5.2. Avoidant processing mode and tasks used in normal populations

Wells and Matthews (1994) suggested an explicit dissociation of general bias from anxiety-specific bias in processing threatening stimuli. To provide evidence for a general bias, they invoked the “perceptual defense” effect (e.g. Kitayama, 1990), where participants demonstrate raised perceptual thresholds towards threatening words, suggesting an automatic filtering of the mildly threatening stimuli (Dixon, 1981). According to Kitayama (1990), affective information can provoke narrowed attention with increased efficiency if the perceptual code for the stimulus is strong, but reduced efficiency if it is weak, such that other activated codes may capture attention. He suggested that the perceptual defense effect depends on participant expectations that influence the strength of pre-attentively processed perceptual codes. Kitayama (1991) found that, while normal participants find it more difficult to perceive visually degraded affective words, stimulus degradation enhances attention to the identification of threatening stimuli in anxious participants (e.g., MacLeod & Rutherford, 1992). Nevertheless, anxiety may be associated with a weakening or reversal of perceptual defense (Mathews & MacLeod, 1985; see also Williams et al., 1997, p. 82).

The dot probe task elicits an interesting processing style, one exhibited by normal participants relative to negative stimuli. This task asks participants to respond as quickly as possible to a probe that appears in one of two locations, occupied by two stimuli (one threatening and the other neutral), usually presented for 500 ms. Using emotional word stimuli, MacLeod, Mathews and Tata (1986) showed that anxious patients and control participants employed very different patterns of attention; anxious individuals responded more quickly to probes that replaced threatening stimuli, as compared to neutral stimuli, and normal individuals were faster to respond to probes that replaced neutral stimuli, as compared to threatening stimuli. This suggests that, while anxious individuals orient attention towards the location at which a threat has occurred, normal participants orient their attention away from those locations. Similar results have been reported by Bradley, Mogg, White, Groom and Bono (1995), Bradley et al. (1997), Bradley, Mogg and Miller (2000), and Yiend and Mathews (2001), using the human threatening face or threatening pictures as stimuli. Bradley et al. (1997) found a significant attentional bias away from threatening faces in non-dysphoric participants, and Yiend and Mathews (2001) discovered a significant interaction between threatening stimulus position and target position for the low-anxious group but not for a high-anxious group. These studies indicate not only that normal individuals exhibit a different pattern of attention to that found in dysphorics, with regard to threatening information, but also that they use a particular pattern of attending to threatening information relative to neutral or positive information.

Research using the dot probe task provides important information about general attention, but provides only a snapshot of events that occur some 500 ms after the onset of the word stimulus pair (Hermans, Vansteenwegen & Eelen, 1999, see also Eysenck, 1992). It is unclear whether the dot probe task effect implicates attentional avoidance or the facility to disengage attentional resources from threatening information in normal participants (see Fox, Russo, Bowles & Dutton, 2001). Finally, the effect implicates the initial orientation of attention and its costs and benefits relative to target processing but does not address the issue of attentional resources that impact the overall speed of target processing.

The effects of the emotional Stroop task reveal another aspect of attention. In this task, which constitutes the method most frequently used to investigate attentional bias to emotional stimuli, participants are asked to name the colour of a word that varies in emotional valence (negative, neutral, and positive). In general, words with a threatening or negative meaning cause slower colour-naming latencies relative to matched neutral words in emotionally disturbed populations (Williams, Mathews & MacLeod, 1996).

Mathews and MacLeod (2002) noted that attentional biases in dysphoric populations were observed only when at least two competing processing options were present. The effect of the emotional Stroop task is based on selectivity and competition between two such processing operations: one for emotionally irrelevant information and the other related to the demands of the task. If an irrelevant stimulus can be processed strongly and capture attention, and thus compete with target processing, it will impede target processing. Indeed, the vigilant attentional processing mode in response to threatening stimuli, typical of dysphoric individuals, could explain longer colour-naming latencies in the emotional Stroop task (Mathews & MacLeod, 2002; Williams et al., 1996, 1997).

The absence of this effect in normal populations does not readily suggest any particular attentional pattern. The absence of an emotional Stroop effect could simply imply that normal participants exhibit no specific response pattern to negative stimuli relative to neutral and positive stimuli. Alternatively, the absence of an effect could imply strong control in task demand processing; that is, although there is a specific pattern of attention to negative information in normal participants, strong perceptual input or compelling target-related processing can intercept the effects of emotionally irrelevant processing, regardless of valence. According to this hypothesis, attention to negative information expressed by normal individuals will not be observed when directly competing processing options are present. Finally, the emotional Stroop task does not seem suitable to demonstrate attention to emotional stimuli in normal individuals.

There is much support for the notion that normal and dysphoric populations exhibit distinctively different attentional response patterns to emotionally negative stimuli. The attentional pattern of normal participants differs from that of dysphoric individuals with respect to specific characteristics, such as avoidant processing, in which attention is allocated more to target-related processing by inhibiting the processing of irrelevant threatening stimuli. While there is much more evidence of a relatively clear attentional pattern to threatening stimuli in dysphoric participants, the comparable attentional pattern in normal participants has not been directly investigated.