塞翁失馬,焉知非福:公平性覺察與負面情緒提升決策表現的模型估計與反饋負波證據

Misfortune might be a blessing in disguise: Novel evidence from model fitting and feedback-related negativity indicates the impact of fairness perception and negative emotional experience in the enhancement of decision-making performance

黃尹笛(Yin-Dir Hwang)

指導教授 ： 徐永豐

共同指導教授 ： 賴文崧(Wen-Sung Lai)

國立臺灣大學/理學院/心理學研究所/碩士(2016年)

https://doi.org/10.6342/NTU201610060

摘要

人生中，我們常因一些外在因素影響自身的心情，例如比較的心態、或是面對生活中各種不公不義，進而影響決策的判斷。本篇研究利用兩階段的實驗設計，藉由數理模型、事件關聯電位揭露情緒對於行為決策影響。受試者首先面臨「命運遊戲」，63位受試者隨機分派到「仁慈組」、「中立組」或是「苛刻組」。每組在過程中受試者會被迫接受一系列的資產分配，覺察其中分配的公平性，進而產生對應的情緒。從「命運遊戲」中所獲得的資產，將為第二階段「機會遊戲」的起始籌碼。「機會遊戲」是一連串的二選一選牌遊戲，同回合呈現的兩張牌會有不同的機率被扣資產。受試者的遊戲目的為最大化自己的總資產，實驗長度約莫兩小時。分析方法上，我們以反饋負波來量測與「預測錯誤」有關之情緒反應，並採用貝氏方法之增強學習模型分析行為資料。結果發現「苛刻組」於「機會遊戲」保留的總資產最高，他們的「非預期」反饋負波值較其他兩組為大。另外，在模型參數表現上他們有較高的趨避損失傾向、較低的即時更新速率、以及較高的選擇一致性。

關鍵字

情緒調節 ； 事件關聯電位 ； 反饋負波 ； 內側額葉負波 ； 預測錯誤 ； 增強學習模型

並列摘要

Equality and fairness in social interaction often elicit affective arousal and show a great impact on decision making. The present study aims to un- cover the mechanism behind such daily-life experiences using the behavioral, model-fitting, and electrophysiological approaches. In the first session of the experiment, Sixty-three paid participants were randomly assigned to one of the “Neutral,” “Harsh,” and “Kind” groups to undertake a different level of perceived fairness. Then a probabilistic gambling task with different proba- bilities of negative-reward assignments was applied to each participant to ex- amine the impact of emotional experience on her choice behavior. Altogether the procedures lasted about two hours. Trial-by-trial data were fitted by a hybrid reinforcement learning model using the Bayesian estimation approach. Brain activities were measured via event-related potentials. Our analyses re- vealed that, compared with other two groups, participants in the Harsh group retained more task scores, exhibited distinct parameter values of the reinforce- ment learning model, and demonstrated a larger feedback-related negativity to unexpected outcomes, suggesting a higher sensitivity to prediction error and a tendency to loss aversion.

並列關鍵字

affective modulation ； event-related potential ； fairness ； feedback-related negativity ； medial frontal negativity ； prediction error ； reinforcement learning model

參考文獻

Ahn, W.-Y., Busemeyer, J. R., Wagenmakers, E.-J., & Stout, J. C. (2008). Comparison of decision learning models using the generalization criterion method. Cognitive Science, 32, 1376–1402.

Ahn, W.-Y., Vasilev, G., Lee, S. H., Busemeyer, J. R., Kruschke, J. K., Bechara, A., & Vassileva, J. (2014). Decision-making in stimulant and opiate addicts in protracted abstinence: Evidence from computational modeling with pure users. Frontiers in Psychology, 5, 849. doi: 10.3389/fpsyg.2014.00849

Amiez, C., Joseph, J.-P., & Procyk, E. (2005). Anterior cingulate error-related activity is modulated by predicted reward. European Journal of Neuroscience, 21(12), 3447– 3452.

Barth, C. M., & Funke, J. (2010). Negative affective environments improve complex solving performance. Cognition and Emotion, 24(7), 1259–1268.

Bayer, H. M., & Glimcher, P. W. (2005). Midbrain dopamine neurons encode a quantita- tive reward prediction error signal. Neuron, 47(1), 129–141.