5.4. General Discussion

Mathews and MacLeod (2002) noted that attentional biases to negative information are observed only when at least two competing processing options are present. Mathews and MacKintosh (1998) supported this assumption, proposing that internal representations corresponding to the two (or more) processing options compete for attention (modeled as activation levels) via inhibitory links. According to this hypothesis, controlled task demands can increase activation of the intended target, which then inhibits attention to irrelevant distractors such as the emotional content in an emotional Stroop task. But in dysphoric individuals, threatening information can receive additional activation from a postulated threat evaluation system, and so will compete more effectively for attention in reducing the efficiency of target processing (Mathews & MacKintosh, 1998; Mathews & MacLeod, 2002). This attentional pattern, associated with a vigilant processing mode, reduces target processing efficiency. Normal or non-dysphoric participants may more effectively inhibit attention to negative stimuli irrelevant to task; there is evidence that they orient attention away from, and easily disengage from, a threatening stimulus. However, no particular pattern of attention to threatening stimuli is observed in a task, such as the emotional Stroop task, that includes direct processing competition between a target and an emotional irrelevant stimulus. We suggest that the absence of an effect of emotion in normal participants within an emotional Stroop task is due to the direct competition between strong target processing and an emotional distractor; strong target processing may override an irrelevant distractor, regardless of its emotional valence.

In the present study, we utilized a modified emotional Stroop task in which the target and the emotional distractor were processed separately with an ISI of 90 ms. This could permit emotional content to be processed more efficiently, without the direct competition or interception of target processing. As expected, with an ISI of 90 ms between the emotional face and colour patch, subjects exhibited faster colour-naming latencies after processing a negative face. That is, when an irrelevant distractor could be processed and target processing was postponed, normal participants increased their target processing efficiency, allocating more attention to target processing with negative information. The processing style observed with negative information at the 90-ms ISI level is consistent with an account that posits avoidant attention when an emotional task-irrelevant stimulus competes with a neutral target stimulus, because participants avoid negative information and allocate more attention to the target. The controlled task demand processing receives additional activation and competes more effectively with a negative stimulus input for attentional resources.

Interestingly, participants showed the opposite pattern of attention in response to positive information. They exhibited a vigilant processing style with the positive information at the 90-ms ISI level; they may have allocated more attention to positive information (relative to neutral or negative information), thus reducing colour-naming efficiency. These results are similar to those of Bradley, Mogg, Falla and Hamilton (1998), who used a dot probe task with positive, neutral and negative faces in normal, low-anxious individuals, although their results implicate an initial orienting of attention in competition between task-irrelevant stimuli.

This effect, however, disappeared in the null ISI condition; that is, when there was strong competition between processing targets and irrelevant distractors. These results strongly suggest that ongoing controlled task demand processing per se can eliminate the influence of negative information when there is direct competition. The absence of an emotional Stroop effect in normal participants can also be explained by this processing pattern. Finally, contrary to the attentional bias effect in dysphoric individuals, no such attentional bias effect in non-dysphoric individuals was observed when there were directly competing processes.

The pattern of attention to negative stimuli can be changed dynamically by emotional events or contexts. As the results of Experiments 1 and 3 indicate, normal participants are likely to employ negative information avoidance as a default processing mode. This style seems to be used more frequently after positive pictures have been presented, according to the results of Experiment 4, because the colour-naming latency pattern observed with positive emotional induction is very similar to that demonstrated in Experiments 1 and 3 at the 90-ms ISI level. After presentation of negative pictures, however, negative face priming reduced target processing efficiency. Participants were more likely to attend to the negative face after viewing negative pictures, and to allocate less attention to task processing. This pattern is consistent with the “vigilant” processing mode discussed by Williams et al. (1997) and Mathews and MacLeod (2002). Normal participants, however, might dynamically change their attentional processing style in response to emotional information across environmental emotional contexts.

LeDoux (1996) noted that an emotion can change the responsiveness to emotional stimulus input and can facilitate or inhibit the detection of threatening information. Izard (1993) conceptualized emotions as states of readiness for adaptive action with two opponent motive systems: an appetitive system that governs approach behavior and engagement with the environment and a defensive system that promotes avoidance and protection against danger. Survival in the natural environment depends on a dynamic balance between the two (Lang, Bradley & Cuthbert, 1997). However, this dynamic balance does not seem to function well in affective disorders; in clinically dysphoric populations, the level of threat sufficient to activate a vigilant processing mode seems to be set too low (Mathews & MacLeod, 2002). In these individuals, the vigilant processing mode is the more dominant processing style, used in response to negative information, regardless of current controlled goals or the environmental context. Individuals from the clinically normal population are more likely to dynamically change their negative information processing styles across emotions and environmental contexts or events, which constitute an ability with adaptive value for the individual.

The task utilized in the present study seems more sensitive to attentional patterns in normal participants in relation to negative information than the emotional Stroop task, but there are still many areas that need further investigation. First, although we found a common tendency for response latencies with a negative face to be faster than those with a neutral face, the processing of emotional stimuli seems to be very sensitive to environmental context, and this needs additional research.

Second, we found a relatively strong impact resulting from a positive face; normal participants seem to pay more attention to positive information, which can lead to more interference with target processing, as compared to negative or neutral stimulus priming. Unfortunately, there is very little research on the influence of positive stimuli on immediate attention, and more is needed.

Third, the emotional stimulus presentation method utilized, and its duration, can also produce very different results (e.g. Fox, Rosso & Dutton, 2001). In the present study, we utilized 30 ms of presentation time for the emotional face, and 90 ms of ISI between emotional and target stimuli, to measure the immediate emotional impact on attention in preventing the processing of emotional information being influenced by participant’s top down control. This method permits us to measure the emotional effects on normal participants, but since only 52% of participants correctly noticed that there were three faces, this presentation time was not entirely suitable neither for conscious processing nor for non conscious processing. It will be interesting to see how emotional effect can be modified by varying stimulus presentation durations and intervals, differentially implicating conscious and pre-conscious processing.

Emotion and cognition are the different mechanisms that enable an organism to adapt socially and biologically to meet environmental challenges. They both mutually influence and compete with each other to control information processing. For example, when a mildly threatening stimulus is present, normal participants with neutral emotions can easily or automatically (as their default processing mode) inhibit or ignore the threat when it is not relevant to the processing goal, but when a very aversive stimulus is present, processing of this kind of stimulus can intercept processing and control, and overtake awareness. Furthermore, when an individual becomes anxious or aroused by his or her environmental emotional situation or context, this also modifies the processing pattern in response to emotional information, even when that emotional stimulus cannot be consciously identified. This dynamic change of processing styles across emotions, and top down cognitive controls, convey an advantage not only in natural and biological environments but also in social and psychological ones.