10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018, pp.483 - 487
Abstract
When making repeated decisions, individuals can learn about associations between actions and outcomes through obtained feedbacks. Such a learning process can occur based on individuals' direct experiences in the past, or simply on observed social others' actions and outcomes. Previous computational and neuroimaging studies have shown that one's learning performance is dependent on her sensitivity to reward (or punishment) and reward prediction error, the differences between experienced and expected rewards (or punishments). However, it remains unknown whether individuals' experience-based and observational learning have common or differential cognitive characteristics (e.g., value sensitivity) that affect the learning performances. Here, we use a probabilistic reward learning task, a choice task with different types of uncertainty, and computational modeling approach to quantify individuals' value sensitivity and learning performances. We further examine associations between performances in observational- and experience-based learnings with individuals' psychopathological symptoms. Particularly, depression, a most prevalent symptom in modern society and known factor that affects reward sensitivity, is used as a psychopathological measure of interest. The current study contributes to understanding how individuals' psychopathological symptoms affect their experience-based and observational reward learning.
Publisher
Institute of Electrical and Electronics Engineers Inc.