A number of control analyses were performed to corroborate these findings.
As illustrated in Fig. 4.2, pleasant, neutral, and unpleasant prime pictures were associated with the expected pronounced differences in posterior negativity. One might therefore hypothesize that the effects of the emotional content of the prime stimuli on the processing of target images, rather than revealing reduced processing resources, were due to differences in baseline ERP activity before target picture onset.
However, the emotional prime and target effects were opposite in directions and calculating the baseline with respect to the prime picture is a conservative approach, rather underestimating the prime effect of emotional pictures. Consistent with this reasoning, re-analysis of the data with a baseline locked to the target pictures closely replicates the above reported findings, although the emotional prime effect appears enlarged in amplitude. In order to circumvent the problem of baseline-differences, we calculated a baseline-free estimate of the posterior negativity by determining the gradient (slope) of the negative potential based on the first derivative of the ERP waveform. Confirming the emotional modulation of target images reported above, emotional target pictures were associated with significantly steeper gradients of the ERP wave compared to neutral contents. Interestingly, this analysis also closely replicated the effects ascribed to the emotional content of prime pictures. Specifically, emotional prime pictures were reflected in current target processing by less steep ERP gradients compared to neutral primes. Taken together, determining the slope of the posterior negativity fully replicated the reported results for the amplitude measure using a baseline-free estimate of the scalp potential.
Discussion
Previous studies utilizing rapid picture presentations have established that emotionally significant stimuli are discriminated from neutral pictures, with both pleasant and unpleasant pictures prompting increased occipital negativity in a 150 - 300 ms time window (cf., Junghöfer et al., 2001; Schupp et al., 2003a). These results were replicated in the present study: target pictures that were pleasant or unpleasant prompted greater occipital negativity than neutral pictures, despite the rapid rate of picture presentation. Moreover, the present study found that the magnitude of posterior negativity for a target picture varies systematically as a function of the preceding prime. When a prime picture was emotional (and itself eliciting heightened posterior negativity), posterior negativity of the current target picture was reduced. This effect of an emotional prime was identical regardless of whether the target picture was pleasant, neutral, or unpleasant. Thus, the data reveal no evidence for affective priming in the perceptual domain. Rather, irrespective of whether the hedonic valence of the prime matched or mismatched the target, occipital negativity was decreased if the prime picture was affectively engaging. Taken together, these results demonstrate that enhanced processing of affective primes diminishes subsequent target processing, regardless of the hedonic content of the target.
The novel finding of the present study is that current target processing is affected not only by the emotional content of the target image but also systematically varies with the emotional content of the preceding prime picture. Previous studies already suggested that current emotional pictures draw emotional resources to a greater extent than neutral images (Schupp et al., 2003b). The present findings imply that the capture of processing resources extends in time and interferes with successively presented stimulus materials. These data are also consistent with the notion that successively presented pictures are characterized by a distinct neural representation, which may compete for a limited pool of processing resources (cf., Kastner & Ungerleider, 2000; Keysers & Perrett, 2002; Potter et al., 2002). From the perspective of competition for resources, recent studies explored the hypothesis of diminished capacity for selective emotion processing predominantly by presenting emotional stimuli and non-emotional task stimuli concurrently (Anderson et al., 2003;
Pessoa et al., 2002b; Vuilleumier et al., 2001). For instance, a recent ERP study observed that a frontal positivity (100 ms) elicited by fearful faces was only present
when the faces were presented at attended locations while absent when presented at non-attended locations (Holmes et al., 2003). Similar findings have been observed in another ERP study exploring the processing of emotional facial expressions (Pourtois et al., 2004). Thus, selective emotion processing might be attenuated by concurrently presented stimuli commanding processing resources.
While broadly consistent with these findings, the present study design differs in several ways from these studies that merit reemphasis. Rather than being limited to concurrent stimulus processing, the present results suggest that processing resources allocated to target pictures systematically vary with the amount of processing resources captured by preceding emotional prime stimuli. Furthermore, previous studies assessed competition effects mainly by introducing an explicit attention task.
In contrast, the pattern of results was observed here in the absence of any active task instruction, presumably reflecting the implicit capture of attention by emotional cues (cf., Öhman et al., 2000a). Finally, it has to be stressed that interference effects on selective emotion processing affects different ERP components depending on experimental manipulation and task demands. The present findings demonstrate interference effects of current target processing due to emotional prime pictures in the first ERP component sensitive to the emotional content of complex natural pictures.
The similarities of increased posterior negativities brought out by either emotional significance or explicit attention to selected stimulus categories and characteristics has been noted previously (cf., Schupp et al., 2003b), contrasting with spatial attention effects affecting considerably earlier ERP components (cf., Hillyard et al., 1995).
Thus, interference effects in selective emotion processing may be reflected by specific ERP components assessing various processing stages depending on experimental context (cf., Luck et al., 2000).
The present data provided no evidence for affective priming effects in the first 300 ms of perceptual processing. Specifically, the processing of pleasant or unpleasant cues was not modulated when presented after pictures with the same compared to opposite emotional valence. If at all, valence effects are suggested in that prime and target effects appeared more pronounced for pleasant pictures. However, this effect might be due to the large number of high-arousing erotic images in the present study, which has been shown in previous research to strongly engage various emotional response channels (cf., Bradley et al., 2001; Sabatinelli et al., 2005; Schupp
et al., 2004b). Furthermore, affective priming was also absent when analyzing longer series of picture presentations, i.e., up to three repetitions of the same picture valence.
Similarly, affective priming effects were not observed for the latency of the occipital negativity, which did not vary as a function of prime and target valence (see Fig. 4.5).
These findings contrast with a recent ERP study reporting early affective priming effects. Specifically, while fearful and happy faces were not differentiated by ERP components, priming happy and fearful faces with expressions of the same compared to different valence was reflected by a relative negative deflection around 150 ms post-stimulus (Werheid et al., 2005). Several issues render a comparison of both studies difficult. With regard to experimental methods, Werheid and colleagues implemented an active task set, possibly augmenting affective priming effects (see Klauer & Musch, 2003). However, the priming effects reported by Werheid et al. were not observed to affect an ERP component sensitive to emotional content. Thus, it is possible that the effects reflect differences in resource competition among pairs of same/different facial expressions rather than the emotion expressed by the faces.
Support for this caveat is provided by a recent study demonstrating competition in perceptual processing (reduced N170 amplitudes) when faces were presented concurrently with artificial objects (“greebles”) trained to expertise (cf., Rossion et al., 2004). The absence of affective priming effects in the present study may be informative given specific features of the study. The stimulus materials included pictures of high emotional intensity depicting representative exemplars of human emotional experience, which are known to reliably engage a wide spectrum of emotional responses (Lang et al., 1997). The inclusion of neutral control stimuli enabled the establishment of selective emotion processing of the stimuli, i.e., increased posterior negativity to emotional target stimuli. Emotional prime effects were seen as reduction of the posterior negativity, demonstrating interference by emotional primes irrespective of emotional valence rather than facilitation of processing by same-valence primed target cues. Clearly, these data provide no support for the notion of affective priming at the level of early perceptual encoding. However, it needs to be determined whether later components of the ERP waveform (late positive potential, N400), which were not assessed in the present study, reflect affective priming effects (cf., Cuthbert et al., 2000; Schirmer & Kotz, 2003).
In summary, the present results extend the concept of motivated attention by considering emotional stimulus processing in the visual cortex as a function of the hedonic valence of preceding stimuli. Previous studies observed the rapid and facilitated processing of emotional cues even when stimuli are presented as a rapid and continuous stream. This notion was extended here by demonstrating that emotional prime picture processing interfered with the processing of subsequently presented pictures. Future studies need to determine whether resource competition among successively presented pictures provides a useful conception for the present findings.