القاء تعصب خوش خیم تفسیری اضطراب خصلتی را کاهش می دهد

If negative interpretational bias causes emotional vulnerability, reduction of this bias should reduce proneness to anxiety. High trait-anxious volunteers were trained over four sessions to resolve descriptions of ambiguous events in an increasingly positive manner. This group subsequently made more positive interpretations of novel descriptions than did those in a test–retest control condition. Furthermore, trait anxiety scores reduced more in the trained group than in untrained controls. These results confirm earlier findings that modifying interpretation biases produces congruent changes in emotional vulnerability, and suggest a possible role for similar training methods in controlling pathological anxiety.

It is now well established that those prone to negative emotional states are less likely than are other groups to interpret ambiguous events in a relatively positive manner (Eysenck, Mogg, May, Richards, & Mathews, 1991; Lawson, MacLeod, & Hammond, 2002). In the study reported by Eysenck et al. (1991), anxious patients and nonanxious controls first listened to sentences, some of which were ambiguous and could be interpreted in a more or less threatening manner (e.g., the doctor examined little Emma's growth). The way in which the critical sentences were understood was assessed by presenting a series of similar items in which the ambiguous meaning had been resolved in either a threatening or benign direction. In a later recognition test (e.g., by referring to Emma's tumor or her height), nonanxious participants were less likely to endorse threatening than benign meanings as matching the original sentence, whereas anxious patients endorsed the threatening interpretations as often as they did benign meanings. Such findings suggest—but do not prove—that a (relatively) negative interpretational style could be a contributory cause of vulnerability to anxiety. However, it could be that the apparent associations arise because a preexisting state (or vulnerability) leads to less positive interpretations being made, rather than the other way around. Furthermore, both vulnerability and interpretation style could be independently produced by other processes, so that there might be no direct causal link between mood and interpretation. More convincing evidence of a causal link between the two has been provided by recent studies in which interpretational biases were experimentally manipulated (Grey & Mathews, 2000; Mathews & Mackintosh, 2000; Yiend, Mackintosh, & Mathews, 2005; Wilson, MacLeod, Mathews, & Rutherford, 2006). In these studies, nonanxious volunteers were randomly allocated to conditions in which they made either negative or benign interpretations of ambiguous text. For example, Mathews and Mackintosh (2000) presented nonanxious volunteers with short texts describing ambiguously threatening social situations, with the emotional outcome being resolved only by the final word, which was presented in fragment form. Participants were required to complete this fragment, and then to answer a question designed to reinforce the designated emotional meaning. Those allocated to a condition in which outcomes were nearly always negative, were subsequently more likely to interpret new ambiguous descriptions in a similarly negative fashion than were those previously exposed to more benign outcomes. Importantly, anxious mood also changed congruently, but only in those who were trained under conditions requiring the active generation of meaning: that is, those who had to complete the resolving fragment and question. In other conditions, participants exposed to the same information—but who did not have to generate it for themselves—developed the same interpretative bias for new descriptions, but did not change in mood. Thus, although active generation of emotional meanings during training can alter mood, such mood changes are not a necessary condition for the induction of interpretative bias. Some forms of training produced interpretation biases even in the absence of mood change. To further illustrate this point, consider the results reported by Wilson et al. (2006). The training used in this study was based on that developed by Grey and Mathews (2000), in which volunteers were presented with homograph primes having both threatening and benign meanings, followed by a word fragment to complete corresponding to one of these meanings (e.g., the homograph “sink” followed by fragments corresponding to either “drown” or “basin”). No mood change was observed during training, even after prolonged practice with either threatening or benign completions. Despite this, tests with new homographs revealed that a differential interpretative bias had indeed been induced. Participants then viewed a series of videos of real-life accidents and the group assigned to prior practice with threat completions reported greater increases in anxiety than did a comparison group that had practiced benign completions. Thus an interpretative bias can be induced experimentally without necessarily changing mood at the time, but changes in emotional vulnerability can be observed if the induced bias influences how potentially emotional events are processed. Although these findings provide evidence that induced negative interpretative style can increase vulnerability to anxiety, they do not demonstrate that pre-existing vulnerability to anxiety can be reduced by similar methods. The designs used have involved the comparison of groups assigned to either threatening or benign training, so it remains unclear whether differences were mainly due to negative changes occurring in those trained in the threat condition, rather than to positive effects due to benign training. Furthermore, participants with elevated negative emotionality scores were generally excluded from these studies for ethical (and other) reasons, so little can be concluded about changes in such vulnerable groups. Encouraging results with high trait anxious individuals have been obtained, however, in parallel work carried out using a method developed by McLeod and colleagues to train attention towards or away from threat cues (MacLeod, Rutherford, Campbell, Ebsworthy, & Holker, 2002. Student participants saw two words displayed simultaneously, one threatening and one benign, followed by a to-be-responded-to target in the location of one of the words. In one group the target was nearly always in the location of the threatening word (attend-threat training), while in the other group the target was in the benign location (avoid-threat training). Once again no differential emotional changes were observed during training, but differences did emerge when a post-training stressful task was performed, with more anxiety elevation in the attend-threat group. More critically for current purposes, subsequent studies involving highly anxious students showed that repeated sessions of avoid-threat training led to greater reductions in anxiety than did a neutral control condition (for more details see Mathews & MacLeod, 2002). Although we (Mathews & Mackintosh, 1998) had proposed that both biased attention to threat and biased interpretation favoring threatening meanings could arise through similar underlying processes, in preliminary unpublished studies we have found little evidence that the two types of bias are correlated within nonclinical populations. Perhaps both types of bias can independently enhance emotional vulnerability but without having reciprocal influences on each other. In any event, at this stage we would not want to predict that induction of one type of bias would necessarily produce consistent effects on others. None the less, if reduction of a bias to attend towards threat in anxiety-prone individuals reduces vulnerability, then the assumption that interpretative bias also has a causal role in anxiety suggests that training designed to induce more positive interpretations should also have beneficial effects. In order to test this possibility, we re-designed the training material originally developed by Mathews and Mackintosh (2000). This original material consisted of a series of about 100 descriptions of ambiguous social events that participants were to read and imagine, and which could be resolved in either a benign or negative manner. In pilot work we redesigned this material so that the ambiguous events previously resolved in a negative or benign manner were now concluded more positively. Additional training items involving ambiguous physical threats were added to the prior social descriptions so as to expand the total number to 200, allowing 50 to be used in each of four sessions. Finally, a matched set of supposedly neutral control descriptions were developed, from which we attempted to remove explicitly emotional wording. In a pilot study of high anxious volunteers allocated at random to four sessions of positive or neutral training, we found some differences in final anxiety scores favoring the positive group. However, inspection of actual mean scores revealed that the difference was almost entirely due to paradoxical increases in anxiety in the control group, with little or no improvement in the positive group. In retrospect, we concluded that removal of explicit emotional references was not sufficient to prevent threatening interpretations about inherently ambiguous situations, particularly in highly anxious populations. To clarify this point, in order to match the ambiguous training content, control descriptions included events such as meeting new acquaintances, or walking in town alone. Including this content may have allowed spontaneous threatening interpretations, but (unlike positive training) the ambiguity cannot be resolved positively (to avoid training effects), nor can it be resolved negatively (for several reasons including ethical considerations). On the other hand, substituting completely different neutral content would not control for the effects of repeated exposure to similar material. Seeing no obvious solution to this dilemma, in the present study we decided to first compare a revised positive training condition with a test–retest control group who did not receive any training, which at least avoided the risk of inadvertently increasing anxiety. Of course, the corresponding disadvantage is that any positive results could be attributed to exposure to materials and the test situation in the trained group, although our experience of unexpected negative consequences following such exposure suggests this is unlikely. In thinking about why positive training effects were not more apparent, we noted that some participants reported finding it difficult to identify themselves with the very positive event outcomes described in training, and a few even reported being irritated by having to endorse such (for them) unrealistic interpretations. This suggested the possibility that it would be more effective to introduce positive outcomes in a more graded fashion, beginning as nonnegative and gradually becoming explicitly positive.