Primacy of emotional vs. semantic scene recognition in peripheral vision

Abstract

Emotional scenes were presented peripherally (5.2° away from fixation) or foveally (at fixation) for 150 ms. In affective evaluation tasks viewers judged whether a scene was unpleasant or not, or whether it was pleasant or not. In semantic categorisation tasks viewers judged whether a scene involved animals or humans (superordinate-level task), or whether it portrayed females or males (subordinate-level task). The same stimuli were used for the affective and the semantic task. Results indicated that in peripheral vision affective evaluation was less accurate and slower than animal/human discrimination, and did not show any advantage over gender discrimination. In addition, performance impairment in the peripheral relative to the foveal condition was greater or equivalent for affective than for semantic categorisation. These findings cast doubts on the specialness and the primacy of affective over semantic recognition. The findings are also relevant when considering the role of the subcortical “low route” in emotional processing.

Keywords:

• Emotion
• Recognition
• Scene perception
• Peripheral vision
• Subcortical
• Primacy

Acknowledgments

This research was supported by Grant PSI2009-07245 from the Spanish Ministry of Science and Innovation to MGC, Aivo Aalto Consortium grant from the Aalto University, and Academy of Finland Grant #217995 to LN.

Notes

¹A 150 ms display at a 5.2° distance from the fixation point has been found to prevent fixations on peripheral pictures (Calvo, Nummenmaa, & Hyönä, 2008). In that study, the mean latency of the first saccade towards the picture was 175 ms (hence above the 150 ms display duration used here); the probability of fixating the picture during the 150 ms display was less than 1%; and whenever fixations on the scenes occurred, their mean duration was only 4 ms.

²As acknowledged by Nummenmaa et al. (2010), there is obviously no universal answer to the primacy of affect versus cognition in the most general sense. Some cognitive operations may occur earlier (e.g., visual recognition; this study) and some are bound to occur later (e.g., abstract reasoning) than affective processing. In the context of the present study, this primacy issue is applicable to (a) the initial processing of sensory input in the visual domain, and (b) broad categories (pleasantness vs. unpleasantness, animal vs. human, female vs. male), with all involving biologically and socially relevant stimuli.

³Nevertheless, Nummenmaa et al. (2010) found that, at very brief display durations (20 ms) where semantic categorisation was already possible from the first presentation of a novel visual scene, affective evaluation was not, even after the third repetition. Only at 40 ms exposure was affective valence accurately recognised above the chance level. So, even if affective processing could be accomplished earlier than semantic processing, a representational product of it cannot be brought to consciousness before the information that is produced by semantic processing.