认知与发展心理学研究室知识库

想象这样一个场景，一只狮子正在扑向一只羚羊。两者之间可能的碰撞，对羚羊而言，意味着性命枚关，而对狮子而言，意味着食物的猎取。对双方来说，对碰撞时间(time-to-collision, TTC)的估计都有非常重要的生存意义。己有不同理论认为，TTC的估计可以通过刺激与个体之间的距离和刺激的靠近运动速度来预测;或者依赖于刺激以及当前环境所提供的光学特征;还可以取决于靠近运动的刺激在任意两点间的距离上视角大小和视角变化率的比值。这些观点都认为TTC的估计主要依赖于较低水平物理属性的加工。

然而，有一些研究发现情绪对TTC估计会产生影响。例如，相对于中性图片，人们都会更低估威胁场景图片(Brendel, DeLucia, Hecht, Stacy, & Larsen, 2012)和威胁动物图片(Vagnoni, Lourenco, & Longo, 2012)的TTC。不过，研究者还未厘清个体低估威胁刺激的TTC的原因;对社会性威胁刺激(如，愤怒面孔)的TTC是否也存在低估的结果也不一致;通过在计算机显示器上放大刺激的方式模拟物体的靠近运动或多或少与实际碰撞情景不符。

针对以上问题与不足，本论文主要探讨关于TTC估计的两个问题:(1)为什么人们会更低估威肋、刺激的TTC(研究一)?(2)虚拟现实环境中对威肋、刺激TTC估计是否表现出类似的低估效应(研究二)?

研究一包含四个实验。首先，实验1重复了Vagnoni等(2012)的经典TTC估计实验，在实验中，被试需要想象每一个靠近的刺激消失后还是继续按照之前速度靠近，并在刺激与自身发生碰撞时做按键反应。结果发现，相对于非威肋、刺激(蝴蝶或兔子)，中国大学生同样更加低估威胁刺激(蜘蛛或蛇)的TTC;实验2想探讨对威胁刺激TTC的低估是否来源于个体对于威胁刺激的主观时距估计的扭曲。采用时间二分任务，要求被试判断刺激的呈现时间更接近先前学习过的较长的，还是较短的标准时间。如果被试对威肋、性刺激有普遍性的主观时距扭曲，那么他们对于威胁性刺激的时距判断将与对非威胁刺激有显著差异。结果发现，被试对于威胁刺激和非威胁刺激的主观时距知觉没有差异。因此，普遍的主观时距知觉扭曲无法解释个体对威肋、刺激TTC低估的现象;实验3的目的是探讨刺激的唤醒度对TTC估计的作用。该实验比较了两类高唤醒度的刺激的TTC估计，其中蛇与蜘蛛作为威肋、刺激，美食与钞票作为正性非威肋、刺激。结果发现，两类刺激的TTC估计无显著差异。与实验1结果的比较发现，被试不仅更低估威肋、刺激的TTC，也同等程度地更低估正性非威肋、刺激的TTC，因此唤醒度在刺激TTC低估中可能有重要作用;实验4采用愤怒，高兴和中性面孔图片作为刺激，要求被试完成同样的TTC估计任务。结果发现，相对于其他两类面孔，个体并没有更加低估愤怒面孔(社会性威胁刺激)的TTC。这表明人们并不是对所有威胁性刺激都更低估其TTC.

研究二包含两个实验。实验_5在VR环境下基本重复了实验1，但结果与实验1不同，总体上被试对威胁刺激和非威胁刺激的TTC估计不存在差异。但表现出性别差异，即仅女性被试对威胁刺激的TTC产生低估。实验6在VR环境下重复实验4，其结果与实验4一致，被试对三类情绪面孔的TTC估计仍然没有差异。

本论文发现，个体总是低估靠近物体的TTC。并且，相对于自然非威胁刺激(蝴蝶和兔子)，个体更加低估自然威胁刺激(蛇和蜘蛛)的TTC。然而，对社会性威肋、刺激(如愤怒面孔)的TTC的估计与社会性非威肋、刺激(高兴和中J陛面孔)没有差别。VR条件下，对蛇和蜘蛛的TTC低估效应有减弱趋势，并表现出性别差异。同时，VR条件下愤怒面孔的TTC与中性和高兴面孔的TTC没有差别。我们猜测，蛇与蜘蛛的威肋、性效应可能是由于其进化意义更为凸显，在生死枚关情境下，个体更需要为即将到来的威胁做出准备;另外，VR条件下TTC低估效应减弱可能是由于VR场景中深度线索比平面显示器要更清楚真实。总之，本论文揭示了情绪在个体时间知觉，尤其是碰撞时间估计中的重要作用。

Just imagine that a lion is pouncing on an antelope. The possible collision between the two means a threat for the antelope and a meal for the lion. For both, the estimation of time-to-collision (TTC) is very important for survival. According to different theories, estimation of TTC can be predicted by distance (between organism and object) and looming velocity; or depends on optic information from looming stimuli; or relies on the ratio of the visual angle subtended by the distance between any two points on an object divided by the rate of the change of this angle. From these points of view, the estimation of the TTC to approaching objects is mainly based on physical features at low level.

However, TTC estimation can be affected by emotion. For instance, threatening animal (spider and snake) pictures and scenes shortened estimation of TTC. But the reasons why human observers underestimate TTC for these threatening stimuli are still unclear. In addition, studies have investigated the judgment of TTC for threat facial expression (anger faces), but obtained inconsistent results. In addition, researchers always simulated the approaching movement by expanding the size of visual stimufi across different frames on computer displays in previous studies, which is different from the scene in the real life.

Regarding the disadvantages mentioned above, the current thesis addressed these two questions: (1)What cause people underestimate the TTC of threatening stimuli (Study 1)? (2) Whether the estimation of TTC for threatening stimuli in virtual reality environment shows similar underestimation effect (Study 2)?

There were four experiments in Study I. In Experiment 1 we intended to replicate the TTC judgement task used by Vagnoni et al. (2012). In the task, participants were instructed to imagine each stimulus continuing to approach them after it disappeared and press a button to indicate the moment when the stimulus would collide with them. It was found that Chinese college students also underestimated TTC of threatening stimuli (spiders or snakes) compared with non-threatening stimuli (butterflies or rabbits). In Experiment 2 we aimed to explore whether the underestimation of TTC for threatening stimuli resulted from their general distortion of temporal perception. A temporal bisection task was used. Participants were instructed to judge whether the duration of the stimuli presented was more similar to the standard long duration or to the standard short duration that they just learned. If the participants had a general distortion in temporal perception for threatening stimuli, their judgment of duration for threatening stimuli would tend to be longer than that for non-threatening stimuli. But the results showed that there was no difference between the judgment for the threatening and non-threatening images.

Therefore, the underestimation of TTC for threatening stimuli cannot be explained by a general distortion of temporal perception. The purpose of Experiment 3 was to explore the effect of arousal on TTC estimation. We used spider and snake pictures as threatening stimuli and an equal number of photographs of money and foods as positive no-threatening stimuli and the arousal of these two kinds of stimuli is equal. The participants were also instructed to perform the TTC judgment task. The results revealed that the estimation of TTC for threatening and positive no-threatening pictures had no significant difference, and both estimations were shorter than those for the threatening stimuli in experiment 1 .This result indicated arousal might play a key role on the underestimation of TTC. In Experiment 4, three types of emotional faces (anger faces for threat, happy and neutral faces for non-threat) were used in the TTC judgment task. The results showed that participants did not underestimate the TTC of angry faces (threat stimulus). This finding suggested that TTC is not underestimated for all kinds of threatening stimuli.

Study II contained two experiments. The purpose of Experiment 5 was to repeat Experiment 1 in the VR condition. Inconsistent with Experiment 1，we did not found any difference in TTC estimation between threatening and non-threatening stimuli among all subjects. Interestingly, there was a gender difference, i.e. only female subjects underestimated TTC of the threatening stimuli. The aim of Experiment 6 was to repeat Experiment 4 in VR condition. The results were in line with that of experiment 4, there was no difference in TTC estimation for the three emotional faces.

In this thesis, we found that looming stimuli shorten our TTC estimation. Compared with non-threatening stimuli (butterflies and rabbits), participants tended to underestimate TTC of natural threatening stimuli (spiders and snakes). However, there was no difference between TTC estimation of social threatening stimuli (angry faces) and social non-threatening stimuli (happy and neutral faces). Under VR conditions, the effects of snakes and spiders seem to decrease, and a gender differences was presented. Moreover, the TTC estimation of social threatening stimuli under VR condition was not different from that of social non-threatening stimuli. We supposed these results can be explained from an evolutionary perspective. The threat from snakes and spiders may be of prominently evolutionary implications. In life-and-death situations, individuals need to prepare in advance for the coming threats. In addition, the weakening of TTC underestimation under VR condition may be due to the fact that cues of depth in VR scenes are clearer and more authentic than common computer displays. Therefore, participants may make TTC estimation in different ways in these two conditions. In conclusion, this thesis revealed the important role of emotion in time perception, especially in TTC estimation.