جانبداری حافظه در هراس اجتماعی تعمیم یافته: خاطره عبارات عاطفی منفی

In two experiments, the authors examined memory for facial emotional expressions in patients with generalized social phobia (GSP) and in nonanxious control (NAC) participants. Three main questions were addressed. First, do patients with GSP differ from NAC participants in their overall memory for facial expressions? Second, do patients with GSP exhibit a memory bias for negative versus nonnegative expressions? Third, if such a bias exists, is it specific to angry expressions? The results of both experiments indicated that patients with GSP have better memory for all facial expressions than do NAC participants. Results of experiment 2 suggest that patients with GSP exhibit enhanced recognition for negative compared with nonnegative expressions; in contrast, NAC participants did not exhibit such enhancement. Results concerning specificity were equivocal. The importance of examining cognitive biases in patients with GSP via the use of facial expression is discussed.

Socially anxious individuals show attention, interpretation, and evaluation biases to socially relevant information (e.g., Amir, Foa, & Coles 1988, Butler & Mathews 1983, Butler & Mathews 1987, Foa, Franklin, Perry, & Herbert 1996, Gilboa-Schechtman, Franklin, & Foa 2000, Hope, Rapee, Heimberg, & Dombeck 1990, Lucock & Salkovskis 1988 and Mattia, Heimberg, & Hope 1993). The findings concerning memory biases in social anxiety are less conclusive. Although some studies have found enhanced memory for negative or threat-relevant information, others did not. For example, Claeys (1989) reported better recall for negative self-descriptive adjectives in a depth-of-processing task in highly socially anxious individuals compared with individuals with low social anxiety. In contrast, in a series of four studies using a variety of memory tasks and stimulus types, Rapee, McCallum, Melville, Ravenscroft, and Rodney (1993) failed to identify memory biases for negatively valenced stimuli in individuals with social phobia (SP). With the exception of one study (Lundh & Ost, 1996), investigations of memory bias in persons with SP used verbal stimuli. Recently, researchers have noted the advantage of using more ecologically valid stimuli in studying cognitive biases in persons with SP (e.g., Clark and Wells 1995, Lundh and Ost 1996 and Bradley, Mogg, Millar, Bonham-Carter, Fergusson, Jenkins, & Parr 1997). Facial expressions of emotions seem particularly well suited for this purpose. First, these stimuli are ubiquitous, biologically significant, and their meaning is context free Hansen & Hansen 1994, Ekman & Friesen 1975 and Bradley, Mogg, Millar, Bonham-Carter, Fergusson, Jenkins, & Parr 1997. Second, facial expressions of emotions constitute the actual cues to which persons with SP seem to be particularly attuned, whereas words constitute only indirect representations of these cues. Third, social evaluations are often expressed only nonverbally. Because persons with SP tend to be excessively preoccupied with such evaluations, biases in the information processing of facial expressions may play an important role in the maintenance of SP. To assess recognition memory for facial expressions, Lundh and Ost (1996) presented 20 photographs to persons with SP and to control participants, asking them to state whether the person in the photograph was “generally critical towards others … or generally accepting and tolerant … ” (p. 789). After a distracter task, participants were presented with 20 photographs of individuals encountered in the initial task and 60 distracter photographs. Participants were asked to identify the faces they had seen in the initial task. Persons with SP recognized more faces they had rated as “critical” than faces they had rated as “accepting,” whereas control participants exhibited the opposite pattern. Although these results are interesting, the design of the study made it impossible to determine whether a response bias or a memory bias underlies the preference of those with SP for critical faces. If the results were the result of a response bias, we would expect persons with SP to designate critical faces as familiar regardless of whether they had seen them before. To explore this alternative explanation, a comparison of percent recognition of previously presented faces with percent recognition of new (distracter) faces should have been conducted. However, because the participants in Lundh and Ost's study did not rate the critical versus accepting nature of the distracter faces presented during the recognition task, such analysis could not be conducted. Thus the results of the study do not constitute conclusive evidence for a memory bias. The purpose of the present studies was to examine memory for emotional expressions in individuals with generalized social phobia (GSP) and in nonanxious control (NAC) participants. Three main questions were addressed. First, do patients with GSP differ from NAC participants in their overall memory for facial expressions? Clark and Wells (1995) proposed that in social contexts, patients with GSP tend to focus on and monitor their own feelings, sensations, and other aspects of self-presentation, thus allocating fewer resources to the processing of others' reactions. Accordingly, patients with GSP would be expected to exhibit inferior memory for facial expressions. Clinical observations, however, suggest that patients with GSP tend to ruminate about social interactions, particularly about others' emotional reactions to their performance. These observations have led us to hypothesize that the memory of patients with GSP for facial expressions would be superior to that of NAC participants. Second, do patients with GSP exhibit a memory bias for negative emotional expressions? Several investigators have suggested that anxious individuals tend to avoid elaboration on threat-relevant information (e.g., Williams, Watts, MacLeod, & Mathews, 1988). Accordingly, patients with GSP would be expected to have poorer memory for negative facial expressions than for positive or neutral expressions. Again, clinical observations seem to contradict this view: patients with GSP appear to be preoccupied with negative social interactions rather than with positive ones. These observations as well as the findings by Lundh and Ost (1996) suggest that patients with GSP are particularly sensitive to negative interactions, and are therefore expected to have an enhanced memory for negative facial expressions. In the present study, we examined the hypothesis that patients with GSP would exhibit enhanced memory for negative (e.g., anger and disgust) compared with nonnegative (e.g., happy and neutral) emotional expressions, whereas NAC participants would not exhibit such a bias. If patients with GSP have a bias for remembering negative expressions, is this bias specific to anger, or does it extend to other negative expressions as well? Social rejection and disapproval are expressed in multiple ways, the most obvious ones being anger, contempt, and disgust. Thus patients with GSP may have an enhanced sensitivity to all expressions connoting disapproval. However, an evolutionary perspective would favor a bias specific to anger. Indeed, anger is more directly related to harm from a dominant member of the group or to expulsion from the group than is disgust (e.g., Trower & Gilbert, 1989). We therefore hypothesized that patients with GSP would exhibit a stronger memory bias to anger than to disgust expressions. To examine these questions, we conducted two experiments using two groups, patients with GSP and NAC participants. Experiment 1 was designed to test the first two hypotheses (the enhanced memory hypothesis and the negative bias hypothesis) within a memory task for proper names. In experiment 2, we tested all three hypotheses using an intentional memory task for facial expressions.