Citron, F. M. 1 , Gray, M. A. 3, 2 , Critchley, H. D. 2 , Weekes, B. S. 4, 1 & Ferstl, E. C. 1
1 School of Psychology, University of Sussex, Brighton, UK
2 Psychiatry, Brighton & Sussex Medical School, Universities of Brighton and Sussex, Brighton, UK
3 Experimental Neuropsychology Research Unit, School of Psychology and Psychiatry, Monash University, Victoria, Australia
4 Faculty of Education, The University of Hong Kong, China
Human emotions can be conceptualized within a two-dimensional model constituted by emotional valence and emotional arousal (intensity). Few neuroimaging studies have manipulated each dimension independently, in contrast to many previous emotion-specific or valence-driven neuroimaging studies. The present study manipulated valence and arousal orthogonally using the attributes of verbal stimuli measured from statistics taken from a large corpus of behavioural ratings (Citron et al., 2009). Lexico-semantic properties known to affect word processing (e.g. word frequency, imageability) were carefully controlled. Written emotion words were presented for identification during an fMRI experiment to examine differences in brain activation associated with known effects of emotional arousal and valence on word recognition.
Behaviourally, a recognition advantage for positive words was observed, which replicates previous findings suggesting that positive words are more richly interconnected within the mental lexicon and therefore easier to process. However, we also found that, words eliciting “conflicting” orientations (positive and highly arousing, or negative and low in arousal) elicited greater BOLD signal activation in insula cortices than words eliciting congruent orientations, thus extending earlier behavioural observations. Our findings validate a two-dimensional approach to the study of emotion and word recognition and show the dimensions of valence and arousal interact during word recognition within brain regions that are also implicated in motivational decision-making.