Abstract: Compared with neutral stimuli, emotional stimuli usually gains a cognitive processing priority. In studies of investigating the relations between attentional and emotional processing, emotional materials are observed to cause attentional bias which leads to faster and more effective reacts to emotional materials than neutral ones for participants. However, it is still unsolved that how dose the emotional bias happen. Researchers have been trying to investigate this issue but got no consistent conclusions until now, e.g., some of them suggest that emotional bias comes from attention captures, while others insist that emotional bias derives from disengagement of attention. In the current study, we recorded undergraduate students’ eye movements by EyeLink II eye tracker to explore how emotional pictures induce emotional bias. Emotional scene pictures (22 pictures for positive, negative and neutral scenes each) selected from the Chinese Affective Picture System (CAPS) were used as experiment materials. An antisaccade task and Go/No-go tasks were adopted in Experiment 1 & 2 with 12 and 13 participants respectively. In Experiment 1, for antisaccade task, when a picture (positive, negative or neutral) appeared on the left or the right of the central fixation cross, participants were required to ignore the picture and fixate its mirror location (on its opposite side to the central fixation cross) as soon and precise as possible. In Experiment 2, for Go/No-go task, a picture (positive, negative or neutral) appeared at screen center together with a square displayed at its left or right side with blue or yellow color as the condition clue, which means that the Go and No-go conditions were determined according to the color of the square for participants, and the corresponding relationship between the colors and the meanings of the squares was balanced between subjects. In the Go condition, participants were asked to move their eyes to the color square from screen center as soon as possible, while in the No-go condition, participants were required to ignore the color square and staring at the central picture. We collected correct trial latencies and directional error rates for both experiments. The correct trail latency refers to the time interval between the target onset and the start of the first saccade which is made to the correct direction. The directional error rate refers to the percentage of the first saccades which are made to the incorrect direction accounted for the total number of the fist saccades. Repeated measures ANOVAs of emotion type (positive, negative and neutral) and picture position (left and right) were conduced on the above mentioned eye movement measures for Experiment 1 (antisaccade task). The results showed that there was a main effect of emotion type on correct trial latencies, F (2，11) =8.39，p<.05, ?2=.43, post-hoc analyses further suggested that negative pictures got longer saccade latencies than neutral pictures, p<.05. The main effect of emotion type on directional error rates was also significant, F (2，11) =7.68，p<.05, ?2=.41, further post-hoc analyses revealed higher saccade error rates for emotional pictures (positive and negative) than neutral pictures, ps<.05. These results implying that the emotional pictures can capture participants’ attention even though they were asked not to look at them. The interaction between emotion type and picture position on direction error rates was also significant. F (2，11) =12.36，p<.05，?2=.53, Simple effects analysis showed that the directional error rates was higher for positive pictures displayed on the right side than left, p<.05, while the directional error rates showed no difference between negative pictures appeared on the right and left sides. Repeated measures ANOVAs of emotion type (positive, negative and neutral) were conducted for Experiment 2. In the go condition, there was a main effect of emotion type, F (2，11) =4.54，p<.05, ?2=.29, which showed that, compared with neutral pictures, when emotional pictures (positive and negative) were presented in screen center, participants moved their eyes towards the peripheral targets more slowly. In the no-go condition, the participants were less likely to make saccades to the peripheral target squares which were indicated from a main effect of emotion type on directional error rates, F (2，11) = 44.59，p<.05, ?2=.80. The results of Experiment 2 reflects that emotional pictures can hold attention and participants disengage their attention from emotional pictures more difficultly. In conclusion, we found that emotional pictures capture participants’ attention more quickly, and it is more difficult for participants to disengage their attention from emotional pictures compared with neutral stimuli, which indicated that emotional information affects both the capture and the disengagement of attention. In addition, there was a brain lateralization of attentional bias for positive pictures but not for the negative and neutral pictures.

Key words: emotional pictures, attentional bias, eye movements, antisaccade, Go/No-go

摘要： 注意实验中，与不带情绪色彩的刺激相比，具有情绪意义的刺激可引起注意偏向。本研究以情绪场景图片为材料，通过眼动技术分别记录被试在反向眼跳任务和Go/No-go任务中的眼动数据，考察了情绪图片的注意偏向。结果发现：反向眼跳任务中，对情绪图片的眼跳潜伏期比中性图片更长，眼跳错误率也更高；在Go任务中，相对于呈现在中央的中性图片，情绪图片引起的对靶子的眼跳潜伏期更长；在No-go任务中，相对于呈现在中央的中性图片，情绪图片引起的眼跳错误率更低。实验结果说明，情绪图片可引起注意偏向，表现为更快的捕获注意并且注意更难从情绪图片上转移。

关键词: 情绪图片, 注意偏向, 眼动, 反向眼跳, Go/No-go