تنظیم احساسات و پردازش اطلاعات عاطفی: اثر تعدیلی هشیاری عاطفی

Abstract The aim of this study was to examine the moderating role of emotional awareness in the relationship between emotion regulation strategies and emotional information processing. A total of 120 female students regulated emotions while watching an unpleasant film. Before and after emotion induction, participants completed a set of tasks that required matching facial expressions. The results demonstrated that participants who were high in emotional awareness showed a significantly smaller increase in error responses (i.e., incorrect matches) than participants who were low in emotional awareness. However, this effect was observed only in suppression (i.e., inhibition of an emotionally expressive behavior), masking (i.e., emotion experienced with a happy expression) and control (i.e., no regulation) conditions. Among reappraisers, who were instructed to adopt a neutral attitude toward the film, regardless of whether they were high or low in emotional awareness, there was not a significant increase in error responses. This study shows that the potentially damaging impact of negative emotions on the processing of emotional information can be prevented by a high emotional awareness or with the implementation of reappraisal as an emotion regulation strategy.

. Introduction Emotion regulation (ER) refers to the process by which we influence what emotions we experience, when we experience them, and how we express them (Gross, 1998). In his model, Gross (1998) distinguishes two major classes of ER strategies that occur in the emotion-generating processes. Antecedent-focused ER acts early in the emotion generation process (i.e., before the emotion response tendencies become fully activated and have changed behavioral and peripheral physiological responding). Response-focused ER refers to regulatory processes that occur after an emotion has been generated and involves emotion modification once an emotion has been elicited and once response tendencies have been fully activated. Two ER strategies have been widely studied in recent years: Expressive suppression and cognitive reappraisal (Gross, 2002). Reappraisal is a cognitive ER strategy that involves changing the way one thinks about a potential emotion-eliciting situation to reduce its emotional impact. Suppression is a behavioral ER strategy that aims to inhibit ongoing emotionally expressive behaviors. These strategies have different consequences. Reappraisal leads to a decrease in both expressive behavior and the experience of negative emotion whereas suppression leads to a decrease in behavioral responses, yet it fails to decrease the emotional experience. There is also evidence that suppression increases physiological responding and that it impairs cognitive functioning whereas reappraisal does not have such an effect (Gross, 2002). With previous research concentrated mainly on the consequences of using ER strategies, there has been relatively little investigation into the role of individual differences that may influence ER outcomes. The concept of emotional awareness (EA), introduced by Lane and Schwartz (1987), provides a promising outline of this idea. Lane and Schwartz (1987) define EA as the ability to identify and describe one’s own emotions and other people’s emotions. According to their model regarding the cognitive-developmental levels of EA, EA is a cognitive skill that undergoes a process, which is structurally parallel to Piaget’s stages of cognitive development. The five levels of EA are hierarchically arranged in the following manner: bodily sensations, action tendencies, single emotions, blends of emotions, and combinations of blends of emotional experiences (see Lane & Schwartz, 1987, for a comprehensive description). According to Lane and Schwartz, the degree of differentiation and integration within the structural organization of EA is reflected through verbal descriptions of emotional experiences. Lane, Quinlan, Schwartz, Walker, and Zeitlin (1990) developed the Levels of Emotional Awareness Scale (LEAS) to measure individual differences in the differentiation in the use of emotional words while describing one’s own emotional experiences and the emotional experience of others. This instrument is scored from an analysis of an individual’s verbal descriptions of emotions, which are provided in response to short scenarios depicting emotion-eliciting situations. High scores on the LEAS, which indicate greater EA, have been associated with greater self-reported impulse control (Lane & Pollermann, 2002). Lane and colleagues (1998) observed a positive relationship between LEAS scores and increased activity in the dorsal anterior cingulate cortex (dACC) during emotional stimuli processing. Dorsal ACC is recognized as crucial for the cognitive control of emotion (Medford & Critchley, 2010). These results suggest that individuals higher in EA recruit more dACC activity during emotional stimuli processing, which possibly optimizes their behavior in response to emotional stimuli. The aim of the present study was to examine whether EA moderates the effects of ER processes. We designed an experiment in which a film that activated intense aversive emotion (i.e., disgust) was presented to participants who were allocated to one of four groups differing in ER strategy: reappraisal, suppression, masking (i.e., expressing positive emotions regardless of real feelings), and control (i.e., no regulation). We included the masking condition in an attempt to broaden the knowledge about the consequences of response-focused regulation. Thus far research has focused mainly on suppression, however, many a time people want to not only suppress an emotion but also cover felt negative emotions with a happy expression (Ekman & Friesen, 2003). To measure emotional information processing (EIP), a set of 12 tasks, which involved matching the facial expressions of emotions, was created (i.e., the Facial Expressions Matching Test, FEMT). The level of performance on these tasks (i.e., the number of errors) was measured before emotion induction (i.e., six tasks, FEMT1) and after emotion induction (i.e., six tasks, FEMT2). On the basis of previous results showing that negative affective states impair the ability to decode facial expressions of emotion (Chepenik et al., 2007 and Schröder, 1995), we assumed that emotion induction would lead to poorer EIP. Prior to the experiment, we conducted a pilot study. A total of 24 female students (mean age = 24.17 years): completed the FEMT1, viewed a film inducing disgust while instructed to allow their emotions to arise, and they completed the FEMT2. The percentage of error responses in the FEMT2 was higher than in the FEMT1 (M = 41.25, SD = 18.09, M = 29.17, SD = 18.01, respectively), t(40) = 3.24, p < .001, Cohen’s d = .53. There was not a significant difference between mean times needed to perform the task [FEMT1 = 5.71 s. (2.51), FEMT2 = 5.55 s. (2.29), t < 1]. These results confirmed our assumption that intense emotion leads to poorer EIP. To rule out other explanations of performance deterioration (e.g., boredom or tiredness) we conducted a follow-up study in which we asked 32 female students (mean age = 23.12 years) to complete the FEMT1, view a neutral film and complete the FEMT2. There was no significant deterioration in the performance (FEMT1: M = 26.37, SD = 14.58; FEMT2: M = 29.31, SD = 16.25; t < 1). The results of the pilot study are consistent with previous research showing that intense negative affective states reduce available attentional resources for performing prospective tasks ( Eysenck & Derakshan, 2011). The next question is: Can ER prevent this unfavorable impact of negative emotions on emotional processing? Previous research shows that reappraisal increases activation in the dACC ( Miller & Cohen, 2001) and diminishes self-reported negative affect ( Gross, 1998), physiological responding to emotion-triggering events ( Gross, 2002), and amygdala activity ( Gross, 2002), which plays a crucial role in encoding emotional stimuli. Accordingly, one can assume that negative emotion is lessened by reappraisal and presumably interferes less with the attentional resources needed for performing incoming tasks ( Baumeister, Heatherton, & Tice, 1994). Therefore, we hypothesize that reappraisers would demonstrate a smaller increase in error responses than participants in the other three conditions (Hypothesis 1). The ability to represent emotion in a differentiated and complex way has been cited as pivotal for effective and adaptive coping with emotions and is a necessary prerequisite for executing regulatory processes (Lindquist & Barrett, 2008). Thus, we assume that EA provides compensatory resources that prevent the deterioration of emotional processing that is caused by negative affect. We predict that individuals high in EA will show a smaller increase in FEMT error responses, as measured by a difference score calculated from the error responses in the FEMT2 and FEMT1, than individuals low in EA. However, we expect an interaction effect, in that the beneficial effects of EA will be observed particularly in the control, suppression, and masking conditions (Hypothesis 2). Because of gender differences in EA ( Ciarrochi, Caputi, & Mayer, 2003) only women were recruited for this study.

Results 3.1. Preliminary results Responses on both the FEMT1 and FEMT2 were scored according to the errors responses (i.e., incorrect matches), which were subsequently changed to percentages. Table 1 presents descriptive statistics for the error responses (in %) for each FEMT. There were no differences between conditions in FEMT1 error levels (F < 1). The results showed that there was not a significant change in the percentage of FEMT error responses (t < 1) in the reappraisal condition. However, there was an increase of errors in the three other conditions: suppression (t(29) = 4.36, p < .001, Cohen’s d = .79), masking (t(29) = 5.64, p < .001, d = 1.05), and control (t(29) = 3.61, p < .01, d = .67). Table 1. Means and standard deviations for the error responses (in %) in the first and second FEMT for each experimental condition. Condition FEMT1 FEMT2 M SD M SD Reappraisal 31.66 20.22 33.89 15.47 Suppression 30.00 14.12 47.22 16.99 Masking 31.66 17.15 53.89 14.96 Control 27.61 16.19 39.44 16.47 Note. N = 120. n = 30. Table options There were no significant differences in mean time needed to perform the task between groups on the FEMT1 (F < 1) and FEMT2 (F < 1). The four experimental groups did not differ in EA levels, as shown in the LEAS scores: reappraisal (M = 61.07, SD = 9.73), suppression (M = 57.37, SD = 8.5), masking (M = 58.17, SD = 9.94), and control (M = 58.00, SD = 9.27), F < 1. 3.2. Main results To test the research hypotheses, we performed a hierarchical multiple regression analysis. The variables were entered into the regression equation in three steps. In the first step, three dummy coded contrasts were entered. In the dummy variable coding, reappraisal represented the baseline condition (coded as 0) against which suppression, masking and control conditions were compared (coded as 1). In the second step, we entered centred LEAS results. In the third step, we added three slope-dummy interaction products (contrasts × LEAS). As a dependent variable, we used the difference score between the means for the error responses (expressed in %) on the FEMT2 and FEMT1. The dependent variable was z standardized. The results above zero indicated an increase in the FEMT error response level. To compute the results in the regression model, we used statistical procedures recommended by Aiken and West (1991). The results of the regression analysis are shown in Table 2. First, the three regression contrasts were significant (adj. R2 = .10, F(3, 116) = 5.33, p < .01), which indicates that reappraisal led to a smaller increase in the level of FEMT error responses as compared to the other conditions. This result confirms our first hypothesis. Second, when the LEAS was added as a predictor, this increased the level of explained variance (adj. R2 = .13, F(4, 115) = 5.26, p < .001). The LEAS score was negatively related to an increase in FEMT errors; however, this correlation lost its statistical significance when interactions between contrasts, and the LEAS score were entered into the regression equation. In the third step of the regression analysis, we found that all expected interactions between contrasts and the LEAS score were significant (adj. R2 = .17, F(7, 112) = 4.47, p < .001). Table 2. Moderated regression of errors in FEMT. Variable ΔR2 B SE B β Step 1 .12⁎⁎ 1. Contrast 1: reappraisal vs. suppression .71 .25 .31⁎⁎ 2. Contrast 2: reappraisal vs. masking .94 .25 .41⁎⁎⁎ 3. Contrast 3: reappraisal vs. control .52 .25 .23⁎ Step 2 .03⁎ 1. Contrast 1 .63 .24 .28⁎ 2. Contrast 2 .88 .24 .38⁎⁎⁎ 3. Contrast 3 .46 .24 .20⁎ 4. LEAS −.40 .19 −.19⁎ Step 3 .07⁎ 1. Contrast 1 .67 .24 .29⁎⁎ 2. Contrast 2 .95 .24 .42⁎⁎⁎ 3. Contrast 3 .53 .24 .24⁎ 4. LEAS .49 .35 .23 5. Contrast 1 × LEAS −.71 .27 −.30⁎⁎ 6. Contrast 2 × LEAS −.58 .25 −.28⁎ 7. Contrast 3 × LEAS −.59 .26 −.27⁎ Note. N = 120. Dependent variable was the difference between FEMT2 and FEMT1, which indicated an increase in the level of error. ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. Table options Simple slope analyses clarified the nature of these interactions. As expected, low EA individuals who were assigned to the reappraisal condition demonstrated a smaller increase in FEMT error responses compared with individuals in the suppression (β = .91, t(112) = 4.38, p < .001) and masking conditions (β = .84, t(112) = 4.21, p < .001). The same favorable effect of reappraisal was observed when reappraisal was compared with the control condition, β = .72, t(112) = 2.41, p < .001 (see left part of Figure 1). There was no significant difference between the conditions with regard to an increase in error responses in individuals with high EA (see right part of Figure 1), ts < 1. Increase in error responses as a function of condition and emotional awareness ... Fig. 1. Increase in error responses as a function of condition and emotional awareness (as measured with the LEAS). Figure options Furthermore, we observed the following expected simple slope effects: an increase in FEMT errors was negatively predicted by the LEAS score in the suppression (β = –.36, t(112) = 2.12, p < .05) and masking conditions (β = −.32, t(112) = 1.96, p < .05), which means that individuals assigned to these conditions who were low in EA demonstrated a significantly greater increase in error responses than individuals who were high in EA. An analysis of the next interaction also revealed a significant simple slope effect in that the LEAS score was negatively related to the FEMT error increase level in the control condition, β = –.35, t(112) = 1.99, p < .05. There was no significant difference in the increase in the FEMT error responses between individuals who were low in EA and those who were high in EA in the reappraisal condition (t < 1). These results fully support Hypothesis 2.