理解双人肢体运动表达的友好和敌对意图的ERP证据

Brain spatio-temporal dynamics of understanding kind versus hostile intentions based on dyadic body movements

已有的意图理解神经成像研究大多关注理解单一个体的中性或负性意图的脑功能定位,而大脑理解依靠双人肢体运动表达的友好和敌对意图的动态时间过程尚不清楚。本研究记录了20名健康被试完成三种不同意图推理任务时的脑电成分。三种意图分别为:1)友好意图; 2)敌对意图; 3)中性无互动意图。行为结果发现理解敌对意图的反应时最短。电生理学结果表明,在额中区的N250 (170～270 ms)上,中性意图比友好和敌对意图均更负,且友好意图比敌对意图也更负;在大脑右半球的P300(270～450ms)上,敌对意图比友好和中性意图均更正,且友好意图比中性意图也更正。对友好和敌对意图的N250和P300进行溯源分析分别定位于额中回(BA10)和脑岛(BA45)。结果表明大脑在多个阶段对双人肢体运动表达的互动意图进行分类理解,对负性敌对意图表现出较早的理解且晚期持续性的评价加工。

Previous social neuroscience studies focused mainly on the neural networks that sustain an understanding of a single individual’s neutral or negative intentions. However, few studies have explored those of different types of social interactive intentions performed by two or a number of individuals and their whole body movements. In addition, the spatio-temporal dynamics of understanding the kind and hostile intentions in the human brain also remain unclear. In order to address the above issue, the electroencephalograph (EEG) technique was employed to explore the dissociative neural correlates of understanding kind versus hostile intention. Twenty healthy participants were recruited for the experiment. Their behavioral data (accuracy and reaction time) and electrical brain activities were recorded while they were watching colorful photos depicting two actors’ actions and performing an intention inference task (IIT). There were three different types of action intentions: kindness, hostility and non-interactiveness (neutrality). The ERP data was analyzed using the Scan and sLoreta software in an off-line way. The Univariate Analysis of Variance (ANOVA) with repeated measures of mean accuracy showed no significant difference among three conditions, while a main effect of condition existed for reaction time. The reaction times of hostile intention were shorter than those of the kind and neutral intentions, while no significant difference was found between the latter two conditions. The ERP data were analyzed using a three-way repeated measure ANOVA. The ANOVA factors were intention condition (kindness, hostility and neutrality), laterality (left, midline and right areas) and caudality (frontal, central and parietal areas). Electrophysiological results showed, over the frontal area, a significantly more negative amplitude of N250 (170~270 ms) for neutral intention compared to kind and hostile intentions, and the N250 amplitudes for kind intention were also more negative than those for hostile intention, especially on the FZ electrode site. The source localization showed maximum activation in Broadman 10 (BA10), in the vicinity of middle frontal gyrus, for N250 for kind and hostile intentions. At the later stage (270~450 ms), the peak amplitudes of the P300 for hostile intention were more positive than those for the kind and neutral intentions, and the P300 amplitudes for kind intention were also more positive than those for neutral intention over the central, parietal areas as well as the right hemisphere. The maximum activation for P300 of kind and hostile intentions was found in BA45, located in the vicinity of insula, and a stronger activity existed for hostile intention compared to kind intention. These findings show that there is a spatio-temporal dynamic dissociation between kind and hostile intentions understanding in the brain. Altogether, the current study provides electrophysiological evidence underlying the kind, hostile interactive intentions and non-interactive (neutral) intention understanding, and suggests a prioritized and sustained processing for hostile interactive intention. The study enriches the contents of the two-stage intention-understanding model and the putative Hierarchical & Multi-level Cognitive Framework (HMCF) in Theory of Mind.