Abstract
Studies using continuous flash suppression (CFS) suggest that emotion-laden stimuli, such as fearful faces, are prioritized when presented outside of awareness. For example, fearful faces gain faster access to awareness from CFS suppression than other expressions (Yang, Zald & Blake, 2007) and emotionally salient images rendered invisible by CFS guide spatial attention to a subsequent target (Jiang, Costello, Fang, Huang & He, 2006). Such findings have been attributed to a specialized, sub-cortical pathway that evaluates the emotional significance of stimuli, without awareness, to direct processing resources (Tamietto & de Gelder, 2010). In two studies, we tested whether the emotional content of faces presented under CFS (i) provides prioritized access to awareness or (ii) directs spatial attention. In Experiment 1, control images were created via spatial inversion and contrast sign reversal to produce stimuli that maintain the low-level image properties of normal facial expressions, but whose recognizable emotional content is vastly reduced. Normal and control face stimuli were presented briefly (800 milliseconds) under CFS. Fearful faces broke suppression more often than other facial expressions. However, in accordance with previous work (Gray, Adams, Hedger, Newton & Garner, in press) a similar effect was found with the control stimuli, suggesting that this "fear advantage" is explained by low-level image differences across expressions rather than by unconscious evaluation of emotional content. In Experiment 2, we employed an attentional cuing paradigm with CFS to examine the impact of unconsciously presented face stimuli on spatial attention. Fearful faces modulated attention in trials where they broke suppression, such that discrimination of a subsequent, co-located probe was improved. However, facial expression failed to modulate attention when observers were unaware of stimuli. Data from both experiments cast doubt on the notion that the emotional meaning of faces is processed unconsciously.
Meeting abstract presented at VSS 2013