One of the most heavily researched areas of bias in the emotional disorders has been the area of mnemonic biases in depression. For example, an influential study by Lloyd and Lishman (1975) reported that depressed patients were faster to recall unpleasant than pleasant memories in response to neutral cue words. In a cleverly designed application of this procedure, Clark and Teasdale (1982) tested a group of depressed patients who showed considerable diurnal variation in their mood, and found that more unhappy memories were retrieved when participants were at their most depressed, but a greater proportion of positive memories were recalled when depression was at its lowest. Williams (e.g., 1992; Williams et al., 2000; Williams et al., 2007) has further shown that not only do the recollections of depressed and suicidal patients differ in terms of valence from the recollections of normal controls, but they also tend to be general rather than specific memories—for example, recollections along the lines of "I always used to be unhappy" rather than "I was unhappy on that particular occasion". A detailed analysis of the production of these overgeneral autobiographical memories in depression by Dalgleish and colleagues (Dalgleish et al., 2007) has identified a crucial role of restrictions in working memory capacity that result from depression. This working memory capacity reduction leads to the production of overgeneral memories rather than specific memories, which typically require additional retrieval effort for their recollection.
In addition to studies of variations in autobiographical memory, there have also been a considerable number of studies that have investigated verbal learning. Derry and Kuiper (1981) found that from a list of words depressed participants recalled more self-related negative words than positive words, whereas normal controls recalled more self-related positive words. Bradley and Mathews (1983) found that this effect was limited to negative words that were endorsed for the self, but that their depressed participants recalled more positive words that they had endorsed as descriptive of other people. This study therefore demonstrated that the negative recall bias is not a general response bias, but rather may be a feature of particularly salient information such as that related to the self. A subsequent study showed that the negative bias was no longer present once the depressed individuals had recovered (Bradley & Mathews, 1988). Dobson and Shaw (1987) reported similar findings in which currently depressed individuals showed a negative recall bias on the self-reference encoding task, but this effect was no longer present following recovery. These negative biases have been found not only with verbal material but also with memory for facial expressions (Ridout et al., 2003).
These and related findings suggest that there is a significant mood congruency effect in depression, in that material that is congruent with the depressed mood is better learned and better recalled, but there is less evidence for state-dependent learning which would require better recall of material, irrespective of its valence, that was learned when depressed. The effects are stronger for recall than recognition (see Dalgleish & Watts, 1990), which points to the fact that mood has an important influence on retrieval. However, the effects are in part a decrease in the positive biases shown in normal individuals who have been found to forget more negative than positive self-descriptive adjectives in a directed forgetting task (Power, Dalgleish, Claudio, Tata, & Kentish, 2000). There is also evidence that recall of to-be-learned material is improved if the structure is inherent in the material rather than if it has to be provided by the participant (Dalgleish & Watts, 1990). Finally, most of the research on mnemonic biases has focused on so-called explicit or strategic memory tasks. Indeed, theorists such as Williams et al. (1997) have proposed that the memory biases evident in depression are restricted to explicit or strategic memory.
However, there has been recent interest in the question of whether implicit memory biases are evident in the emotional disorders, and Tobias, Kihlstrom, and Schachter (1992) have summarised preliminary data showing that under carefully designed conditions induced sad mood may also lead to retrieval biases. The fact that the data on implicit memory biases in depression are contradictory, with some studies reporting no implicit memory effects (e.g., Denny & Hunt, 1992; Hertel & Hardin, 1990), but other studies obtaining such biases (e.g., Elliott & Greene, 1992; Ruiz-Caballero & Gonzalez, 1994), emphasises that the effects are subtle and may be easily swamped by other retrieval factors (Tobias et al., 1992). A careful study by Watkins, Martin, and Stern (2000) in which four types of implicit memory bias were investigated demonstrated that implict perceptual biases were not obtained in depression, but one type of implicit conceptual bias (a definition-based word retrieval task) was obtained. This study therefore highlights the importance of semantic biases in depression, which can be found at automatic levels especially, though not only for self-related material. However, the effects can often be subtle.
The work on mood-incongruent memory and mood regulation provides some support for our suggestion that the basic emotions modules may be more difficult to alter in individuals prone to emotional disorders. For example, studies of mood-incongruent memory show that the induction of a sad or depressed mood in normal individuals leads to the retrieval of mood-incongruent positive memories as part of mood regulation (e.g., Parrot & Sabini, 1990). Wegner (1994; Wenzlaff & Wegner, 2000) has proposed that this effortful control over mental states will be more successful where more working memory capacity is available, but where there is a reduction in such capacity, control is more likely to backfire and lead to the opposite; thus, the more that depressed individuals attempt to rid themselves of the depressed mood, the more the depressed mood predominates, because of the loading of working memory with preoccupations about the lost role or goal and its implications for the self. In addition, however, we are suggesting that the modular or near-modular organisation of the basic emotions of sadness and disgust in depression leads to their autonomous operation. That is, once a state of sadness-disgust has been entered, the fact that in depressed individuals there is no integration with the positive aspects of the self means that the emotional state is difficult to alter or regulate; thus, the mood-regulation strategies in which normal individuals can retrieve positive memories of the self or focus on other valued goals or roles are less available in a state of depression.
Was this article helpful?