基于他人受益信息的社会交互意图识别

The recognition of intentions of observed actions or behaviors is one of important social functions in human’s visual system. Previous research has invested much effort on how vision recognizes object-directed intentions of actions (i.e., actions are implemented to approach the physical objects without influencing the others). However, actions are also directed to social entities or agents, as to impact others, which is termed as social interaction intention. The current project is aimed to investigate how to recognize social interaction intentions in vision. On the basis of elaborating the properties that multiple individuals involved in social interaction are rational and should maximize the utility of actions in the overall level, this project proposes a hypothesis that when one’s actions bring about benefits for others (short for, information of benefiting for others), they are recognized with social interaction intentions. To test this hypothesis, we plan to use both behavioral and EEG/ERP methods. In specific, by manipulating how two agents move and influence each other, and so on, this project is about to first examine whether the proposed information determines the recognition of social interaction intention, then attempts to investigate the boundary conditions and the mechanisms about how the information of benefiting for others is applied to recognize social interaction intentions. Further, based on experimental findings, we expect to construct a preliminary theory about the visual mechanisms of recognizing social interaction intentions. This project is hopefully to provide a psychological basis for designing robots with the ability to implement social interaction, the treatments for impaired social interaction, and so on.

人类视觉系统的重要社会性功能之一是识别他人动作或行为的意图。与已有研究着重考察视觉系统如何识别客体导向性意图（动作以指向物理对象为目标，而不涉及其他人）不同，本项目拟探讨视觉系统如何对社会性交互意图（动作以指向社会主体为目标且影响对方行为）进行识别。基于对社会交互中的行为在理性上应遵循总体效用最大化的分析，申请者提出当一方因另一方的动作或行为获得收益（简称为他人受益信息）时，视觉系统将其识别为动作发出方对受益方存在以交互为目的的行为意图。为检验该假设，本项目拟采用行为实验与脑电技术相结合的方法，操纵两个行为主体间的动作模式与影响方式等，系统探讨视觉是否可基于他人受益信息对社会交互意图进行识别，并试图揭示该类信息影响交互意图识别的边界条件以及机制。在此基础上，试图构建相关理论，阐明视觉对社会交互意图识别的认知机制，从而为设计具有社会交互功能的机器人以及社交障碍相关疾病的治疗等提供心理学依据。

人类视觉系统的重要社会性功能之一是识别他人动作或行为的意图。与已有研究着重考察视觉系统如何识别客体导向性意图（动作以指向物理对象为目标，而不涉及其他人）不同，本项目探讨了视觉系统如何对社会性交互意图（动作以指向社会主体为目标且影响对方行为）进行识别。基于对社会交互中的行为在理性上应遵循总体效用最大化的分析，本项目提出当一方因另一方的动作或行为获得收益（简称为他人受益信息）时，视觉系统将其识别为动作发出方对受益方存在以交互为目的的行为意图。利用行为实验与脑电技术相结合的方法，采用记忆混淆效应和脑电指标Mu，通过操纵两个行为主体间的动作模式与影响方式等，对上述假设进行了检验。本项目结果为基于他人受益信息（一方因另一方行为受益）进行交互意图识别提供了实验证据。同时，发现他人受益信息作用于社会交互意图识别的边界条件：当受益物品不属于行动者时，必须建立在他人充分获益的前提下才被识别为具有社会交互意图，且他人受益必须由行动者发出而非外界干预所致。最后，本项目揭示社会个体的动作既满足他人受益条件也符合客体导向性意图的识别条件时，其被识别为具有社会交互意图，即他人受益信息作用于意图识别具有一定的优先性。该项目加深了对两个社会个体间动作意图识别的认识，弥补了以往对单个个体动作意图识别研究的不足。且本项目首次将行为理性原则拓展至社会交互领域，提出动作理解中的合理性原则也适用于多人情景，即人们在多方交互主体构成的整体层面进行效用最大化分析，从而识别动作的意图类型。基于上述研究结果，发表期刊论文6篇，其中SSCI/SCI论文5篇。本项目不仅实现了预期目标，同时还丰富和深化了原计划的研究内容，为后续探讨群体行为加工机制奠定了基础。