رویکرد فرد محور به تنظیم احساسات نوجوانان: ارتباط با آسیب شناسی روانی و پدر و مادر

Abstract Adolescence is a unique period of heightened emotional arousal and still-developing regulatory abilities. Adolescent emotion regulation patterns may be critically involved in adolescents’ psychosocial development, but patterns of emotion regulation in youths are not well understood. The current study used latent profile analysis (LPA) to elucidate patterns of emotion expression, experience, and emotion-related physiological arousal in adolescents. A sample of 198 adolescents and their primary caregivers participated in an emotionally arousing parent–adolescent conflict interaction. Adolescents’ observed emotion expressions, emotion experiences, and heart rate (HR) and caregiver parenting behaviors were assessed during and/or after the interaction. Parents reported on adolescents’ internalizing and externalizing symptoms, and youths reported on depressive symptoms. The LPA revealed four emotion regulation profiles: a moderate HR and high expression profile, a suppression profile (with low negative emotion expression and high emotion experience), a low reactive profile, and a high reactive profile. The moderate HR and high expression profile was associated with lower conduct disorder symptoms, the suppression profile was related to lower anxiety symptoms, and the high reactive profile was associated with higher adolescent depressive symptoms. The high reactive profile and moderate HR and high expression profile were associated with more negative/critical parenting behaviors. Findings suggest that profiles of adolescent emotion regulation can be empirically identified and may be significant risk factors for psychopathology.

Introduction Adolescence is a dynamic developmental period during which a series of hormonal, cognitive, and behavioral changes occur, leading to a heightened system of emotional arousal and a still-developing regulatory control system (Galvan et al., 2006 and Steinberg, 2005). Effective regulation of heightened emotional states is critical in navigating the novel stressors of adolescence such as puberty and shifting peer and family relationships (Eccles et al., 1993). Emotion regulation refers to the process by which emotions are automatically or volitionally monitored and modulated to facilitate a desired state or goal through internal processes and also external interpersonal influences (Cole et al., 2004 and Thompson, 1994). Adolescents who experience difficulties in regulating their emotions are vulnerable to poor social relationships and to internalizing and externalizing disorders, and they are more likely to engage in risky behaviors such as substance use (Bradley, 2003 and Silk et al., 2003). Within-person variability in adolescents’ responses across emotion domains (e.g., expression, subjective experience, physiology) may reflect meaningful patterns of emotion regulation (Zalewski, Lengua, Wilson, Trancik, & Bazinet, 2011b), perhaps affecting current and future mental health (Aldao, Nolen-Hoeksema, & Schweizer, 2010). Therefore, understanding types of emerging emotion regulation patterns in youths may provide insight into the development and progression of psychopathology and risk behaviors during a critical period for emotional development. Emotions are considered to be dynamic multisystem responses that include expressive, experiential (e.g., self-reported emotion), and physiological components. Early theories of emotion suggest a concordance model of emotion—that all types of emotion systems respond in concert to produce a coherent emotional response (e.g., Ekman, 1992 and Izard, 1979). For example, in a concordance model, anger may be expected to manifest in facial expressions ( Izard, 1979), appraisal of experience as anger, and increased physiological reactivity ( Ekman, Levenson, & Friesen, 1983). However, accumulating research reports a relative lack of correlation among expected emotion response systems in direction and magnitude ( Mauss & Robinson, 2009). This lack of association is often called emotion discordance (e.g., Evers et al., 2014, Hollenstein and Lanteigne, 2014, Lanteigne et al., 2012 and Mauss et al., 2005). Some research suggests that emotion discordance may result from emotion regulatory processes ( Butler et al., 2013, Hollenstein and Lanteigne, 2014 and Lewis, 2011). That is, the down- or up-regulation of emotion may affect certain emotion domains (e.g., expression) more than others (e.g., physiology), thereby resulting in emotional discordance among systems. For that reason, patterns of emotion discordance may indicate the presence of some form of emotion regulation. Theoretical models of emotion discordance/emotion regulation patterns Varying patterns of activation across emotion domains appear in a number of theoretical models of emotion with implications for emotion regulation and psychopathology. For one, Gross, 1998a and Gross, 2002 process model describes one response-focused emotion regulation strategy, expressive suppression, a strategy that reflects a distinct pattern of emotion discordance. Specifically, expressive suppression refers to high levels of subjective emotion experience and high physiological arousal but low expression of emotion (Gross, 1998a, Gross, 1998b, Gross and Levenson, 1997, Harris, 2001 and Richards and Gross, 1999). Indeed, Butler and colleagues (2013) found that participants prompted to suppress the expression of their emotions showed reduced emotional concordance across continuous measures of emotion expressive behavior and physiology, presumably with low expressive behavior but moderate or high physiology. Importantly, the consistent use of emotion regulation strategies such as suppression is linked with depression, anxiety, and externalizing disorders (Aldao et al., 2010, Gross and John, 2003 and Larsen et al., 2013). A second theoretical model of emotion discordance involves a pattern of high reported negative emotion but low physiological arousal. This model has been proposed to be an under-reactive pattern of emotion regulation (Hastings et al., 2009 and Raine, 2002), and studies have found evidence for this type of emotion discordance in youths. For example, Hastings and colleagues (2009) found that high reported anger experience and low heart rate (HR) were associated with adolescent externalizing problems (ages 11–16 years). This finding is consistent with literature indicating that youths with conduct disorder tend to have lower emotional responsivity and physiological reactivity, including lower HR (Ortiz & Raine, 2004). This under-reactive pattern of emotion regulation may reveal an important risk pattern. Finally, a third theoretical model has proposed that a pattern of high reactive concordant responses—indicated by high levels of expressed emotion, experienced emotion, and physiological arousal—may occur in individuals, particularly when they are in intense emotional states (Mauss et al., 2005) or when resources are unavailable for inhibition or regulation (Lewis, 2011). Patterns of chronically high reactive concordant responses may indicate an under-regulated response style and also place individuals at risk for psychopathology because high negative emotion reactivity is linked with depression and problem behaviors during adolescence (Silk et al., 2003 and Wetter, 2009). Although these theoretical models suggest important patterns of emotion discordance and regulation, they have mostly been supported by self-report studies of emotion regulation styles (e.g., Larsen et al., 2013 and Silk et al., 2003) or are limited to discordance among two emotion domains in youths (e.g., Hastings et al., 2009). The assessment of multiple emotion domains can facilitate inferences about emotion regulation when discordant responses are present (e.g., low expression of negative emotion when physiologically aroused or subjectively experiencing high levels of anger) or when concordant responses are present (e.g., high reactivity across expression, experience, and physiological arousal). For this reason, the current study aimed to advance understanding of adolescent emotion patterns and their relationship with psychopathology symptoms by collectively interpreting multiple measures of emotion to find empirically derived person-centered patterns of emotion discordance/regulation. Person-centered approach An ideal strategy for understanding patterns of emotion discordance and regulation is through the use of person-centered statistical approaches such as latent profile analysis (LPA). Person-centered analyses may give way to a substantively richer understanding of emotion discordance and regulation by empirically identifying heterogeneous, within-person emotion patterns. LPA can simultaneously model activation across multiple emotion subsystems, allowing profiles of reactivity and the theorized latent variable, emotion regulation, to be revealed. Similar to cluster analysis, LPA groups individuals by the patterning of variable values but differs by using a more rigorous and model-based approach to determine subsets of the population (Muthén & Muthén, 1998–2012). To date, few studies have examined emotion regulation by using person-centered approaches, and very few studies have done so with adolescents. In light of this limitation, adolescents’ observed emotion expressions, reported emotion experiences, and HR in response to an emotionally arousing parent–adolescent conflict task were examined, and LPA was used to discern adolescent emotion regulation profiles. Person-centered studies of emotion regulation and psychopathology As noted above, only a few studies have examined emotion regulation patterns, particularly during adolescence, from a person-centered approach and have related those patterns to psychopathology. These initial studies suggest that patterns of high emotionality are linked to greater psychopathology symptoms, peer rejection, and negative appraisal styles. For instance, Zalewski et al., 2011a and Zalewski et al., 2011b examined children’s (ages 8–11 years) observed emotion expression self-reports of emotion and physiological arousal in a frustration-eliciting bead-sorting task and an anxiety-eliciting speech task. Emotion measures were explored through LPA, and results indicated that emotion regulation profiles were identified and differentially associated with coping, appraisal styles, and adjustment. Specifically, the “moderately responsive” and “unregulated responsive” profiles (characterized by moderate and high arousal across frustration emotion domains) were associated with higher conduct problems and depression (Zalewski et al., 2011b). Furthermore, this unregulated responsive profile was correlated with an appraisal style in which youths negatively evaluated others (Zalewski, Lengua, Wilson, Trancik, & Bazinet, 2011a). LPA also identified anxiety and frustration groups indicating “low response or well-regulated,” “response-regulated,” and “moderately responsive–expressive” children (Zalewski et al., 2011b, p. 958). In a second study using LPA, Smith, Hubbard, and Laurenceau (2011) investigated profiles of second-grade children’s anger control in a laboratory task where children lost a board game to a cheating child confederate. Based on children’s anger report, skin conductance, and anger expression, five groups were identified: “physiology and expression controllers,” “expression-only controllers,” “non-controllers,” “non-reactive,” and “non-reporters” (Smith et al., 2011, pp. 221–222). Although this study did not explicitly examine psychopathology symptoms in children, findings revealed that the expression-only controllers (low expression, high physiological arousal, and high self-report) and non-controllers (high levels across anger domains) were considered to be more aggressive and were more disliked by their peers. Finally, Lanteigne and colleagues (2012) used cluster analyses to determine subgroups of emotion discordance patterns in a small sample (N = 49) of adolescent girls (ages 12–17 years). In response to a standardized speech task, the “experience-expressive” cluster (with high expressed emotion, high self-reported self-conscious emotions, and lower physiological arousal) was associated with greater difficulties in regulating emotions and more internalizing problems relative to a higher physiological arousal group called the “arousal” cluster ( Lanteigne et al., 2012). Each of these studies highlights the importance of using person-centered approaches that include multiple emotion domains and finds that high reactive emotion profiles may be maladaptive. However, beyond standard stress- and emotion-eliciting tasks, there is a need to understand emotion patterns within significant interpersonal relationship contexts during adolescence, especially because external social influences (e.g., parenting behaviors, responses to adolescent emotion) may actually regulate youths’ emotional states and youths’ emotion regulation patterns may influence future interpersonal relationships (Cook et al., 2013 and Thompson, 1994). Adolescents encounter many new social circumstances with the reorganization of family roles, increasing rates of family conflict, and more time and value placed on peer relationships during adolescence (Larson & Richards, 1991). Although peer relationships are significant and evolving during this developmental period (Brown & Larson, 2009), the parent–adolescent relationship is also a central socializing agent for youths’ emotional functioning throughout adolescence (Eisenberg, Cumberland, & Spinrad, 1998). Emotionally arousing interpersonal contexts with caregivers are of particular relevance because family difficulties have been implicated in adolescent psychopathology (Lewinsohn et al., 1994, Sheeber et al., 1997 and Steinberg et al., 1994). Given this key role, parent–adolescent interactions may reveal maladaptive patterns of adolescent emotion that may place adolescents at risk for psychopathology. Hence, this study builds on existing person-centered studies by exploring adolescents’ patterns of emotion responses in an ecologically valid parent–adolescent conflict interaction. Emotion regulation profiles and parenting Emotion regulation patterns, such as the profiles described above, are influenced by or related to a number of social processes within the family environment (Eisenberg et al., 1998). Although the nature of parent–child relationships changes during adolescence, parenting style, practices, and parental emotionality are considered to be important interpersonal influences on child emotion regulation throughout development (Morris, Silk, Steinberg, Myers, & Robinson, 2007). However, the majority of research investigating parental factors and children’s emotion regulation has focused on infancy through early childhood (Bariola, Gullone, & Hughes, 2011). To our knowledge, no person-centered studies of emotion regulation during adolescence have examined correlations between emotion regulation profiles and parenting or parents’ emotional functioning. Existing variable-centered research examining parenting styles and young children’s emotion regulation shows that observed or self-reported negative parenting (e.g., over-reactive discipline, negative control, hostility) is related to poor emotion regulation, as indicated by youths’ poor effortful control (Karreman et al., 2008 and Morris et al., 2002), low vagal tone (Calkins, Smith, Gill, & Johnson, 1998), and higher frequency of negative emotion displays (Del Vecchio & Rhoades, 2010). Although parenting style may affect youths’ emotional development over time, youths’ expression of dysregulated negative emotion and problem behaviors also elicit negative parenting behaviors (Huh et al., 2006 and Patterson, 1982). Given the important correlations between parenting and the development of emotion regulation, the current study explored the relationship between negative parenting behaviors in a conflict interaction and adolescent emotion regulation profiles in that interaction. Emotion regulation profiles and gender Another factor that may influence adolescent emotion reactivity and regulation is gender. Gender has been widely examined in relation to styles of emotion reactivity and regulation. Girls are often considered to be more emotionally expressive than boys, particularly for positive emotion and for “softer” negative emotions such as sadness and anxiety (Brody and Hall, 2000, Buck, 1979, Chaplin and Aldao, 2012 and Gross and John, 2003). Thus, gender may play a role in an adolescent’s pattern of emotion responses and emotion regulation, with girls potentially showing greater emotional reactivity patterns. Because gender may be associated with differential styles of expression, reactivity, and regulation, the current study examined gender as a correlate of emotion profiles. The current study To examine profiles of emotion regulation during adolescence, multiple indicators of adolescents’ responses to an ecologically valid emotion elicitor (parent–adolescent conflict interactions) were measured. We examined observed negative and positive emotion expression, internal emotion experience reported by the adolescent, and HR reactivity. In our view, each measure reflects a distinct component of the emotion response. Emotion expression refers to the observable and behavioral manifestation of the emotion experience, including facial and vocal cues. We chose to include both negative and positive emotion expression because adolescents may possess advanced means of masking emotion to meet social norms, potentially by expressing positive emotion in this interpersonal context. An emotion experience relies on the individual’s appraisal of his or her emotional state, including the perception of physiological manifestations, emotion expression, and associated cognitions. Although not specific to negative emotional arousal, HR reactivity was conceptualized as a measure of emotion-related physiological arousal. HR reactivity is related to both sympathetic and parasympathetic activity and was measured through change in HR (Mauss & Robinson, 2009). The current study’s first aim was to determine whether subsets of adolescents could be identified based on emotion variables through LPA. We theorized that the latent variable, emotion regulation, can be inferred from profiles of emotion expression, experience, and physiological arousal. We predicted, based on past theory and research, that profiles of over-regulation (suppression), under-regulation (high reactive), blunted physiology (low HR), and normative arousal (low reactive) would emerge. The over-regulation profile, or “suppression” profile, was expected to consist of adolescents who showed low negative emotion expression and possibly high positive emotion expression, reported high levels of negative emotion, and exhibited heightened physiological arousal. This profile is theoretically analogous to expressive suppression. The under-regulation profile, or “high reactive” profile, was expected to consist of adolescents who showed elevated responses across all emotion domains and were hypothesized to represent ineffective regulation. The blunted physiological profile, or “low HR” profile, was expected to consist of adolescents who showed lower HR reactivity relative to their expression and experience of emotion. Finally, a profile of moderate or average responses across levels of emotion variables was expected to emerge, indicating a normative arousal or regulated response to a conflict interaction. Our second aim was to determine whether such profiles differentially relate to adolescent internalizing and externalizing symptoms and parenting. First, we hypothesized that adolescents with more psychopathology symptoms would fall outside of the low reactive profile. Specifically, we expected adolescents with greater internalizing problems to fall within a suppression profile or a high reactive profile given theory and research linking internalizing problems to emotion suppression and to rumination/high reactivity (Larsen et al., 2013 and Nolen-Hoeksema, 2000). We expected adolescents with greater externalizing symptoms to exhibit an under-regulated response and fall within a high reactive profile or, alternatively, a low HR profile given evidence linking under-reactivity of emotion and low HR with conduct problems (Ortiz & Raine, 2004). We expected a moderate or low reactive profile to represent a well-regulated group of adolescents and adolescents with fewer symptoms to fall within this profile. Second, we hypothesized that observed negative parenting behaviors (e.g., criticism) would be associated with less regulated or “at-risk” emotion profiles—the suppression, high reactive, and low HR profiles. Finally, given gender differences found in emotion expression and regulation, we examined associations between gender and the latent profiles.

Results Descriptive statistics and correlations Table 1 provides descriptive statistics and correlations among adolescent emotion response variables, parenting, and adolescent psychopathology. Among emotion variables, self-reported anxiety and anger showed the strongest positive correlation. In addition, adolescent negative emotion expression was positively correlated with anger self-reports and negatively correlated with positive emotion expression. Correlations among emotion variables and parenting revealed two small yet significant relationships: Negative parenting was positively associated with adolescent-reported anger and negative emotion expression. Finally, correlations among emotion variables and psychopathology yielded significant relations between depressive symptoms and adolescents’ higher anxiety self-reports, anger self-reports, and negative emotion expression. Despite some small to moderate correlations, relationships among study variables support further examination using a multi-method, person-centered approach. Table 1. Means, standard deviations, and bivariate correlations among study variables. Mean (SD) Correlation 1 2 3 4 5 6 7 8 9 1. Adolescent-report anxiety 5.12 (2.20) 2. Adolescent-report anger 8.12 (3.68) .58⁎⁎ 3. Negative emotion expression 2.03 (1.02) .10 .43⁎⁎ 4. Positive emotion expression 2.70 (0.87) −.05 −.09 −.15⁎ 5. HR reactivity 3.40 (5.85) −.00 −.04 .03 .02 6. Negative parenting 1.72 (0.88) .03 .23⁎ .25⁎ .07 −.07 7. Depressive symptoms 7.15 (6.33) .19⁎⁎ .20⁎⁎ .21⁎⁎ −.04 .03 −.05 8. Generalized anxiety symptoms 10.34 (2.97) .05 .08 .10 −.11 .06 −.14⁎ .24⁎⁎ 9. Oppositional defiant symptoms 13.47 (4.41) −.00 .09 .09 −.14 −.13 .09 .18⁎ .49⁎⁎ 10. Conduct symptoms 15.99 (1.78) .03 .04 .04 .01 −.08 .16⁎ .25⁎⁎ .33⁎⁎ .57⁎⁎ Note. HR, heart rate. ⁎ p < .05. ⁎⁎ p < .01. Table options Latent profile analysis LPA identified a four-class model using the emotion variables with gender entered as a covariate. The best fitting model was determined by assessing multiple fit statistics and the substantive meaning of latent classes. Classes were added iteratively, and statistical information criteria such as Akaike’s information criterion (AIC; Akaike, 1987), Bayesian information criterion (BIC; Schwarz, 1978), and the adjusted BIC (Sclove, 1987) were considered, with lower information criterion values indicating greater model fit. The Bootstrapped Log Likelihood Ratio Test (BLRT) was also considered because a significant BLRT p-value suggests that the number of classes explains the model significantly better than one less class ( McLachlan & Peel, 2004). In addition, the entropy statistic provided information concerning group fit, with values approaching 1 suggesting greater distinction between classes ( Celeux & Soromenho, 1996). The four-class model demonstrated decreased BIC, AIC, and adjusted BIC along with a significant BLRT p-value (p < .001). Log likelihood ratio tests did not replicate for the five-class model, suggesting issues with convergence. Based on the aforementioned fit statistics, the four-class model was determined to be the best fit in addition to being theoretically meaningful ( Table 2). Table 2. Summary of fit statistics for latent profile analysis. Number of class AIC BIC Adjusted BIC Entropy BLRT One-class 4556.21 4595.67 4557.66 – – Two-class 4081.20 4137.10 4083.25 .94 .00 Three-class 4047.05 4125.96 4049.93 .94 .00 Four-classa 4019.12 4121.06 4022.85 .90 .00 Note. AIC, Akaike information criterion; BIC, Bayesian information criterion; BLRT, Bootstrapped Log Likelihood Ratio Test. a Five-class model did not replicate. Table options Emotion regulation profiles were inferred from the patterning of emotion variable means. Table 3 presents indicator variable means for each profile in the four-class model. The standard deviations from the sample mean for each profile are presented in Fig. 1. t-Tests were conducted to compare latent profile indicator means with the grand mean. Table 3. Latent profile means for adolescent report of emotion, observed emotion expression, and HR reactivity. Moderate HR and high expression Suppression Low reactive High reactive Profile 1 (14%) Profile 2 (15%) Profile 3 (62%) Profile 4 (9%) Adolescent-report anxiety 4.25 7.80 4.43 6.78 Adolescent-report anger 7.43 13.10 6.09 15.09 Negative emotion expression (scale: 1–5) 3.20 1.89 1.53 4.14 Positive emotion expression (scale: 1–5) 2.45 2.81 2.79 2.30 HR reactivity 5.20 2.56 3.14 4.11 Note. HR, heart rate. Table options Latent profile standard deviations from the sample mean for adolescent expressed ... Fig. 1. Latent profile standard deviations from the sample mean for adolescent expressed negative and positive emotion, heart rate reactivity, and self-reported anxiety and anger. Figure options Profile 1 comprised 14% of the sample and was characterized by moderate positive and high negative emotion expression, low to moderate emotion experience, and high HR reactivity relative to other groups. For this group, negative emotion expression levels were significantly higher than the grand mean, t(27) = 12.46, p < .001, self-reported anxiety was lower, t(27) = −2.74, p < .05, and HR reactivity was higher at a level approaching trend significance, t(27) = 1.62, p = .116. We called Profile 1 a “moderate HR and high expression” subset of adolescents. Profile 2 represented 15% of the sample and was characterized by moderate positive and negative emotion expression, high reported emotion (anxiety and anger), and moderate HR reactivity. In this profile, self-reported anxiety and anger were significantly higher than the grand mean, t(28) = 6.66, p < .001, and t(28) = 12.16, p < .001, respectively. Comparisons among standardized emotion indicators within this profile revealed that self-reported anxiety and anger were significantly higher than negative emotion expression, t(28) = 13.68 and t(28) = 26.84, respectively, p < .001. This profile may represent a “suppression” profile given that negative emotion expression did not meet the level of intensity of emotions experienced. Profile 3 contained 62% of the sample and was characterized by low to moderate responses across emotion indicators. Negative emotion expression was significantly lower than the grand mean, t(123) = −11.52, p < .001, self-reported anxiety was lower, t(123) = −4.86, p < .001, and self-reported anger was lower, t(123) = –16.71, p < .001. This profile may represent a “low reactive” response pattern and may signify a low reactive and/or well-regulated subset of adolescents. Profile 4 contained 9% of the sample and was characterized by high negative emotion expression, low positive emotion expression, high reported negative emotion, and moderate HR reactivity. In Profile 4, positive emotion expression was significantly lower than the grand mean, t(16) = −2.44, p < .05, negative emotion expression was significantly higher, t(16) = 12.36, p < .001, self-reported anxiety was higher, t(16) = 3.51, p < .01, and self-reported anger was higher, t(16) = 11.22, p < .001. This profile is considered the “high reactive” profile and may represent high emotional arousal and/or under-regulation of emotion. Multinomial regression Relationships between auxiliary variables and latent classes were examined through multinomial logistic regression with gender as a covariate in the LPA model. Table 4 presents odds ratios (ORs) and confidence intervals (CIs) of the associations between latent profiles and auxiliary variables, including adolescent depressive, GAD, CD, and ODD symptoms and negative parenting. Profile 3, the low reactive profile, was used as a reference group in regression analyses. Table 4. Odds ratios (95% CIs) of the relationship between auxiliary variables and latent profile membership. Moderate HR and high expression Suppression High reactive Profile 1 Profile 2 Profile 4 Depressive symptoms 1.06 [0.98–1.14] 1.05 [0.98–1.12] 1.10 [1.03–1.17]⁎⁎ Generalized anxiety symptoms 0.74 [0.60–0.91]† 0.78 [0.66–0.92]⁎⁎ 1.00 [0.81–1.23] Oppositional defiant symptoms 1.02 [0.90–1.15] 1.00 [0.90–1.12] 1.06 [0.92–1.29] Conduct symptoms 0.53 [0.32–0.89]⁎ 0.80 [0.60–1.07] 0.87 [0.53–1.41] Negative parenting 2.32 [1.32–4.10]⁎⁎ 1.65 [0.93–2.97]† 1.90 [1.06–3.43]⁎⁎ Note. CI, confidence interval; HR, heart rate. The reference group was Profile 3 (low reactive). † p < .10. ⁎ p < .05. ⁎⁎ p < .01. Table options Gender Multinomial regression indicated that girls were significantly more likely than boys to be members of the high reactive profile than the low reactive profile (OR = 2.77, p < .05, 95% CI [1.58, 4.87]). Internalizing symptoms Adolescents with higher depressive symptoms were more likely to be members of the high reactive profile relative to the low reactive profile (OR = 1.10, p < .01). Depressive symptoms did not significantly relate to the moderate HR and high expression profile or suppression profile. Adolescents with higher levels of GAD symptoms were less likely to be members of the suppression profile relative to the low reactive profile (OR = 0.78, p < .01). GAD symptoms did not significantly relate to the moderate HR and high expression profile or high reactive profile. Externalizing symptoms Adolescents with higher CD symptoms were less likely to be members of the moderate HR and high expression profile relative to the low reactive profile (OR = 0.53, p < .05). CD symptoms did not significantly relate to the suppression or high reactive profile, and ODD symptoms did not significantly relate to emotion profiles. Negative parenting Relative to the low reactive profile, higher ratings of observed negative parenting were associated with the moderate HR and high expression profile (OR = 2.32, p < .01) and to the high reactive profile (OR = 1.90, p < .01). For every 1-unit increase in observed negative parenting (measured on a scale from 1 to 5), adolescents were approximately two times more likely to be in the moderate HR and high expression profile and high reactive profile relative to the low reactive profile.