Eye-catching: Right hemisphere attentional bias for emotional pictures

Abstract

To examine whether healthy individuals spontaneously direct attention to emotional pictures, eye movements were monitored continuously during the presentation of pairs of emotional and neutral pictures. A right hemisphere advantage was expected for initial orienting to emotional pictures. When an emotionally arousing picture was presented in the left visual hemifield, i.e., to the right hemisphere, the initial saccade was more often issued towards it, irrespective of valence. Negative pictures did not lead to avoidance throughout the 8-second trials; rather, both positive and negative valence held attention. This study adds to the evidence that overt attention to visual stimuli is boosted by emotion and that there is a right hemisphere advantage for orienting to emotional cues.

Acknowledgements

I thank Nora Walz for her help with the data collection, Dr. Jochen Müller for important discussions on the topic, Antje Gerdes for comments on a previous version of the paper, and most importantly, Prof. Dr. Paul Pauli for his continuous academic support.

Notes

¹Positive: 1600, 4659, 7200, 8030, 8400, 8021, 8041, 7410, 8510, 8502, 8420, 8370, 8034, 8033, 2080, 1999, 1640, 1650, 4250, 4599, 7330, 8080, 4002, 7400, 7580, 2250, 4653, 4660, 7350, 8180, 7450, 4611, 4680, 8200, 4614. Negative: 1070, 2800, 3000, 3230, 9300, 3060, 6212, 6560, 9433, 1090, 3010, 3350, 6020, 2205, 3500, 6312, 6550, 9040, 9250, 1300, 3261, 3150, 3530, 9405, 3400, 6313, 6360, 9253, 2120, 3100, 3130, 6230, 9490, 3071, 6510. Neutral: 5731, 2190, 7500, 7004, 5510, 5120, 2221, 2200, 7000, 7020, 7150, 7006, 7185, 7009, 7175, 5520, 5130, 7010, 7050, 7160, 7234, 2850, 7031, 7060, 7090, 7002, 7130, 7040, 7950, 2840, 7110, 5500, 7080, 7170, 7034, 9210, 7207, 7560, 2410, 1390, 1670, 2210, 2620, 2870, 2880, 2890, 5250, 5390, 5530, 5720, 5531, 5750, 5900, 6150, 7035, 7100, 7140, 7190, 7205, 7233, 7490, 7491, 7550, 7620, 7640, 7710, 7820, 7830, 7920, 8311.

²We tested the raw ratios in this case because a number of participants always initiated the initial fixation to the emotional picture when it was presented on the left, and the planned Fisher-z transformation would have led to a loss of these participants. However, we verified in the follow-up tests that the significant ratios remained highly significant when they were Fisher-z transformed: positive–neutral, t(23) = 5,756, p<.001; positive–neutral, t(20) = 6,725; p<.001. Moreover, positive results remained highly significant when a Bonferroni adjustment was carried out.

³Dwell time data were confirmed by the number of fixations on each of the two pictures. In the repeated measures ANOVA, the significant intercept indicated that overall the emotional preference ratios were significantly different from zero, F(1, 27) = 50.55, p<.001. The significant Emotion effect indicated that there were more fixations on the negative pictures than on the positive ones both relative to neutral pictures, F(1, 27) = 6.52, p=.017. There was no significant Location effect and no Location by Emotion interaction.

⁴A MANOVA of average luminance, red, green, and blue obtained from Adobe Photoshop® histograms revealed significant differences, F(8, 270) = 3.22, p=.002. Follow-up tests showed that positive pictures had higher values on all variables than negative and neutral pictures, which did not differ. An ANOVA with standardised RMS contrast as calculated according to Bex and Makous (2002) showed a significant difference between emotions, F(2, 280) = 3.42, p=.036. Post hoc Scheffé tests showed that negative pictures had a higher contrast (M=0.797) compared with positive pictures (M=0.621), p=.039; neutral pictures were in between (M=0.729).