افکار سرکوب، تعصب حافظه برای مواد تهدیدی را بهبود می بخشد

The current study examined the impact of thought suppression on indices of anxiety, including memory indices (implicit and explicit memory biases) and physiological indices (heart rate). The participants, 81 undergraduates scoring in the top quartile of a self-report measure of trait anxiety, were randomly assigned to one of three experimental groups: thought suppression (TS), thought concentration (TC), and thought wandering (TW). The TC and TW groups were included to control for the effects of effortful processing and exposure to stimuli, respectively. One block of threat words and one block of neutral words were presented under conditions of cognitive load, and participants’ physiological responses and memory biases were measured. The thought suppression group exhibited an enhanced overall memory bias for threat words, driven by an elevated explicit memory bias, relative to the other two groups, a result that has implications for ironic processes theory and may inform information-processing models of anxiety.

In daily life, virtually all people experience unwanted thoughts (Wegner, 1994). For example, in unselected samples, cognitive intrusions are generally reported by 99% of respondents (Freeston, Ladouceur, Thibodeau, & Gagnon, 1991; Purdon & Clark, 1993). Although the occurrence of such thoughts is commonplace, the various anxiety disorders are characterized by recurrent and persistent intrusive thoughts related to worry, anxiety, and fear (American Psychiatric Association, 2000), and thereby referred to as “threat-related”. Individuals often try to suppress unwanted, aversive, and distressing thoughts (e.g., threat-related thoughts; Hayes, Wilson, Gifford, Follette, & Strosahl, 1996; Wegner & Zanakos, 1994). Thought suppression entails effortful cognitive attempts to not think about or avoid a given thought (Wegner, 1994). However, attempts to use effortful cognitive strategies, such as thought suppression, to manage intrusive thoughts are associated with increased difficulty in eliminating the thoughts as compared with non-effortful responses (Freeston et al., 1991). Ironic processes theory (Wegner, 1994) These observations are consistent with an extensive body of research relating to Wegner's ironic processes theory (Wegner, 1994). The theory postulates that thought suppression can generate hyperaccessibility, or enhanced cognitive accessibility of to-be-suppressed thoughts. After a period of attempted suppression, a rebound effect has been observed that takes the form of the suppressed thought being more accessible than it would have been without such suppression (Wenzlaff & Wegner, 2000). The theory was first supported by investigations conducted with unselected college samples, demonstrating that suppression of neutral thoughts leads to elevated subsequent thought rebound using think-aloud paradigms (e.g., Wegner, Schneider, Carter, & White, 1987). Findings also showed that hyperaccessibility is pronounced under conditions of cognitive load, in which participants suppress while performing an additional cognitive task (e.g., rehearsal of a nine-digit number; Wegner & Erber, 1992). Thought suppression is hypothesized by Wegner to involve dual processes: an effortful operating process that searches for distracter thoughts to promote a desired state, and a non-effortful monitoring process that remains vigilant for unwanted thoughts in order to signal success or failure in suppression and to sustain a preferred state. Cognitive load is believed to enhance ironic effects by creating competing cognitive demands that impair effortful attempts at suppression, while leaving the monitoring process free to search for unwanted thoughts that may then intrude upon conscious awareness (reviewed in Wenzlaff & Wegner, 2000). Ironic processes theory and anxiety disorders Several investigations have evaluated ironic processes theory within the context of anxiety disorders. A correlational study by Wegner and colleagues showed an association between the tendency to suppress obsessional thoughts and self-reported occurrence of such thoughts in a sub-clinical college sample (Wegner & Zanakos, 1994). Other studies have examined suppression of threat-related thoughts characteristic of anxiety disorders, generally using self-reported thought occurrence paradigms. For example, Fehm and Margraf (2002) compared groups of participants with agoraphobia, social phobia, or no anxiety disorder, with instructions to suppress three statements: one relevant to fears in agoraphobia, one to fears in social phobia, and one to financial problems. Participants attempted to suppress each topic for 5 min while pressing a button each time the topic came to mind. The agoraphobic group showed a heightened frequency of thoughts related to the agoraphobia topic during attempted suppression, whereas participants with social phobia showed elevated frequency of thoughts across all topics. Additional studies have employed instructed suppression and thought-monitoring methodology using target thoughts and samples representative of anxiety disorders, such as topics of worry in generalized anxiety disorder (e.g., Becker, Rinck, Roth, & Margraf, 1998; Mathews & Milroy, 1994), traumatic intrusions in posttraumatic stress disorder (e.g., Ehlers & Steil, 1995), feared stimuli in phobias (e.g., Wenzel, Barth, & Holt, 2003; Zeitlin, Netten, & Hodder, 1995), and obsessions in obsessive–compulsive disorder (e.g., Salkovskis & Campbell, 1994; reviewed in Purdon & Clark, 2000). Overall, the findings have been relatively inconsistent across the anxiety disorders, with more support for elevated thought frequency under conditions of suppression in posttraumatic stress disorder and obsessive–compulsive disorder than in specific phobias and generalized anxiety disorder (reviewed in Purdon, 1999). Differing methodologies may also contribute to the discrepant findings. Certain measures of thought frequency, for example, may be subject to self-reporting biases. More specifically, participants with anxiety disorders or high in trait anxiety may avoid reporting intrusive thoughts on think-aloud or thought recall tasks as such tasks require reproduction of potentially anxiety-provoking stimuli (Gosselin et al., 2007; Hayes et al., 1996). Trait anxiety, as indicated by self-report in the present study, refers to a general propensity to respond fearfully to aversive or threat stimuli (Spielberger, 1985). In the present study, both an implicit memory measure and a recognition memory measure were used, such that participants were never required to reproduce the words presented during the experiment. As previously stated, thought suppression is posited to enhance subsequent cognitive accessibility of to-be-suppressed thoughts. After a period of attempted suppression, a rebound effect has been observed whereby the suppressed thought appears more accessible than it would have been without such suppression (Wenzlaff & Wegner, 2000). For example, Harvey and Bryant (1999) presented high- and low-trait anxious participants with either a neutral or distressing film, and then instructed participants to either suppress thoughts of the film or think about anything for two 3 min periods, pressing a response button each time they thought about the film. Harvey and Bryant obtained a delayed rebound effect for frequency of film-related thoughts across films during the second 3-min period, mediated by trait anxiety. The evidence for thought rebound effects raises the possibility that suppression may lead to measurable memory biases for previously suppressed thoughts. Memory biases and anxiety disorders Memory bias has been operationalized as increased recall of threat-related material relative to non-threat material, and biases may be explicit and consciously retrieved or implicit and tested indirectly (see Coles & Heimberg, 2002, for a review). Memory biases for threat material have generally been studied in anxious populations, as threat-related processing is highly relevant to such groups; however, memory biases for threat have only been obtained in a minority of such studies. As reviewed by Coles and Heimberg (2002), some prior research has obtained evidence of an explicit memory bias in panic disorder, posttraumatic stress disorder, and obsessive–compulsive disorder, and of an implicit memory bias in each of the anxiety disorders, but other studies have found no such effects. The majority of research has tested memory biases following concentration on threat-related words. Interestingly, in the few studies of memory biases following directed forgetting, an instructional condition similar to thought suppression in which participants are told to either forget or remember words just presented, researchers have found enhanced memory for threat-related words that were instructed to be forgotten (McNally, Metzger, Lasko, Clancy, & Pitman, 1998; Wilhelm, McNally, Baer, & Florin, 1996). However, as yet, memory biases for threat words have not been investigated following instructed thought suppression as operationalized by Wegner and colleagues. Bjork and Bjork (2003) suggested that implicit biases for threat-related thoughts may be particularly susceptible to ironic effects following attempted thought suppression, as suppressed material may remain activated in memory while conscious access to the material may be impaired, although clearly there is a need to evaluate both implicit and explicit memory biases. Goals of the present research In addition to exploring memory effects, the present study also explored potential physiological effects of threat-related thought suppression. Several studies have investigated the impact of suppression of negative thoughts on emotion, most using self-report data. For example, Purdon (2001) and Markowitz and Borton (2002) found that instructed suppression of obsessions and negative self-referent thoughts, respectively, led to increases in self-reported anxiety and frustration. Spontaneous suppression of personally relevant intrusive thoughts appears to have a similar impact on emotional distress and discomfort (Freeston et al., 1991; Marcks & Woods, 2005; Salkovskis, Westbrook, Davis, Jeavons, & Gledhill, 1997). However, only a few studies have examined the effects of thought suppression on physiological indices of emotion. Wegner and Zanakos (1994) demonstrated that individuals who chronically suppress certain thoughts tend to respond to reminders of those thoughts with elevated physiological arousal. Likewise, attempts to suppress expressions of negative emotion elevate physiological activation, measured via cardiovascular activation, skin conductance level, and somatic activity (Gross & Levenson, 1993; Gross & Levenson, 1997). These results provide some support for the predictions of ironic processes theory and align with aforementioned findings regarding self-reported emotion. As yet, however, the effects of suppressing threat-related thoughts on physiological indices of anxiety, such as heart rate, have not been examined directly. As self-report measures may be influenced by factors such as response biases, introspective limits (Nisbett & Wilson, 1977), or demand characteristics (Orne, 1962), physiological data may be particularly useful as a more objective measure as well as a component of anxious processes (Rosen & Schulkin, 1998). The present study investigated the ways in which threat-related thought suppression may impact memory and physiological indices relevant to anxiety. Instructed thought suppression methodology was adapted from prior research conducted by Wegner and colleagues’ (i.e., “try to not think about the word”; reviewed in Wenzlaff & Wegner, 2000) and involved exposure to threat words, stimuli that have been used extensively in previous research (e.g., Mogg et al., 2000; Mogg, Bradley, Millar, & White, 1995). Participants high in trait anxiety, as assessed by a self-report measure, served as a sample. Physiological indices and memory biases were assessed during and following the instruction to suppress, concentrate on, or just notice the words. Building on ironic processes theory and prior information-processing research, it was first hypothesized that an implicit memory bias and explicit recognition memory bias would be most elevated in the thought suppression group compared with the other two groups. Second, it was hypothesized that physiological arousal as assessed via heart rate would be most elevated in the thought suppression group.

All analyses were conducted in SPSS Version 14.0. An α level of 0.05 was used for all statistical tests. Statistical power was estimated using standard formulas, and was in the range of 67% for the predicted memory bias results but only 10% for the predicted heart rate result. Preexisting and baseline differences A χ2 test revealed no group differences in proportion of females, χ2 (2)=0.033, ns, proportion of each ethnicity, χ2 (6)=0.664, ns, mean age, F (2, 78)=0.539, ns, or mean income, F (2, 78)=1.738, ns. Likewise, one-way ANOVAs revealed no significant differences in BIS subscale scores across groups at the time of recruitment, F (2, 78)=1.970, ns, or at the time of participating in the study, F (2, 78)=1.497, ns. One-way ANOVAs revealed no significant differences across groups in HRT, F (2, 77)=0.754, ns, or RSA, F (2, 77)=0.257, ns, during the baseline recording period prior to the start of the experimental manipulation. One-way ANOVAs revealed no significant differences across groups for PANAS-positive emotion, F (2, 78)=2.213, ns, or PANAS-negative emotion, F (2, 78)=0.513, ns, prior to the experimental manipulation, and no significant differences in WBSI scores across groups, F (2, 78)=2.495, ns. Since no group differences were observed for these variables, they were not used as covariates in subsequent analyses.