From the work of Esteves & Öhman (1993), visual masking has been used to explore the processing of facial expressions at different levels of conscious awareness with growing evidence leading to the conclusion that emotion processing is automatic and can occur without conscious awareness. fMRI and patient studies of these phenomena have resulted in the putative sub-cortical pathway to the amygdala for threat related stimuli to be processed without conscious awareness. However, little is known about the temporal evolution of such processing.
We investigated this in an electroencephalographic study using a backward masking paradigm. Participants were presented with six different identities each displaying three facial expressions of emotion: fear, disgust and happiness and asked to categorize them by expression. Stimuli were displayed for 8, 24, 40 or 64ms and immediately replaced by a 200ms noise mask (with the same phase structure of a face, ie. an inverse square relation to frequency).
Considering two key face processing ERP components, the N170 and P3 complex, we found effects of both presentation time and expression. Specifically, for all expressions, we observed a parametric relationship between presentation time and the amplitude of the two components, with largest N170 amplitudes occurring to the most briefly presented stimuli and largest P3 amplitudes to the stimuli presented for the longest times. This double dissociation highlights key differences in the processing of facial expressions of emotion at different levels of awareness. Furthermore, differential processing of the three facial expressions was observed as a function of stimulus presentation time, with significant differences in all three expressions for the longest presentation times and only for fear for short presentation times.
These results will be discussed in the context of current theories on the processing of facial expressions of emotion.
Esteves, F., Öhman, A. (1993). Scand. J. Psychol, 34:118.