Attentional Bias of Spatial Location in Viewing Scenes:Evidence From Eye Tracking Research
Yang Meng,Liu Danwei,Cai Anni,Zhou Renlai
2012-01-01
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.