摘要
目的:本研究首次在中国健康大学生样本中应用概率奖赏任务(Probabilistic Reward Task,PRT),并对其奖赏强化学习行为的机制进行探讨。方法:127名健康大学生被试完成计算机汉化版的概率奖赏任务(长短嘴版本,300试次),采用重复测量方差分析对反应偏向、鉴别力、准确率、反应时等指标进行统计。结果:反应偏向与鉴别力的组块主效应显著(F(2,252)=20.05,P<0.001;F(2,252)=13.02,P<0.001),随操作次数增多,反应偏向与鉴别力逐渐提高。反应时和准确率的刺激类型主效应显著(F(1,126)=84.09,P<0.001;F(2,126)=143.08,P<0.001),奖多刺激(获奖赏机率高的刺激)的反应时显著低于奖少刺激(获奖赏机率低的刺激),奖多刺激的准确率显著高于奖少刺激(P<0.001)。反应时和准确率的组块和刺激类型的交互作用均显著(F(2,252)=14.15,P<0.001;F(2,252)=17.79,P<0.001)。在奖多刺激条件下,反应时组块1>组块2=组块3(P<0.05),准确率组块3>组块2>组块1(均P<0.001);而奖少刺激条件下,三个组块之间的反应时及准确率均没有显著差异。结论:中国大学生群体在PRT任务操作过程中,表现出明显的追求更高机率奖赏的动机与欲望,并存在渐进累积的奖赏学习效应。
Objective： Based on the first application of Probabilistic Reward Task（PRT） in Chinese undergraduate stu- dents, this study introduced the PRT paradigm per se and the reward reinforcement learning mechanism it illustrated. Meth- ods： Totally 127 healthy undergraduate students completed computer-based PRT including 300 trials. Results： Signifi- cant main effects of block on response bias and discriminability were observed. A significant main effect of stimulation type was also detected on the reaction time. The RTs for the rich stimulus（higher possibility for rewarding） was significantly short- er than that for lean stimulus（lower possibility for rewarding）, The discriminability for the rich stimulus was significantly higher than that for lean stimulus. A significant interaction between block and stimulation type was found on RTs, showing that significant differences among three block on RTs were presented for the rich stimulus type, but not for the lean stimulus type. Additionally, there was significant interaction between block and stimulation type on accuracy, showing that significant differences among three block on accuracy were peresented for the rich stimulus type, but not for the lean stimulus type. Conclusion： Through the performance of PRT, Chinese undergraduate students show obvious motivation and desire to pursuit high probability of rewarding. Additionally, they show enhanced reward learning effect along with the gradual accu- mulation of experiences.
出处
《中国临床心理学杂志》
CSSCI
CSCD
北大核心
2017年第1期1-5,24,共6页
Chinese Journal of Clinical Psychology
基金
国家自然科学基金(31671144
81401120)
教育部人文社会科学研究规划基金(13YJA190015)
新世纪优秀人才支持计划(NCET-12-0557)
关键词
概率奖赏任务
奖赏学习
积极强化
Probabilistic reward task
Reward learning
Positive reinforcement
