自由观看实景图片时注意的空间偏向

杨萌1，2 刘丹玮1，2 蔡安妮3 周仁来＊＊1，2

( 1 北京师范大学心理学院应用实验心理北京市重点实验室，北京，100875)

( 2 北京师范大学认知神经科学与学习国家重点实验室，北京，100875) 

(  3 北京邮电大学多媒体通信与模式识别实验室，北京，100876)

摘要 本研究采用眼动仪直接记录40 名在校大学生及研究生(  平均年龄21． 9 岁，SD = 1． 8) 自由观看实景图片的眼动情况，考察自然状态下视觉注意在真实场景中的空间偏向。为了控制图片不同区域刺激特征显著性对结果的影响，将被试随机分为两组，一组观看正向图片，一组观看倒置图片。对图片四个象限上的首次注视概率、首次眼跳方向和注视时间百分比分别进行2( 水平方向) ×  2( 垂直方向) ×  2( 呈现方式)  的重复测量方差分析，结果表明，存在独立于刺激特征的空间注意偏向，表现在水平方向具有左侧注意偏向，垂直方向具有上侧注意偏向。

关键词视觉注意偏向空间位置实景图片眼动

Attentional Bias of Spatial  Location in Viewing Scenes:Evidence  From Eye Tracking Research

Yang Meng1，2，Liu Danwei1，2，Cai Anni3，Zhou  Renlai1，2

( 2 Beijing Key Lab of Applied Experimental  Psychology，School of Psychology，Beijing Normal  University，Beijing，100875)

( 1 State Key Lab of Cognitive Neuroscience and  Learning，Beijing Normal University，Beijing，100875)

( 3 Multimedia Communication and Pattern Recognition  Labs，Beijing University of Post and  Telecommunications，Beijing，100876)Abstract  There are two main types of visual-field  asymmetries: right /left hemi-field asymmetries and upper /lower  hemi-field asymmetries. In view of scenes differing from other types of visual  stimuli chosen by previous papers in a number of important ways，it is still an  open question whether there is an overall attentional bias of horizontal and  vertical visual fields in viewing natural scenes. 40 college students  ( 20  females) viewed 40 real-world pictures with free viewing  allowed，and their eye movements were monitored by Tobii T120. In  order to control the effect of feature saliency，half of the  participants viewed the normal picture while the other half viewed its mirror  reversed in the horizontal and vertical orientation. The dependent variables  were locations of the first fixation，as a measure of  endogenous attentional bias before the pictures showed up ( to make sure of  that，there was a 600-800ms blank screen between the central  fixation cross and picture) ; initial saccade direction，as a survey of the  initial distribution of attention during observers’first fixations on  pictures; and fixation duration，as a reflection of  visual asymmetries in the whole process. To determine whether the presence of  pictures would affect the result，a repeated measure  analysis of variance ( ANOVA) with horizontal (  left and right) and  vertical( upper and lower)  bias as the within-subject factors and the  presence mode ( normal and reversed) as a  between-group factor was designed. With regard to locations of first  fixation，there was a significant main effect of horizontal  bias，F( 1， 38) = 10. 437，p ＜. 01，η2p = .  215，sp  = . 883. However，there was no  vertical bias differences，and no significant interaction was found between any two  or among three variables except the one between vertical bias and presence  mode，F( 1， 38) = 4. 123，p ＜ . 05，η2p = .  098，sp  = . 508. The simple effect test  demonstrated that when the observers viewed the normal pictures，they showed a  striking upward bias，F( 1， 38) = 7. 756，p ＜ . 01，η2p = .  170，sp  = . 775; when they viewed  the reversed pictures，the upward bias disappeared. With regard to the initial  saccade direction，there was a significant main effect of horizontal  bias，F( 1， 38) = 5. 573，p ＜ . 05，η2p = .  128，sp  = . 633. However，there was no  vertical bias difference，and no significant interaction between any two or among  three variables except the one between vertical bias and presence  mode，F( 1， 38) = 7. 143，p ＜ . 05，η2p = .  158，sp  = . 740. The simple effect test  demonstrated that when the observers viewed the normal pictures，they did not show  vertical bias; when they viewed the reversed pictures，upward bias  appeared，F( 1， 38) = 25. 497，p ＜ . 001，η2p = . 402，sp = . 998. With regard to fixation duration，there was a  significant main effect of horizontal bias，F( 1， 38) = 11.  540，p  ＜ . 01，η2p = . 233，sp = . 911，and vertical bias  F(  1，  38)  = 5. 037，p ＜ . 05，η2p = .  117，sp  = . 590. No significant interaction was  found between any two or among three variables. Our study revealed that there  are leftward and upward attentional biases in viewing natural scenes. The  present findings play an essential role in driving，aviation，advertisement and  other fields where visual attention is important to safety and  efficiency.Key words visual attentional bias，spatial  location，natural scene，eye  movementJournal of Psychological Science  2012，35( 2) : 258 －  263