A number of paradigms have been employed to test for possible biases in attentional processes in both anxiety and depression. The most widely applied task has been the emotional Stroop (see Chapter 6 for a detailed discussion of this task) in which disorder-relevant words are tested for their degree of interference with colour naming the ink in which such words are presented. Gotlib and McCann (1984) reported Stroop interference effects that distinguished dysphoric students from controls. Gotlib and Cane (1987) repeated this study with clinically depressed individuals and again found an interference effect with the naming of negative content words. However, when the task was repeated on recovery no effect was found. These studies can be criticised for not partialling out the effects of anxiety levels which are often higher in depression and, as Mogg et al. (1991) found, may well be the source of the effects reported in Gotlib's studies.
Thus, MacLeod, Mathews, and Tata (1986) used their dot probe task with groups of both depressed and anxious individuals (see also Chapter 6). In this task, the individual is presented with a pair of words one of which may be threatening; on the critical trials the threat word is replaced with a dot to which the participant must respond as quickly as possible, the reaction time being used as an indicator of the focus of attention. MacLeod and colleagues found an attentional bias towards threatening words in anxious participants, but no such bias in depressed participants. Because this study is one of the few in which both clinically depressed and clinically anxious participants have been compared on the same task, the results have had more impact on views about attentional bias than is warranted. Major problems, which we have also noted in Chapter 6, that limit the interpretation of the study are: that the delay of over 500 ms between word presentation and the dot probe could allow two or more changes of attentional focus; that the key words themselves were threat-related rather than depression-related; and that the overall reaction times for the depressed group were considerably slower than those for the anxious patients. Mogg, Bradley, and Williams (1995) found attentional bias for negative words at 1000 ms, but none at shorter periods with masked stimuli, again suggesting that biases may only occur in controlled but not automatic processes in depression, and that such biases may lead to depressed individuals having problems with disengaging from negative stimuli once these have been attended to (Compton, 2000).
In a variant on the dot probe task Gotlib, MacLachlan, and Katz (1988) found with more appropriate negative and positive words that normal participants attended more to positive words than to negative words, whereas mildy depressed participants attended equally to both. Gotlib and his colleagues interpret these findings as evidence that normal individuals show a positive bias, but no bias is shown by mildly depressed individuals who are therefore "realistic". Of course this interpretation runs headlong into the problems faced by the "depressive realists" which we have mentioned above and which we will cover in detail in the later section on judgement and reasoning. Overall, though, the findings emphasise that attentional biases can be found in depression with carefully chosen depression-related content, especially at longer display times. These findings concur with results obtained from the emotional priming paradigm with clinically depressed patients (Power et al., 1996) in which depression-related primes were found to increase the speed of endorsement of negative adjectives by depressed patients. Such results emphasise the need for careful selection of appropriate materials, and suggest, for example, why Clark et al. (1983) failed to find effects of depressed mood on lexical decision times for negative words.
Other tasks that have been used to assess attentional biases include the dichotic-listening task and an eye fixation task. For example, McCabe and Gotlib
(1993) reported that depressed participants were distracted by depression-relevant words occurring on the non-attended channel in a dichotic-listening task, although following recovery this effect no longer occurred. The recovered participants were, however, more distracted by positive words on recovery, although the effect just failed to reach significance. Interestingly, the depressed participants in the Power et al. (1996) priming study were found to show positive priming effects for positive adjectives. The possibility therefore that recovered depressives may be more positive on some tasks needs further exploration. Preliminary results with an eye fixation task reported by Matthews and Antes (1992) showed that dysphoric students made more eye fixations towards sad visual material than did normal students, although Lough-land, Williams, and Gordon (2002) found poorer overall scanning of facial features compared to controls.
Was this article helpful?