فراحالت خصلتی، پاسخ جنسیتی و EEG در طول تنظیم احساسات

Abstract Effective emotion-regulation is central to emotional intelligence. Relationships between the Trait Meta-Mood Scale (TMMS) and wellbeing may reflect individual differences in the strategies used for regulating negative emotions. The present study (N = 136) manipulated emotion-regulation strategy during viewing of a fear-inducing film clip. EEG response was assessed across five frequency bands in reappraisal, suppression and control conditions. The TMMS predicted higher power in theta and gamma bands, a pattern of response that may represent directed attention to emotional processing. Gender differences included elevated theta in females in the reappraisal condition, but effects of gender and Trait Meta-Mood appeared to be dissociable.

1. Introduction One of the primary functions attributed to emotional intelligence (EI) is emotion regulation (Wranik, Barrett & Salovey, 2007). More emotionally intelligent individuals should be better able to recognize and understand their emotional states, and to modify them to support their current goals. The broad tendency for EI to be associated with a variety of measures of emotional wellbeing and adaptive coping is consistent with this hypothesis (Zeidner, Matthews, & Roberts, 2012). Scales for EI correlate with self-report measures of emotion-regulation (Mikolajczak, Nelis, Hansenne, & Quoidbach, 2008), but rather little research has addressed the relationship between EI and specific emotion-regulation processes. Gross (2002) discriminated several strategies that support emotion-regulation, including reappraisal and suppression of emotions. The aim of the present study was to use electroencephalographic (EEG) methods to determine whether relevant facets of EI are associated with the neural processes underpinning emotion-regulation. EI refers to a multitude of constructs that are only loosely related (Matthews, Zeidner & Roberts, 2002). Emotion-regulation may be supported by both ability EI, especially objective emotion knowledge (Wranik et al., 2007), and personality factors described as ‘trait EI’. The present study focused on a trait EI measure, Salovey, Stroud, Woolery, and Epel’s (1995) Trait Meta-Mood Scale (TMMS), because it assesses three facets of EI that are directly associated with emotion-regulation. The first is attention; regulating emotion requires that the person is aware of their emotional experience. Next is clarity, referring to discriminating and understanding one’s different emotions clearly. The third aspect is repair, being able to change one’s emotional state constructively, for example, through reappraisal. The TMMS is quite effective in predicting scales for wellbeing and coping ( Extremera et al., 2011, Matthews and Fellner, 2012 and Zeidner et al., 2012). Salovey, Mayer, Goldman, Turvey, and Palfai (2002) found that the TMMS was associated with reduced psychophysiological reactivity to laboratory stressors, assessed using cortisol and blood pressure response, with attention being the most predictive subscale. The TMMS also has the advantage of being only moderately correlated with the Big Five traits (e.g., Bastian, Burns, & Nettelbeck, 2005). However, with a few exceptions ( Salovey et al., 2002), the validation evidence for the TMMS is limited by its reliance on self-report criteria. We based this study on Gross’s (2002) theory of emotion-regulation for three reasons. First, the theory specifies different regulative strategies in terms of information-processing model. Reappraisal is a strategy that modifies the encoding of an emotive stimulus, typically towards constructing a more positive meaning. By contrast, suppression operates later in processing, following extraction of meaning, such that the person attempts to inhibit behavioral expressions of emotion. Consistent with the Gross (2002) theory, use of reappraisal tends to be related to mental health, and suppression to psychopathology ( Nolen-Hoeksema, 2012). Second, building on Lazarus’ (1966) early work on appraisal and emotion, Gross and his colleagues have developed methods for manipulating strategies experimentally. Thus, we could test whether EI moderated responses to experimentally-induced regulative strategies. Third, studies have addressed the neuroscience of emotion regulation. Brain imaging studies (Buhle et al., 2013) suggest that reappraisal is supported by cognitive control regions, including lateral areas of prefrontal cortex, and lateral temporal cortex. Reappraisal may then actively modify semantic representations of emotional stimuli, leading to changes in emotional experience via modulation of activity in the amygdala (Buhle et al., 2013). The neural bases of suppression have been much less researched (see Ochsner, Silvers, & Buhle, 2012). Reappraisal has also attracted more attention than suppression in EEG studies. Reappraisal instructions tend to elevate the late positive potential (LPP) evoked response, and to suppress event-related frontal alpha, especially in the left hemisphere (Parvaz, MacNamara, Goldstein, & Hajcak, 2012). The lateralization may reflect verbal mediation of reappraisal. Dennis and Solomon (2010) suggested a capability model of frontal EEG. Individual differences in response reflect the interaction between the emotional demands of a specific situation and the person’s capabilities for regulation. Dennis and Solomon (2010) used film clips to induce emotion. Individual differences in change in frontal alpha, relative to a baseline measure, were associated with attentional interference, taken as an index of emotional dysregulation. This experimental methodology is straightforward to adapt to examine effects of EI on multiple spectral bands of the frontal EEG response. Studies of EI and EEG have focused mostly on evoked responses to emotive stimuli. Jaušovec and Jaušovec (2005) and Jaušovec and Jaušovec (2010) showed that an ability test for EI predicted greater event-related gamma activity, together with upper alpha desynchronization, during an emotion perception task. Freudenthaler, Fink, and Neubauer (2006) obtained a similar result for event-related alpha, and Aftanas and Varlamov (2007) found elevated alpha response in alexithymics exposed to emotive films. Trait EI measures, including the TMMS, are positively associated with amplitudes of P2 and P3 evoked potential components during processing of emotive stimuli, suggesting heightened attention (Raz, Dan, Arad, & Zysberg, 2013). These effects may reflect individual differences in elaborative emotional information processing, which may support cognitive reappraisal of emotional stimuli. Studies of event-related responses demonstrate that EI is associated with immediate stimulus processing. However, efforts at mood-regulation are often more protracted in time, lasting for minutes or longer. Also, the target of emotion-regulation is often internal affective states rather than external stimuli, so that it is difficult to examine the event-related response. Thus, the present study assessed EEG continually during exposure to a fear-inducing film clip. We also aimed to investigate gender differences in emotion-regulation capability, as indexed by the EEG. Typically, both ability and trait EI measures show higher mean scores for females, although gender differences vary across different facets of trait EI (Fernández-Berrocal, Cabello, Castillo, & Extremera, 2012). On the TMMS, only the attention scale reliably shows female superiority (Bastian et al., 2005 and Thompson et al., 2007). Gender differences in both emotion-regulation and in regulative brain activity are complex. Gross, Richards, and John (2006) found only limited gender differences in regulation. More recent research (Nolen-Hoeksema, 2012) shows that women report higher usage of reappraisal (along with various other regulative strategies), but there is no reliable gender difference in suppression. Studies of the EEG also suggest that gender may play a role in both EI and emotion-regulation, but findings are somewhat inconclusive. For example, Parvaz et al. (2012) failed to find any gender difference in LPP and frontal alpha responses to cognitive reappraisal. Jaušovec and Jaušovec (2010) reported complex gender differences in their studies of EEG responses evoked by processing emotive stimuli. They found that relationships between both cognitive and emotional intelligences and EEG may vary with gender. They also reported evidence for higher early gamma amplitude in females, a finding they attribute to enhanced early visual processing of emotion in females. Thus, it remains unclear whether gender differences in emotion-regulation, where found, can be attributed to higher EI in females. The aim of the current study was to investigate whether the TMMS and gender predicted frontal EEG response to reappraisal and suppression emotion-regulation strategies. Following baseline recording, participants viewed a fear-inducing clip from the film, the ‘Silence of the Lambs’. Three participant groups were instructed to use one of the two strategies, or no instruction was given (control group). It was hypothesized that the mood-repair subscale of the TMMS would predict EEG response, given that mood-repair refers to active management of emotive state. It was further hypothesized that mood-repair would predict EEG during reappraisal, because this strategy is more effective in up-regulating a negative mood than is suppression (Gross et al., 2006). Previous studies of evoked responses have most commonly implicated alpha activity in emotion-regulation (e.g., Parvaz et al., 2012), but theta and gamma bands are also sensitive to cognitive processing of emotive stimuli (Aftanas and Varlamov, 2007, Balconi and Pozzoli, 2009 and Jaušovec and Jaušovec, 2010). Given the inconsistency of previous findings, we did not make specific predictions of gender differences. We aimed to test the extent to which females showed EEG responses similar to those associated with higher EI, as assessed by the TMMS.

. Results 3.1. Correlations between TMMS scales and EEG In males, means (and SDs) of scores for attention, clarity and mood-repair were 50.1 (9.6), 42.7 (7.4) and 25.9 (7.2), respectively. Corresponding statistics for females were 58.8 (11.0), 43.8 (8.1) and 30.2 (6.5). The genders differed on attention (t(134) = 4.89, p < .01) and on mood-repair (t(134) = 3.68, p < .01). Women tended to score higher on these scales. Bivariate correlations between the three TMMS subscales and the EEG SPD measures were calculated. At baseline, significant correlations did not exceed chance levels. For the film-clip EEG measures, correlations were found for the whole sample, and within each experimental condition. Clarity was unrelated to any EEG measure. Attention and mood-repair correlated with theta and gamma measures as shown in Table 1. In the whole sample, both TMMS scales tended to be associated with higher bilateral gamma; mood-repair also correlated positively with theta. The TMMS scales were more strongly correlated with EEG in the reappraisal compared to the other conditions. Attention correlated with theta only in the reappraisal condition. Table 1. Correlations between two TMMS subscales and theta and gamma spectral power measures. Attention Mood-repair Theta Gamma Theta Gamma Left Right Left Right Left Right Left Right Whole sample .10 .11 .19⁎ .20⁎ 23⁎ .19⁎ .22⁎ .20⁎ Control −.16 −.11 .15 −.01 .18 .19 .24 .21 Suppression .05 .05 .14 .34⁎ .17 .07 .23 .16 Reappraisal .37⁎⁎ .32⁎ .20 .19 .37⁎ .36⁎ .26 .29⁎ ⁎ p < .05 ⁎⁎ p < .01. Table options Multiple regressions tested whether associations between TMMS and theta and gamma during the film clip were moderated by experimental condition. TMMS attention and mood-repair were centered. Condition was represented by two dummy variables. The first dummy was set to 1 in the reappraisal condition, and 0 in the others. The second dummy was set to 1 for suppression and 0 for the other conditions. Four product terms (2 TMMS scales × 2 dummy variables) were calculated to represent interactions. Because correlations were similar for left and right hemispheres, we also averaged attention and mood-repair measures across hemisphere, giving two criterion variables for the regressions. We also included EEG values at baseline to control for any pre-existing individual differences in theta and gamma. Two regression models were tested, with each of the film-clip EEG measures as the criterion. The linear model included five predictors: baseline EEG (theta or gamma), the two dummies for experimental condition, and the two TMSS variables. The interaction model added the four product terms. The linear model was significant for both theta (R2 = .14, p < .01) and gamma (R2 = .10, p < .05). Addition of the four product terms did not add significantly to the variance explained for either theta (ΔR2 = .02) or gamma (ΔR2 = .01). Thus, the linear model was preferred to the interaction model. In the linear model for theta, significant predictors were baseline theta (β = .30, p < .01) and repair (β = .19, p < .05). In the linear model for gamma, only repair (β = .19, p < .05) was independently significant. 3.2. Effects of gender and regulation strategy on EEG A series of t-tests (Bonferroni-corrected) showed no effects of gender on the EEG measures at baseline. Similarly, a series of one-way ANOVAs showed no effect of regulation strategy on EEG. Film clip EEG data were analyzed using mixed-model 2 × 3 (hemisphere × regulation strategy) ANOVAs, with repeated-measures on hemisphere. Significant effects were as follows: Theta. The main effect of gender was significant (F(1,130) = 4.76, p < .05, View the MathML sourceηp2 = .035), and the gender × strategy interaction was close to significance (F(2,130) = 2.93, p = .057, View the MathML sourceηp2 = .043). Follow-up ANOVAs tested the effect of gender at each level of regulation strategy. The gender difference was significant in the reappraisal condition (F(1,45) = 12.94, p < .01, View the MathML sourceηp2 = .223), but not in the other conditions. Mean log SPDs are shown in Fig. 1. Theta power was higher in females than in males in the reappraisal condition; means for the two genders were similar in the other conditions. Lower alpha power as a function of experimental condition and gender, in left ... Fig. 1. Lower alpha power as a function of experimental condition and gender, in left and right hemispheres. All error bars are standard errors. Figure options Alpha-1. The main effect of regulation strategy (F(2,130) = 3.38, p < .05, View the MathML sourceηp2 = .049) and the gender × strategy interaction (F(2,130) = 7.06, p < .01, View the MathML sourceηp2 = .098) were both significant. The hemisphere × strategy interaction (F(2,130) = 3.07, p < .05, View the MathML sourceηp2 = .045) was also significant. Follow-up ANOVAs to test the effect of gender in each condition showed a significant effect of gender in the control condition only (F(1,46) = 8.37, p < .01, View the MathML sourceηp2 = .154). Fig. 2 shows the means by condition; males showed higher alpha-2 power than females in the control condition. The hemisphere × strategy interaction appears to reflect a reduction of alpha-2 in the reappraisal condition in the right hemisphere only. Theta power as a function of experimental condition and gender, in left and ... Fig. 2. Theta power as a function of experimental condition and gender, in left and right hemispheres. Figure options Alpha-2. The only significant effect was the hemisphere × gender interaction (F(1,130) = 4.13, p < .05, View the MathML sourceηp2 = .031). Left hemisphere mean SPDs were similar for males (.77) and females (.80), but in the right hemisphere, the mean was lower for males (.76) than for females (.83). Beta. There were no significant effects. Gamma. The main effect of hemisphere was significant for gamma (F(1,130) = 29.93, p < .01, View the MathML sourceηp2 = .187), as was the hemisphere × gender interaction (F(1,130) = 4.92, p < .05, View the MathML sourceηp2 = .036). Left hemisphere means were similar for males (.67) and females (.69), but the mean for the right hemisphere was lower for males (.52) than for females (.62). Analyses thus far show that theta was elevated both in females, and in individuals high in mood-repair. We ran an ANCOVA to test whether the gender difference in theta could be attributed to the higher mood-repair score of females. The ANCOVA repeated the 2 × 3 (hemisphere × regulation strategy) ANOVA, with repair included as a covariate. It showed both a significant gender x strategy interaction, F(2,129) = 3.26, p < .05, View the MathML sourceηp2 = .046, and a significant effect of repair, F(2,129) = 4.63, p < .05, View the MathML sourceηp2 = .035. Thus, strategy-dependent effects of gender cannot be directly attributed to variation in repair