Abstract: The processing of emotion-laden stimuli is often proposed to take place in a prioritized and obligatory fashion. The evidences from both electrophysiological and neuroimaging studies in which the spatial focus of attention is explicitly manipulated have showed controversial results on the subject of whether attention is required for emotional stimuli processing. Although this issue can be partially reconciled by adapting the concept of attentional load, the manipulation of attentional resources may not be subtle enough to effectively restrain the privileged processing of emotional stimuli. The present study is aim to further investigate the mechanism of emotion processing by combining the Load Theory and attentional blink (AB) paradigm. We hypothesize that in attentional blink paradigm the prioritized processing of emotional T2 stimuli may be suppressed by the double limitation of increased T1 perceptual load and short latency of T1-T2 interval.
In present study, each trial consisted of a first target (T1, a scrambled face colored in green) and a second target (T2, either a fearful or a neutral face) embedded in a rapid series visual presentation (RSVP) of 18 scrambled faces. The Eriksen flanker task was used to distinguish the level of T1 perceptual load in which the participants were asked to determine the orientation of central arrow between other 4 congruent (low-load condition) or incongruent (high-load condition) arrows. All 20 items were presented for 67ms on the black background of computer screen and immediately followed by sequential items. The possible intervals between T1 and T2 were Lag-2 (SOA 134ms), Lag-3 (SOA 201ms), Lag-5 (SOA 335ms) and Lag-8 (SOA 536ms). 30 undergraduate and graduate students were instructed to make the identification response to the central arrow of the only green item (left or right) and the detection response to the other target face with intact features (present or absent). T1 load conditions were separately presented in two blocks and the order of these two blocks was counterbalanced across the participants. The T1 identification accuracy and the T2 detection accuracy in all conditions were recorded respectively.
The analysis of behavioral data revealed that for low-load conditions, fearful faces were detected more often than neutral faces, therefore replicating previous reports of the privileged emotion processing in AB. However, this advantage was hampered significantly in the high-load condition and the detection of neutral faces was not affected by the increased T1 load, suggesting that the privileged access of fearful faces to awareness is more sensitive to the current available processing resources. Most importantly, the attenuated emotional impact in AB was merely observed in the condition of high T1-load and short T1-T2 lag, indicated that the prioritization of emotion-laden stimuli processing are restricted by both the depletion of attentional resources induced by T1 perceptual load and the ineffective modulation of positive attentional set mechanism toward emotional stimuli.
In conclusion, the present study demonstrated that faces with emotional expressions are not a category of objects which immune to the effects of attention. Instead, the processing of emotional stimuli such as facial expression depends on the available attentional resources and is modulated by attentional control.

Key words: emotion, attention, perceptual load, attentional blink