and physiological mechanisms.

Key words: mnemonic discrimination, reward prediction error, reward conditioning, episodic memory, eye-tracking

摘要: 记忆辨别反映了个体对所经历事件能够形成独特的记忆表征并正确拒斥相似诱饵的能力,是情景记忆最重要的特征之一。研究采用奖赏条件学习范式和记忆相似任务考察奖赏预测误差对记忆辨别的影响。结果显示奖赏预测误差的绝对值越大,记忆辨别值越高;在诱饵击中项目中,高奖赏预测误差绝对值条件下的瞳孔扩张值(pupil dilation, PD)显著高于低绝对值条件;而在诱饵虚报项目中,两者无显著差异;瞳孔扩张值是记忆辨别的正向预测因子。综上,不带符号的奖赏预测误差能够引发个体的意外反应、提高其情绪唤醒水平(瞳孔扩张)、促进个体对细节信息的加工、提高记忆辨别。

关键词: 记忆辨别, 奖赏预测误差, 奖赏条件化, 情景记忆, 眼动追踪