动作理解因境而异:动作加工中情境信息的自动整合

Contextual modulation of action interpretation: Automatic integration of situational contexts during action understanding

针对动作理解的机制,模拟论主张大脑自发模拟他人的动作,就相同的动作其理解也相同,而理论论则认为人们基于合理性原则对他人动作进行推理,相同的动作发生在不同的情境时会有不同的理解。但以往研究所采用动作材料的运动学特性和发生情境存在共变,其难以区分动作理解是支持模拟论还是理论论。通过两项实验,采用动画制作技术来产生有无约束情境下的追逐动作,以指示动作加工过程的脑电μ抑制为指标,对前述两种观点进行了检验。其中,在约束情境中存在障碍物,追逐者需改变运动方向以绕过障碍物,从后方逐渐趋近目标;而无约束情境中不存在障碍物,但追逐者依然保持与存在约束情景下相同的运动模式。结果发现,当追逐动作发生在存在约束的情境时,其基于合理性原则推测可获得清晰的动作目标,该条件下的μ抑制程度高于不存在约束情境的条件(实验1);而当仅追逐者运动,即趋近的目标不确定时,虽然有约束和无约束情境间的物理差异与实验1相同,但条件间μ抑制的差异消失(实验2);且上述μ抑制并非与注意相关的枕叶α活动的泛化。该结果提示,动作发生的情境信息影响人们对动作的理解,即基于推理过程理解动作,支持理论论观点。

Action understanding enables us to predict others’ actions and interact with them smoothly;however, its impairment makes patients unable to take care of themselves leading to cases such as autism. Despite the central importance of action understanding, the cognitive mechanisms involved in it remain highly controversial. Two alternative accounts have been advanced. Simulation theory assumes that we understand actions by simulating the observed behavior through a direct matching process, which has been suggested to usually activate the mirror-neuron circuit. The alternative interpretive account(i.e., theory) assumes that action understanding is based on specialized inferential processes, in which a goal is assigned to an action by evaluating its efficiency as an optimal means of obtaining the goal within the specific constraints of the situation. Each account was supported by previous studies, but due to mythological drawbacks, the evidence can be explained by both accounts. Hence, it remains unclear how action understanding is implemented in our cognition. The simulation theory claims that action understanding is sensitive to minor differences in the kinematics of actions, and the theory emphasizes the role of contextual information in action understanding. Namely, according to the theory, even for identical actions, contextual information modulates action understanding. Hence, to examine which account is involved in action understanding, we created a chasing action wherein a chaser pursues a movable target, but the chasing action occurred in different contexts. Specifically, in a constrained context, the chaser had to bypass the obstacles to approach the target, and the chaser’s action was efficient in catching the target, while in an unconstrained context, the obstacles were removed but the chaser still implemented the same action as in the constrained context. All actions lasted for 3 seconds. In both contexts, the chaser and the target had exactly the same kinematics, but were assigned different goals if the inferential process is involved. To identify the outcome of action understanding, we measured μ suppression(electroencephalogram oscillations within the 8–13 Hz range in the sensorimotor regions;namely, C3 and C4 channels) related to action understanding. Participants were asked to count the fillers(i.e., incomplete chasing action) when watching actions presented on the screen.It was found that the chasing action occurred in the constrained context(M = –1.955 μV^2) induced more μ suppression than the action occurred in the unconstrained context(M = –1.913 μV^2), but in both contexts, the evaluated familiarity for them was not significantly different(Experiment 1). Importantly, the occipital α with the same frequency band as μ was not modulated by the contextual information, but this component was suggested to be functional with the attentional mechanisms. In Experiment 2, to further test whether the effect in Experiment 1 was specific to the inferential process, the target was set to be still but the chaser still moved in the same way, which cannot be attributed to an analytical goal. In this case, the possible simulation difference between constrained and unconstrained contexts when the target was still was almost the same as when the target was moving;however, we found that the difference in μ suppression between constrained and unconstrained contexts was insignificant. Our findings showed that contextual information modulates μ suppression, suggesting that the action understanding is sensitive to the context, and the assigned goal for the actions depends on the contextual information. Hence, our findings support the assumption that action understanding is primarily mediated by an inferential interpretive system rather than a simulation process.