A temporal dissociation of subliminal versus supraliminal fear perception: an event-related potential study.

Current theories of emotion suggest that threat-related stimuli are first processed via an automatically engaged neural mechanism, which occurs outside conscious awareness. This mechanism operates in conjunction with a slower and more comprehensive process that allows a detailed evaluation of the potentially harmful stimulus (LeDoux, 1998). We drew on the Halgren and Marinkovic (1995) model to examine these processes using event-related potentials (ERPs) within a backward masking paradigm. Stimuli used were faces with fear and neutral (as baseline control) expressions, presented above (supraliminal) and below (subliminal) the threshold for conscious detection. ERP data revealed a double dissociation for the supraliminal versus subliminal perception of fear. In the subliminal condition, responses to the perception of fear stimuli were enhanced relative to neutral for the N2 "excitatory" component, which is thought to represent orienting and automatic aspects of face processing. By contrast, supraliminal perception of fear was associated with relatively enhanced responses for the late P3 "inhibitory" component, implicated in the integration of emotional processes. These findings provide evidence in support of Halgren and Marinkovic's temporal model of emotion processing, and indicate that the neural mechanisms for appraising signals of threat may be initiated, not only automatically, but also without the need for conscious detection of these signals.

Mapping the time course of nonconscious and conscious perception of fear: an integration of central and peripheral measures.

Neuroimaging studies using backward masking suggest that conscious and nonconscious responses to complex signals of fear (facial expressions) occur via parallel cortical and subcortical circuits. Little is known, however, about the temporal differentiation of these responses. Psychophysics procedures were first used to determine objective thresholds for both nonconscious detection (face vs. blank screen) and discrimination (fear vs. neutral face) in a backward masking paradigm. Event-related potentials (ERPs) were then recorded (n = 20) using these thresholds. Ten blocks of masked fear and neutral faces were presented under each threshold condition. Simultaneously recorded skin conductance responses (SCRs) provided an independent index of stimulus perception. It was found that Fear stimuli evoked faster SCR rise times than did neutral stimuli across all conditions, indicating that emotional content influenced responses, regardless of awareness. In the first 400 msec of processing, ERPs dissociated the time course of conscious (enhanced N4 component) from nonconscious (enhanced N2 component) perception of fear, relative to neutral. Nonconscious detection of fear also elicited relatively faster P1 responses within 100 msec post-stimulus. The N2 may provide a temporal correlate of the initial sensory processing of salient facial configurations, which is enhanced when top-down cortical feedback is precluded. By contrast, the N4 may index the conscious integration of emotion stimuli in working memory, subserved by greater cortical engagement. Hum. Brain Mapping 21:64-74, 2004.