key: cord-0070394-f1vqirkd
authors: nan
title: Top Ranked Abstracts from the 2021 Annual Meeting of the Society for Affective Science
date: 2021-11-23
journal: Affect Sci
DOI: 10.1007/s42761-021-00068-4
sha: 6f16c14daf0a4dba125707c8578f853f53d5e02c
doc_id: 70394
cord_uid: f1vqirkd

nan

Caregiver emotions influence child emotion regulation capabilities via modeling, responsivity, and other pathways which predicate on their internal ER selection and implementation (Thompson, 1991) . This influence is not unidirectional, but child traits and behaviors likewise shape caregiver responses over time (Lengua & Kovacs, 2005) . While ER is adaptive in helping us conform to our environments, best practice in ER strategy selection is likely situationally dependent (Troy et al., 2013) . Despite these understandings, few studies have tested these dyadic contingencies in the naturalistic caregiving context, as a factor of individual caregiver, child, and situational differences.

Aims Our overarching goal is to inform best practice in family intervention by contributing a new level of individual-and situation-dependent understanding of parent-child emotion interaction. Ecological momentary assessment (EMA) measurements will be collected four times per day for one week. These will capture caregiver regulation of their own and their child's emotions, and the influence of these caregiver ER choices on the subsequent dyadic affect and behavior. We will test four main EMA hypotheses regarding these data. These predictions are that 1) caregiver affect will predict child affect at subsequent timepoints, and child affect will predict caregiver affect at subsequent timepoints; 2) caregiver strategy to regulate their own emotions will predict caregiver and child affect at subsequent timepoints; 3) caregiver strategy to regulate their child's emotion will predict child affect at subsequent timepoints; and 4) child affect will predict caregiver selection of strategies to regulate their own and their child's emotions at subsequent timepoints. Methods Approximately 300 caregivers of children between the ages of 18 months and 5 years will be recruited from several geographical regions across the US. Daily EMA data will be collected through SMS surveys delivered four times daily in 3-hour intervals for 7 consecutive days. These surveys will briefly (5-min) gauge participants' affect, use of strategies to regulate their own and their child's emotion, and child affect and behavior. Intake surveys will capture parent and child emotion attitudes, emotional intelligence, parent-reported emotion regulation and behaviors, socioeconomic status (SES), and child temperament.

Proposed Analysis A series of hierarchical linear models with two levels (1-timepoints, 2-caregiver-child dyads) will be run to test our hypotheses. These models will assess how emotional states and behaviors at one timepoint predict change in states at the next timepoint, with 28 timepoints in total. We will control for average caregiver and child affect, time of day, day of week, child age, child sex, and family SES.

Conclusions Our study will clarify the individual-and situation-dependent nature of ER in a caregiving context, which carries important implications for generalizing family intervention across diverse populations. Future iterations will observe the influence of parent-child emotion patterns on long-term developmental outcomes.

Introduction Individuals regulate their emotions across diverse situations in daily life. Theory suggests that the use and function of emotion regulation (ER) strategies are context-dependent (Greenaway et al., 2018) and that flexibly adjusting ER based on environmental demands is beneficial for well-being (Aldao et al., 2015) . However, previous work has either only assessed variability in ER without considering context (which could capture random variability not tied to situational demands) or has focused primarily on younger samples and has not assessed consistency of ER within contexts as a measure of flexible adaption. As people age, they may become more skilled in flexibly adjusting their ER based on environmental factors. Accordingly, ER may be more situation-specific among older adults. Additionally, individuals may adapt their ER to situations in ways that could not be captured by mean-level differences in strategy use when in certain situations.

The present study has three aims. First, we will explore associations between situational context and ER strategies to determine what types of contexts are tied to specific forms of ER. Second, we will assess whether individuals with higher well-being adapt their ER to situational contexts more so than those with lower well-being. Third, we will assess whether older adults adapt their ER more to contexts than younger adults. We will use a relatively novel statistical modeling technique that enables consideration of situationspecific differences in both mean-levels and variability.

Methods A sample of 300 adults (25-85 years old) will participate in an experience sampling study. Participants will be randomly prompted 5x/day for 7 days to report on their momentary use of 14 different ER strategies and several aspects of their situation (location, activity, social context, appraisals). Participants will report on their current location and the main activity they are currently engaged in. To capture social context, participants will be asked to report who, if anyone, they are currently interacting with. To assess appraisals of the situation, participants will be asked to report on the familiarity and controllability of the situation. Participants also complete a depression scale (CES-D; Radloff, 1977) as an indicator of well-being at baseline.

ing multivariate mixed-effects location scale modeling (MELSM; Hedeker, Mermelstein, & Demirtas, 2008), we will assess whether mean levels and variability of 14 ER strategies are related to features of the participant's situation. We will separately assess whether mean-level and variability differences in respective strategies are predicted by a) locations, b) activities, c) social contexts and d) situational appraisals. Reduced variability in strategy use related to these aspects of the situation suggests within-situation consistency of strategies and therefore will be used to index flexible adaption of ER to situational context. We will use multivariate MELSM to test our hypothesis that well-being and age are positively associated with flexibility, as indexed by age and well-being predicting reduced situation-specific variability in ER strategies (i.e., more flexibility). (Gross, 2015) . Until recently, though, most emotion regulation measures have focused only on negative emotions (John & Eng, 2014) . Based on Gross's (2015) process model, the PERCI (Preece et al., 2018) was recently developed as a 32-item self-report measure of emotion regulation ability that integrates the assessment of both valence domains (example items: "When I'm feeling bad, I'm powerless to change how I'm feeling", "When I'm feeling good, I have no control over whether that feeling stays or goes").

Aims Few data are available on the PERCI. We examined its factor structure, measurement invariance, internal consistency, and concurrent validity. A secondary aim of this study was also to help establish what profile of emotion regulation strategy usage characterizes poor or strong emotion regulation ability. Results Confirmatory factor analysis supported the intended eight-factor (subscale) structure of the PERCI (e.g., CFI = .962, RMSEA = .036), thus supporting the separability of the negative/positive valence domains and the various subcomponents of emotion regulation ability. This structure was invariant across age, gender, and education. All PERCI subscale and composite scores had high omega and alpha reliabilities (.81-.94), and correlated with psychopathology, emotional reactivity, and emotion regulation strategies in expected ways. For example, poor overall emotion regulation ability (i.e., high PERCI total scale scores), was associated with a regulation strategy profile characterized by increased behavioural avoidance (e.g., expressive suppression, withdrawal, ignoring) and more unhelpful cognitive patterns (e.g., focused on rumination, blame attribution, and catastrophizing), combined with decreased use of active problem solving and cognitive reappraisal. In regression analyses, the eight PERCI subscales together accounted for 46.1% to 51.8% of the variance in depression, anxiety, and stress symptoms.

The PERCI had strong psychometric properties. Given that previous measures have tended to focus only on negative emotions, future use of the PERCI in research/clinical settings may therefore help to develop a more comprehensive understanding of the multidimensional emotion regulation ability construct.

Introduction Online firestorms are characterized by negative affect, emotion contagion, activism, and rapid information exchanges on social media (Einwiller, Viererbl, & Himmelreich, 2017; Rost, Stahel, & Frey, 2016) . Given the power of online firestorms to affect social and political climates, emotion regulation (ER), the modification of one's emotional response to information, social interactions, or other emotion-provoking stimuli (Gross, 1998) , may play a critical role in elucidating the roles that social media users have in catalyzing and sustaining these online firestorms. ER strategy use, habits, and goals may, for example, reflect engagement and disengagement with emotion expression, influencing the types of content that users will spread versus avoid on social media.

Aims What role does ER play in propagating or attenuating the diffusion of emotional content in online firestorms? We hypothesized that individual differences in ER strategy behaviors would be observable in social media users involved in a online firestorm.

We conducted semi-structured, in-person interviews with 41 participants (31 identified as Female, and 2 as Non-Binary). Participants self-reported their race: 3 identified as Asian; 20 as Black, 15 as White, and 3 as mixed race.

The interview asked open-ended questions designed to capture each participant's personal recollections of, and social media experiences of, the police shooting of Keith Lamont Scott, which gave rise to local demonstration and riots. Next, each participant evaluated 11 authentic tweets from the event and indicated how they would have responded to the tweet (e.g. like or retweet) and why. Interviews were transcribed and coded by researchers for evidence of ER strategy use (e.g., suppression, avoidance).

Transcript analysis revealed 4 groups that differed in behavior and rationale. Each group embodied unique ER habits regarding a willingness to spread emotional content on social media during a online firestorm. Spectators (n=8; 88% female, 63% Black), are individuals whose behavior is observational in nature; they avoid publicly interacting with emotionally charged content and instead prefer to watch online exchanges. The other 3 groups comprise the class of "megaphoners," participants who engage with social media content in unique ways that align with distinct ER strategies. The first group, Emotion Expressors (n=14; 86% female, 50% Black), are motivated to amplify online firestorms by spreading emotional messages due to perceived emotional intensity and shock value. Expressors assume that social media is a place to seek, process, and express emotions. Negative Emotionavoiders (n=16; 69% female, 56% Black); however, avoid negative emotional content and only share messages they perceive as unifying or productive. Finally, Negative Emotion-suppressors (n=3; 33% female, 100% White) act to suppress any intense negative emotion on social media. They view intense negative content as orthogonal to productive discourse and instead advocate for facts; they respond to dominantly emotional content with contempt and suppressive messages.

Our thematic analysis yielded a social media activist typology (SMAT) that provides evidence of system level ER that may sustain online firestorms. Findings also clarify rationales for users' social media engagement and suggest that conflicting ER strategies among users may contribute to online firestorm spread.

Only few studies to date have examined effective connectivity during emotion regulation, and none of these has considered emotional intensity (e.g., Morawetz et al., 2016) . We explored for the first time how effective connectivity at rest may be related to emotion regulation ability of high-and low intensity negative emotions.

Aims We hypothesized that intrinsic effective connectivity of reappraisal-related brain regions would be (i) associated with the prospective reappraisal success and (ii) would be modulated by stimulus intensity.

Bayesian framework that allows one to make inferences about the causal (directed) interactions between brain regions of a coupled system. Here we use a recent variant of DCM, namely spectral DCM (spDCM), that was designed to model resting-state fMRI (rs-fMRI) (Friston et al., 2014) . spDCM estimates the effective connectivity that produces the average functional connectivity over the time period examined. We used fMRI and rsfMRI data of 26 participants collected during 3 scanning sessions separated by 1 week (22 females, mean age = 22.8 ± 3.3 yrs) (SIEMENS Magnetom 7.0 Tesla MR scanner; 32-channel head coil; 371 whole-brain images per run). First, we analysed the fMRI data of a standard emotion regulation task (Morawetz et al., 2016) using SPM12. Next, we extracted rs-fMRI time series from 11 regions of interest (ROIs), defined by the parametrically modulated neural activity during reappraisal of highand low-intensity negative emotions, for the analysis of the intrinsic brain network. Regions included left inferior frontal gyrus (IFG), bilateral precentral gyrus (preCG), bilateral inferior parietal lobe (IPL), middle temporal gyrus (MTG), supplementary motor area (SMA), precuneus and occipital cortex. We then analysed connectivity profiles in the neural circuitry, separately for the regulation of high-and low-intensity stimuli, by estimating hierarchical models over the parameters, which were specified within a hierarchical Parametric Empirical Bayes framework for DCM.

Reappraisal success for both high-and lowintensity stimuli was characterized by a top-down network with excitatory and inhibitory connectivity from the left IFG to preCG, IPL, and MTG to other structures in the network. These intrinsic connections systematically varied as a function of stimulus intensity. Specifically, the left IFG, IPL and preCG had more efferent connections than afferent ones. In addition, the right MTG was identified as integration hub, receiving input from eight other ROIs.

Conclusions This is the first study to unravel how effective connectivity within a reappraisal-related network at rest is linked to reappraisal success of highand low-intensity negative emotions. We demonstrate that the initial intrinsic causal network dynamics in the very same neural network that supports task-related activity during reappraisal might already explain parts of the variance of reappraisal success and varies depending on contextual demands. Emotional demands are part of theatre actors' work. Actors aim to please and animate audience (Green, 2009 ) and, therefore, a great amount of emotional labour is employed for this work (Ressegger-Slone, 2011). Emotional labour was first introduced by Hochschild (1983) and described as "management of feeling to create a publicly observable facial and bodily display" (p. 7). A small amount of research identified that acting and regulating emotions can cause a 'possession syndrome' (Evans & Wilson, 1999) . However, there is no evidence on the specific emotion regulation strategies and the outcomes that actors experience.

Aims Actors employ diverse emotional labour strategies (Blix, 2007) , however, there is a scarcity of research linking emotional labour strategies to the outcomes perceived by theatre actors. Hence, the aim of the study is to design a provisional model of emotional labour strategies and outcomes by tackling two goals:

1. Identifying emotional labour strategies that are employed by theatre actors.

2. Discerning the outcomes of emotional labour as reported by theatre actors.

Methods An explorative semi-structured interview was conducted with 7 theatre actors. Acting experience in years varied from 3 to 42 years. The sample comprised of 4 males and 3 females with a mix of Asian (n=1) Caucasian (n=1), Black African (n=3), and Black Caribbean (n=2), living in Greater London. Participants' age ranged from 22 to 62 years of age. Data was analysed applying principles of a thematic analysis.

As presented in Table 1 , five themes describing emotional labour strategies were identified, i.e. (1) Consulting self, (2) Selcuk et al., 2012) to assess the extent to which PFs promote affective recovery. Each trial consisted of reading a distressing news headline and then viewing a photograph of either a PF or a matched unknown control. For each trial, mood (e.g., "How good do you feel right now?"; 1 (not at all) to 7 (extremely)) was assessed three times: at baseline (T1), after reading the headline (T2), and after seeing a photograph of the PF or matched control (T3). 1 Participants completed four trials total (2 for PF and 2 for control), in a semi-randomized order that ensured that PFs and matched controls were not viewed consecutively. Finally, participants completed a new scale to assess their positive feelings, thoughts, and behaviors with regards to the PFs and matched controls.

Overall, the negative consequences of reading distressing news were dampened by viewing PFs (in 1 Participants were assigned to one of two conditions, but here, we focus on the results of the "recovery" condition, for the sake of simplicity.

comparison to their matched controls) as measured by positive mood change (see Figure 1 ; Conclusions A major source of negative affect in daily life is the news. The present work shows that viewing photographs of PFs helped people recover more quickly from reading distressing news headlines. Given the ubiquity in which reminders of PFs and distressing news cooccur, findings suggest that the affective regulatory power of PFs may be pervasive.

Positive reappraisal, or reframing an event to feel better, consists of separate process steps (McRae & Gross, 2020) yet few studies have distinguished these steps to discern when and how emotion is actually impacted (Tamir, Halperin, Porat, Bigman, & Hasson, 2019) . We aimed to distinguish the generation of positive reappraisals from their implementation. Generating positive reappraisals relies on semantic processing to create an initial set of possible alternative meanings of the negative event. At this stage, these possible alternative meanings have not yet been vetted for their effectiveness in producing emotion change and may co-exist with negative (re)appraisals of the event. Alternatively, the implementation of positive reappraisals requires the selection of potentially impactful alternative meanings, elaborating on those candidate meanings and monitoring for success in changing emotion.

Aims We aimed to experimentally separate the generation from the implementation of reappraisals. We hypothesized that people will report significant emotional changes from when they generate positive reappraisals to when they implement them.

Methods Participants (n = 78) viewed neutral and negative emotional images. Each trial (22 trials for each condition) featured two phases -a generation phase, in which they generated positive and/or negative reappraisals of the emotional image -and an implementation phase in which they were focused on one of those reappraisals to make themselves feel better/worse. On some trials, they just looked at and reacted naturally to the images. On each trial, participants were first told that they were going to either generate positive reappraisals (+ +) and use the positive reappraisal later (use +), generate positive and negative reappraisals (+ -) and either use the posi-tive reappraisal later (use +) or use the negative reappraisal later (use -), or look at negative or neutral images. They then saw the image and generated the reappraisals (or just looked), rated their positive emotion (1-5), saw the image again and implemented the reappraisal (or just looked), and rated their positive emotion again.

As predicted, people's report of their positive emotion significantly and robustly changed from the generation to implementation of positive reappraisals in the expected directions (increase when using positive reappraisals and decrease when using negative reappraisals; Figure 1 ). Conclusions We found that people are not only able to separately generate and implement positive reappraisals, but that it's the implementation of these positive reappraisals that drives emotional change. This suggests the possibility that people can generate 'latent' positive reappraisals in stressful situations that may not necessarily impact their emotional state until they are able to implement them.

Introduction Interoception involves sensing, representing, and appraising the internal state of the body 1 . Interoceptive awareness plays a critical role in one's sense of well-being 2 . Mindful Awareness in Body-oriented Therapy (MABT) is specifically designed to cultivate interoceptive awareness and research links the therapy to reduced depression and improved emotion regulation 3, 4 .

Aims To better understand the neural representations of interoception and the effects of MABT, we use machine learning models to 1) classify whether our participants were attending to internal body sensations or external visual stimuli, 2) predict participants' moment-to-moment attentional states when they were asked to focus on the body, and 3) predict participants' attentional states when they watched sad or neutral film clips.

Methods In a recent study 5 , 24 community participants were randomized to an 8-week MABT or a waitlist control condition and assessed before and after the intervention. To achieve Aim 1, fMRI localizer data were collected when participants attended to internal body sensations or external visual stimuli (Int/Ext) and were either actively engaging in these processes or passively observing them (Act/Pas). Machine learning classifiers are trained to assess how accurately these four conditions can be distinguished from one another. To achieve Aim 2, participants were asked to sustain attention on their body for a few minutes. Machine learning classifiers trained on the localizer data are used to predict their moment-tomoment attentional states during the sustained attention task, in order to characterize how well they attended to their body sensations and whether MABT increased their ability to do so. To achieve Aim 3, participants watched a series of sad and neutral film clips. Machine learning classifiers trained on the localizer data are used to predict their moment-to-moment attentional states when they watched these film clips, in order to examine whether MABT changed the way participants coped with emotional stimuli.

Aim 1 results: Our classifiers can distinguish between the 4 localizer conditions with a 70% accuracy, well above the 25% chance level. Aim 2 results: Sustained attention is classified more as an active state than a passive state; there is no evidence that MABT promotes internal attention.

Conclusions Internal and external attentional states have distinct neural patterns that allow them to be reliably differentiated by machine learning classifiers. The neural patterns of sustained attention look more like an active state, suggesting that active self-monitoring might be employed to maintain attention. Our next step is to predict participants' attentional states when they watched sad and neutral film clips to examine whether MABT changes the way people regulate emotions in stressful situations.

Introduction Emerging research suggests that situation selection may be particularly effective for regulating emotions (Gross, 2014) , and has been associated with less negative affect and lower depressive symptoms in adults (Webb et al., 2017) . A hallmark of adolescence is increased autonomy, which can result in new opportunities to engage in this emotion regulation strategy. Additionally, the teenage years are associated with increased psychiatric risk (Casey et al., 2008) , so it is important to understand how teens regulate their emotions in ways that confer risk for depression and anxiety. Situation selection research has primarily focused on adult samples, so whether teens use situation selection to regulate emotions, and the mental health implications of situation selection strategy use, is unknown.

Aims This study examined situation selection strategy use and mental health in adolescents using a multimethod, longitudinal design. We specifically focused on how adolescents select into positive situations and avoid negative situations.

Methods Participants were 54 adolescents aged 13-17 years at Time 1, which consisted of 2 in-person lab visits and 1 week of daily reports. During the in-lab visits, participants completed questionnaires on their trait situation selection (Webb et al., 2017) , and anxiety and depressive symptoms (Chorpita et al., 2000) . Over the next week, participants completed daily reports about their situation selection use throughout the day. At the end of the week, participants completed the behavioral situation selection task; they were provided a list of short video clips labelled based on whether they were positively or negatively valenced, and were given 10 minutes to watch any videos in any order they wanted.

One year later, participants completed a 15-minute follow-up survey online (N = 40), which included surveys about current anxiety and depressive symptoms. Conclusions This study provides preliminary evidence that adolescents engage in situation selection as a way of regulating emotional experiences. Individual differences in avoiding negative emotional situations was associated with greater mental health symptoms cross-sectionally and one year later. This indicates that situation selection may influence the development and maintenance of mood and anxiety disorders, which may be especially important during the adolescent years when psychiatric risk increases.

Introduction Declining sexual desire is a common problem in long-term relationships. Existing treatments for increasing sexual desire include medication and couples therapy. It might be beneficial to increase sexual desire for the long-term partner specifically and without involvement of the partner. We tested the effectiveness of three cognitive strategies. The first strategy was positive reappraisal of the partner, which entails focusing on the positive aspects of the partner. The second strategy was reappraisal of sexual desire decline, which entails accepting without judgement that sexual desire declines over time in long-term relationships. The last strategy was sexual imagery, which entails evoking sexual thoughts and fantasies related to the partner.

The study aims were to test whether the three strategies increase 1) sexual desire for a long-term partner, 2) infatuation, attachment, and relationship satisfaction, and 3) motivated attention for the partner, as indicated by the late positive potential (LPP) amplitude.

Methods Twenty-five adult participants (6 men) who had been in a relationship for at least two years completed a regulation task while their electroencephalogram was recorded. Each trial consisted of a regulation prompt (see below) for 5 sec, a fixation cross for 500-700 ms, passive viewing of a partner picture for 1 sec, and a blank screen for 1 sec. An example of a positive reappraisal prompt is "Think of something funny your partner says". An example of a reappraisal of sexual desire decline prompt is "It's normal for sex drive to decrease even if the relationship is good". An example of a sexual imagery prompt is "Imagine something naughty your partner could say during sex". The no regulation prompts were three asterisks. After each condition, participants rated their sexual desire, infatuation, attachment, and relationship satisfaction. The LPP in response to the partner pictures was quantified by a mean amplitude measure between 400-1000 ms at electrodes F3/z/4, C3/z/4, and P3/z/4. Results Participants felt more sexual desire for and more infatuated with their partner after sexual imagery than after no regulation, both ps < .03, see Fig.  1 . Sexual imagery had no effect on attachment or relationship satisfaction. Positive reappraisal of the partner and reappraisal of sexual desire decline had no effect on sexual desire, infatuation, attachment, or relationship satisfaction. None of the strategies had an effect on the LPP amplitude.

Conclusions Sexual imagery may be used to increase sexual desire and infatuation in long-term relationships. Advantages of this strategy are that it does not involve medication or involvement of the partner.

Emotion regulation: Current status and future prospects

Three approaches to individual differences in affect regulation: Conceptualizations, measures, and findings

Measuring emotion regulation ability across negative and positive emotions: The Perth Emotion Regulation Competency Inventory (PERCI)

A DCM for resting state fMRI

Changes in Effective Connectivity Between Dorsal and Ventral Prefrontal Regions Moderate Emotion Regulation

Bad and worse: Neural systems underlying reappraisal of high-and lowintensity negative emotions

Stage actors and emotions at work

Fame: The psychology of stardom

The days of our lives: deep acting, surface acting and actors' health: a thesis presented in partial fulfillment of the requirements for the degree of Master of Arts in Psychology

The managed heart

Suicide by occupation

Suicide rates by major occupational group-17 states

Movement for the Actor: A Practical Approach to the Application of Movement Training

Emotion regulation

When there's a will, there's a way: Disentangling the effects of goals and means in emotion regulation

Interoception: The sense of the physiological condition of the body

How do you feel? Interoception: The sense of the physiological condition of the body

Interoceptive awareness skills for emotion regulation: Theory and approach of mindful awareness in body-oriented therapy (MABT)

Mindful awareness in body-oriented therapy as an adjunct to women's substance use disorder treatment: A pilot feasibility study

The adolescent brain

Assessment of symptoms of DSM-IV anxiety and depression in children: A Revised Child Anxiety and Depression Scale. Behaviour research and therapy

Emotion regulation: Conceptual and empirical foundations

Situation selection is a particularly effective emotion regulation strategy for people who need help regulating their emotions

Acknowledgements This project was supported by an NIA grant (R03AG057795) awarded to Tammy English.

Aldao, A., Sheppes, G., & Gross, J. J. (2015) . Emotion regulation flexibility. 

Summary Situation selection is the approach or avoidance of situations depending on their perceived emotional outcome. Studies of situation selection during adolescence are lacking, despite rapid developmental changes in emotional functioning during this period. This multi-method, longitudinal study assessed situation selection using a trait survey, daily reports, and a behavioral task, and measured anxiety and depressive symptoms both cross-sectionally and one year later. Selecting more negative stimuli during the behavioral task and greater daily avoidance of negative situations were associated with greater anxiety and depressive symptoms. Greater daily avoidance of negative situations was also associated with increased anxiety one year later. Results suggest that adolescents use situation selection as an emotion regulation strategy, which has implications for mental health symptoms.Keywords · emotion regulation · situation selection · anxiety · depression · adolescence Suzanna.powell@montana.edu