The Effect of Emotion Regulation on Executive Function

Abstract

Emotion regulation and executive function are associated: adaptive regulatory strategies are linked to better executive functioning while maladaptive strategies correspond with worse executive functioning. However, if – and how – these two processes affect one another has not previously been explored; most studies have employed a correlational approach, leaving the direction of influence unknown. We aim to address this gap by using an experimental design to explore the impact of emotion regulation on executive functioning. Adult participants (N=31) completed an executive functioning task (Computerized Task-Switching Test) under four induced emotion regulation conditions (1) neutral/baseline, (2) positive mood-maintain, (3) negative mood-maintain, (4) negative mood-reduce (conditions 2-4 were randomized). Relative to baseline, participants demonstrated better set-shifting performance across regulation conditions. In contrast, inhibitory control performance was slower, despite anticipated improvement due to practice effects. This suggests that inhibitory control may be more involved in the emotion regulation process than set-shifting when participants have a specific emotion regulation goal to achieve. The present study provides preliminary evidence that individuals’ ability to perform executive function tasks may be affected by concurrent emotion regulation demands; additional experiments are necessary to further probe the complexity of the association between these two processes.

We all experience a range of emotions that provide us with information that helps us to adapt in changing circumstances. However, emotions can also create problems when they are context-inappropriate or at the wrong intensity level (Nelis, Quoidbach, Hansenne, & Mikolajczak, 2011). When this happens, people typically employ strategies to regulate their emotions. The process of emotion regulation (ER) allows people to flexibly adjust and adapt according to external and internal demands (Pruessner, Barnow, Holt, Joormann, & Schulze, 2020). Executive function abilities, including set-shifting, response inhibition, and working memory, play important roles in emotion regulation (Hendricks & Buchanan, 2016; Morillas-Romero, Tortella-Feliu, Balle, & Bornas, 2015; Schmeichel & Tang, 2015).

Executive function capacity changes under different mood states (see Cohen et al., 2016) and the employment of certain emotion regulation strategies, such as rumination or suppression, leads to temporary declines in executive functioning (Curci, Lanciano, Soleti, & Rime, 2013; Lantrip, Isquith, Koven, Welsh, & Roth, 2016). However, better executive function also predicts greater emotion regulation capacity (Coifman et al., 2019; Tortella-Feliu et al., 2014). Although these observations seem contradictory, it may be that the nature of the association depends on the circumstances. One theory is that there are shared neural systems supporting both high-level cognitive processes and emotion regulation (Ahmed, Bittencourt-Hewitt, & Sebastian, 2015), so that when significant cognitive resources are committed to emotion regulation, executive functioning declines (and vice versa). It is important to learn more about how emotion regulation and executive functioning are related; both processes are critical to an individual’s well-being and knowledge about their underlying mechanisms could help us to develop strategies to help prevent the consequences caused by interference between these two processes.

Emotion Regulation

The Process Model of emotion regulation theorizes that emotion regulation occurs in parallel with emotion generation and that different regulatory strategies are recruited at different time points of the emotion-generative process to alter one’s emotional responses (Gross, 2014; Palmer & Alfano, 2017). Based on this model, the time point at which a person begins regulating their emotion could lead to different affective consequences. Being able to react early in the process is likely to lead to more effective emotion regulation because the regulation of emotion after it is fully generated may be less effective and requires more effort (Palmer & Alfano, 2017).

Depending on the emotion regulation goal and the context, a person may choose to employ different regulatory strategies. Strategies like reappraisal, problem solving, and mindfulness are typically considered as adaptive, as they are used to reduce dysfunctional emotions or enhance functional emotions, and have been linked to positive outcomes (Aldao, 2013; Aldao & Nolen-Hoeksema, 2010; Berking & Wupperman, 2012; D’Avanzato, Joormann, Siemer, & Gotlib, 2013; Joormann & Stanton, 2016). On the other hand, strategies like rumination, emotional suppression, and avoidance are considered maladaptive because they tend to increase dysfunctional or reduce functional emotions, and are associated with negative outcomes (Aldao, 2013; Aldao, Nolen-Hoeksema, & Schweizer, 2010; D’Avanzato et al., 2013; Gilbert, Nolen-Hoeksema, & Gruber, 2013). Of note, this distinction is not always clear and it is context-dependent; to be consistent with the extant emotion regulation literature, we will refer to strategies as “adaptive” or “maladaptive” based on typical use.

Emotion Regulation and Executive Function

Adaptive emotion regulation requires one to be aware of one’s emotion, to recognize the need to regulate the emotion, and to flexibly recruit and implement various regulatory strategies according to individual’s emotion-regulation goals and the situational demands (Gross & Jazaieri, 2014). Given the complexity of the process, it is not surprising that set-shifting, response inhibition, and working memory updating (WMU) – are directly involved (Hendricks & Buchanan, 2016; Schmeichel & Tang, 2015). Set-shifting refers to the capacity to shift attention between stimuli or tasks (Hendricks & Buchanan, 2016). This allows an individual to flexibly adjust the regulation of emotion to different emotional contexts. Response inhibition involves the ability to suppress context-inappropriate emotional response and goal-inappropriate regulatory strategies (Ahmed et al., 2015; Schmeichel & Tang, 2015). Working memory updating enables the maintenance of emotion regulation goals, allowing individuals to keep track of one’s current regulation of emotions so that it is consistent with the regulatory goals (Schmeichel & Tang, 2015; Silvers, Buhle, & Ochsner, 2014).

Neural Basis of Emotion Regulation and Executive Function

Evidence from neuroimaging studies indicates that the prefrontal cortex (PFC) is implicated in both executive function and emotion regulation (see Corbalan, Beaulieu, & Armony, 2015; Kolb & Whishaw, 2015; Ochsner & Gross, 2005). In addition to its roles in attention, working memory, and other cognitive processes that comprise the executive function network (Kolb & Whishaw, 2015; Silvers et al., 2014), several prefrontal subdivisions, including the dorsolateral prefrontal cortex, the ventrolateral prefrontal cortex, and the dorsomedial prefrontal cortex have prominent roles in the regulation of emotion (Ahmed et al., 2015; Silvers et al., 2014). These subdivisions are connected through different neural pathways to subcortical structures like amygdala, ventral striatum, and insula that generate emotions (Ahmed et al., 2015; Ochsner, Silvers, & Buhle, 2012). The fact that many of the same prefrontal cortical regions underlie both executive function and emotion regulation, suggests that the two processes share overlapping cognitive resources.

Effects of Executive Function on Emotion Regulation

Given the shared resources underlying emotion regulation and executive function, one could reason that individuals who have strong executive function would be effective at regulating their emotions and vice versa. In fact, greater executive function capacity has been linked to better emotion regulation ability across age groups (Sudikoff, Bertolin, Lordo, & Kaufman, 2015). Emotion regulation strategies also increase in number and become more sophisticated as youth mature and develop better executive function (Lantrip et al., 2016). Moreover, executive function has been found to be positively associated with adaptive emotion regulation and negatively associated with maladaptive emotion regulation (Hoorelbeke, Koster, Demeyer, Loeys, & Vanderhasselt, 2016). Additionally, improving one’s executive function through cognitive training can lower the tendency to employ rumination (typically a maladaptive strategy) when experiencing negative emotions (Hoorelbeke et al., 2016). This suggests that impaired executive function could lead to increased use of maladaptive regulatory strategies and decreased use of adaptive regulatory strategies (Hoorelbeke et al., 2016).

Effects of Emotion Regulation on Executive Function

However, underlying shared resources also mean that when one process is underway, the other may have insufficient resources, causing a decline in performance in both areas (Bonanno, Papa, Lalande, Westphal, & Coifman, 2004; Urry, 2016). Using a computer-based test of cognitive control, Cohen et al. (2016) found that healthy adults demonstrated diminished executive function under a prolonged threat condition, and that their cognitive control improved when an excited mood was induced. One explanation for this discrepancy is that the cognitive resources deployed to regulate emotions may vary under different conditions. For example, more cognitive effort might be put forth to regulate or inhibit negative emotions, which are unpleasant to experience, than positive emotions, which may not be regulated at all (Cohen et al., 2016; Cromheeke & Mueller, 2014; Diener, Kanazawa, Suh, & Oishi, 2015). In fact, emotional state changes have been found to influence the availability of cognitive resources for other types of information processing (Deveney & Pizzagalli, 2008).

Learning more about the relationship between emotion regulation and executive function has important implications, as these processes are implicated in various social processes, such as social cognition and decision making and play important roles in individuals’ functioning and well-being (Garcia-Andres, Huertas-Martínez, Ardura, & Fernández-Alcaraz, 2010). Moreover, research evidence over the past decades has consistently demonstrated the role of maladaptive emotion regulation in the development and maintenance of various psychopathology symptoms, including depression and anxiety (Aldao et al., 2010; Sheppes, Suri, & Gross, 2015). Thus, improved knowledge on the relationship between emotion regulation and executive function could also inform strategies to improve emotion regulation ability, which could benefit mental health.

Present Study

Most studies have explored (and therefore are more likely to find) that executive function either dictates or is strongly associated with one’s ability to regulate emotions. However, there is evidence to suggest that the relation goes the other direction – that emotion regulation affects and influences executive function. These observations obscure the direction of the relation between emotion regulation and executive function. To address this gap in the literature, our study focuses on the impact of emotion regulation on executive function. This approach may have greater utility as emotion regulation is a skill that can be targeted in treatment, whereas executive function – particularly those processes that are involved in emotion regulation (i.e., set-shifting, response inhibition, and WMU) are less malleable (Hendricks & Buchanan, 2016). We hypothesize that participants’ executive function performance on a computerized task-switching test will vary across three different emotion regulation conditions. Based on evidence of the impact of mood on cognitive abilities (Cohen et al., 2016), we expect that participants to have the best executive function performance when engaging in a relatively easy emotion regulation task (maintaining positive mood) and that executive function performance would decline with increased emotion regulation burden (reducing negative mood, maintaining negative mood).

Materials and Methods

Participants

Community participants between the ages of 25 and 40 were recruited through online advertisements. Exclusion criteria included: uncorrected vision impairment, cognitive impairment, current psychiatric diagnosis, or current substance use disorder. Participants were compensated for participating. This study was approved by the Institutional Review Board of Albert Einstein College of Medicine. All participants provided informed consent to participate in this study.

Measures

Demographic Questionnaire

Participants were asked to report on their demographic information, education history, substance use, and psychiatric history.

Emotion Regulation Questionnaire

This brief questionnaire was developed for the present study. It includes a visual analog scale for participants to indicate how well they were able to regulate or maintain their induced mood during the computerized task and asks about strategies used during the experiment. It ranges from 0 (not able to regulate/maintain the induced mood) to 100 (able to regulate/maintain the induced mood very well).

Depression Anxiety Stress Scale-21 (DASS-21)

The DASS-21 (Lovibond & Lovibond, 1995) was used to measure depression and anxiety in order to control for the influence of preexisting internalizing symptoms on participants’ mood during the experiment (Gross, 2014; Millgram, Joormann, Huppert, & Tamir, 2015). This measure consists of 21 items rated on a 4-point Likert Scale from which participants can choose from 0 (Never) to 3 (Almost always) to indicate how they felt anxiety (e.g., “I felt I was close to panic”), depression (e.g., “I felt that I had nothing to look forward to”), and stress (e.g., “I found it difficult to relax”) symptoms over the past week. In the current study, the Cronbach’s alpha was 0.75 for the anxiety subscale, 0.83 for the depression subscale, and 0.81 for the stress subscale.

Cognitive Emotion Regulation Questionnaire (CERQ)

This 36-item CERQ (Garnefski, Kraaij, & Spinhoven, 2001) assesses the cognitive emotion regulation strategies used by participants when facing stressful life events. This questionnaire was only administered once to each participant throughout the experiment session. Items like “I look for positive sides to the matter,” “I feel that I am the one to blame for it,” and “I think about how to change the situation” are included in this measure, and participants can choose from a 5-point Likert scale from 1 (Almost Never) to 5 (Almost Always). Higher scores indicate more frequent usage of certain emotion regulation strategies.

This measure assesses nine emotion regulation strategies. For the purpose of the current study, scores on four regulation strategies (positive refocusing, refocus on planning, positive reappraisal, and putting into perspective) were totaled into an overall adaptive regulation score; scores on the other five regulation strategies (self-blame, acceptance, rumination, catastrophizing, and other-blame) were totaled into an overall maladaptive regulation score (Van Meter & Youngstrom, 2015). The Cronbach’s alpha was 0.86 for adaptive regulation strategy and 0.81 for maladaptive regulation strategy.

Visual Analogue Mood Scale (VAMS)

The VAMS (Stern, Arruda, Hooper, Wolfner, & Morey, 1997) was used to assess participants’ moods at baseline and after each mood induction. Following the mood induction video, participants were required to answer the questions, “How excited do you feel right now,” “How triumphant do you feel right now,” “How strong do you feel right now,” “How happy do you feel right now,” “How fearful do you feel right now,” “How threatened do you feel right now,” “How angry do you feel right now,” and “How sad do you feel right now” to assess their mood. We measured these different moods as they represent the multifaceted aspects of positive and negative emotions. Each mood VAMS scale ranges from 0 to 100 (0 = absence of the mood, 100 = maximum level of the mood). Scores on the excited, triumphant, strong, and happy mood VAMS were summed to create a “total positive mood score” and scores on the fear, threaten, angry, and sad mood VAMS were summed to create a “total negative mood score”. The Cronbach’s alpha was 0.90 for positive mood and 0.74 for negative mood.

Mood Induction Videos

Three short video clips (~5 minutes each) were selected for their ability to reliably induce either a positive or a negative mood. The positive mood inducing video consists of a montage of women announcing pregnancy to their partners. The negative mood inducing video used in the “Negative Mood-Maintain the mood” (NegativeMaintain) condition consists of pet owners grieving over the death of their pets. The video used in the “Negative Mood-Reduce the mood” (NegativeReduce) condition consists of a speech in which a man expresses his guilt and regret for not listening to his late-mother’s words.

Computerized Task-Switching Test (CTST)

The CTST (Stoet, O’Connor, Conner, & Laws, 2013) was used to measure participants’ inhibitory control and set-shifting abilities. During each trial, participants are shown a square frame with the label “shape” on top of the frame and the label “filling” below the frame. Stimuli with different shapes (diamond or square) or fillings (2 dots or 3 dots) were shown in the top or bottom part of this square frame. The location where the stimulus appears determines how participants should respond to the stimulus (“B” keyboard button for square or 2 dots, “N” keyboard button for diamond or 3 dots). These trials are categorized into four different categories for the purpose of this study: congruent trials (shape and filling elicit the same response), incongruent trials (shape and filling elicit opposite response), non-switching trials (subsequent stimuli set is the same as the previous trial) and switching trials (subsequent stimuli set is different from the previous trial). At baseline, participants first completed 40 training trials, followed by 192 performance trials. In each mood regulation condition, participants only completed 192 performance trials. The type of trial (i.e., congruent/incongruent, non-switch/switch) is randomized within the program.

Procedure

Potential participants, who responded to an online ad, were asked to complete an online questionnaire to determine their eligibility. Eligible participants were invited for an in-person experimental session. After the informed consent process during the in-person session, participants completed a practice trial of the CTST. They then completed a baseline mood check and a full trial of CTST to determine their baseline executive function. Next, participants experienced one of three conditions: “Positive Mood-Maintain the mood (PositiveMaintain),” “Negative Mood-Maintain the mood (NegativeMaintain),” and “Negative Mood-Reduce the mood (NegativeReduce).” The mood condition order was randomized across participants to control for ordering effects. The randomization of the order was done using an online random number generator (random.org). All participants completed the same computerized task-switching test under all three conditions. During the transition intervals between conditions, all participants were asked to fill out a series of self-reports, including the demographic questionnaire, the DASS-21, and the CERQ. At the end of each condition, participants were asked to rate how successfully they stuck with the mood regulation task while completing the CTST. Participants were also instructed to count backwards starting from 100 to 50 to minimize the carryover effect of the mood inductions and to help participants’ mood to return to baseline (Figure 1).

Figure 1. Illustration of condition sequence that participants went through during the experiment.

Note. ERQ=Emotion Regulation Questionnaires, VAMS=Visual Analogue Mood Scale

Mood Induction and Regulation

In each mood condition, participants were asked to watch a brief video clip to induce either positive mood or negative mood. In a meta-analysis of mood induction techniques, movie clips were found to be the most effective way to induce mood, with average effect size across studies of r =.73 (Westermann, Spies, Stahl, & Hesse, 1996). After watching the video clip, participants were asked to rate on the VAMS to assess their induced mood. They were then instructed to maintain their positive mood or negative mood or to reduce their negative mood, depending on the condition. Participants were not given specific emotion regulation strategies to employ; Heiy and Cheavens (2014) reported that research participants often do not fully engage in assigned regulatory strategies in a laboratory setting and thus assigning a strategy may not impact behavior. Instead, a spontaneous regulation approach was employed, enabling participants to use the strategies they prefer. Specifically, in the PositiveMaintain condition, participants were told to “maintain your positive mood throughout the computerized task-switching test”. In the NegativeMaintain condition, participants were told to “maintain your negative mood throughout the computerized task-switching test”. In the NegativeReduce condition, participants were told to “reduce your negative mood throughout the computerized task-switching test”. A brief questionnaire was used to assess how well they were able to regulate their mood during the computerized task.

Statistical Analyses

The experiment was powered to detect medium within-person effects of the emotion regulation task on executive functioning (Faul, Erdfelder, Lang, & Buchner, 2007). This is consistent with effects found in an investigation of the influence of attentional control on spontaneous emotion regulation (Morillas-Romero, Tortella-Feliu, Balle, & Bornas, 2015).

Analysis was conducted using the SPSS statistical package and the lmer package for r (Bates et al., 2015). Four difference scores to assess inhibition and set-shifting were calculated using the recorded reaction time (RT) and error rate (ErR) from the CTST (Hendricks & Buchanan, 2016; Schmeichel & Tang, 2015). For inhibition capacity, difference scores of inhibition RT (congruent RT minus incongruent RT) and inhibition ErR (congruent ErR minus incongruent ErR) were computed. For set-shifting capacity, difference scores of set-shifting RT (non-switching RT minus switching RT) and set-shifting ErR (non-switching ErR minus switching ErR) were computed. This approach is consistent with other investigations (Hendricks & Buchanan, 2016; Morillas-Romero et al., 2015). We multiplied all the CTST outcomes with −1 to obtain positive outcome values for ease of interpretation. For all variables, higher scores indicate worse performance (i.e., more errors or longer reaction time) and lower scores indicate better performance (i.e., fewer errors or shorter reaction time) on the CTST. To examine participants’ CTST performance across conditions, taking into account individual differences in executive functioning, we used multilevel models with condition (dummy coded) and emotion regulation strategy scores (adaptive and maladaptive) as predictors. If either emotion regulation strategy score was significant, the interaction of that score and condition was added to the model. All data and research materials that support this study are available from the corresponding author upon reasonable request.

Results

Descriptive Statistics

A total of 31 participants were recruited for the current study. See Table 1 for clinical and demographic information. Participants’ performances on the CTST at baseline were 8.71 (SD=72.2) for inhibition RT, 4.06 (SD=6.4) for inhibition ErR, 430.13 (SD=204.5) for set-shifting RT, and 5.95 (SD=9.3) for set-shifting ErR (Table S1).

Table 1.

Demographic and Clinical Characteristics of Sample

	Participants
n	31
	Demographics n (%)
Sex and Gender
Male	13 (42)
Female	18 (58)
Ethnicity
African American/Black	13 (42)
Caucasian American	10 (32)
Asian/Pacific Islander	2 (6)
Native American/Alaskan	1 (3)
Bi-racial	1 (3)
	Demographics M (SD)
Age	31.65 (4.8)
Total years of education	17.00 (3.3)
	Clinical Characteristics
	Total M (SD)
DASS-21
Anxiety	1.87 (2.6)
Depression	2.61 (3.1)
Stress	4.13 (2.9)
CERQ
Adaptive regulation	47.03 (10.0)
Maladaptive regulation	47.06 (8.9)
	Mood Scores
	Total Positive Mood M (SD)	Total Negative Mood M (SD)
Baseline	152.55 (94.7)	11.77 (18.45)
Positive Mood	143.87 (102.8)	10.00 (18.1)
Negative Mood-Maintain	68.71 (86.1)*	58.06 (48.2)*
Negative Mood-Reduce	82.26 (89.0)*	54.19 (46.7)*

Note: DASS-21 = Depression Anxiety Stress Scale – 21, CERQ = Cognitive Emotion Regulation Questionnaire

^*significantly different from baseline mood at p < .001 level

Mood Manipulation Check

Positive Mood Condition

Participants’ total positive mood scores at baseline and after mood induction in the PositiveMaintain condition did not differ significantly. Related, there was no significant change between participants’ total negative mood scores at baseline and following the positive mood induction. When asked “how well were you able to maintain your positive mood from the video during the task,” participants rated an average score of 63.61 (SD= 22.6) out of 100, suggesting that they were able to maintain their induced positive mood moderately well (Table S2).

Negative Mood-Maintain Condition

Participants’ negative mood score after the mood induction in the NegativeMaintain condition was significantly higher than at baseline, t(30) = −5.95, p < .001. Participants also endorsed significantly lower total positive mood score than at baseline, t(30) = 6.09, p < .001. There was one participant who reported increased positive mood following the negative mood induction. When asked “how well were you able to maintain your negative mood from the video during the task,” participants rated an average score of 58.06 (SD=28.4) out of 100, indicating that they were able to maintain their induced negative mood moderately well (Table S2).

Negative Mood-Reduce Condition

Participants reported significantly higher total negative mood score following the mood induction in the NegativeReduce condition than at baseline, t(30) = −5.52, p < .001. As expected, participants endorsed significantly lower total positive mood score than at baseline, t(30) = 4.99, p < .001. When asked “how well were you able to reduce your negative mood from the video during the task,” participants rated an average score of 61.87 (SD=28.5) out of 100, indicating that they were able to reduce their negative mood moderately well (Table S2).

Differences in Executive Function Performance Across Conditions

Set-Shifting Reaction Time

Relative to the baseline trial, on average, participants had shorter reaction time in the PositiveMaintain condition (B = −65.42, p = .041) and NegativeReduce (B = −66.18, p = .038), indicating participants demonstrated improved set-shifting capacity in these two conditions (Table S3). The set-shifting RT in the NegativeMaintain condition was not significantly different from baseline. Paired comparison indicated no significant differences in performance across the three regulation conditions (Figure 2). Participant adaptive CERQ scores were marginally associated with set-shifting RT (B = 4.51, p = .073), but none of the interaction terms for CERQ score and condition were statistically significant.

Figure 2. Participants’ Reaction Time and Error Rate Across Conditions.

Set-Shifting Error Rate

Relative to the baseline trial, participants made significantly fewer errors in the NegativeMaintain condition (B = −5.54, p = .024). Performance in the PositiveMaintain condition (B = −2.33, p = .337), and NegativeReduce (B = −2.28, p = .348) were equivalent to baseline (Table S3). Paired comparisons indicated there were no significant differences between performance in the three emotion regulation conditions (Figure 2). Adaptive regulatory strategies were associated with fewer errors (B = −0.22, p = .042), there was no association with maladaptive regulatory strategies (B = −0.07, p = .546). The interaction of adaptive regulation with NegativeMaintain condition was significant (B = 0.50, p = .042) indicating that people with strong adaptive ER skills made more errors in the NegativeMaintain condition.

Inhibition Reaction Time

Relative to baseline, participants had slower reaction times in the PositiveMaintain condition (B = 31.65, p = .044) and NegativeMaintain conditions (B = 35.78, p = .023; Figure 2, Table S3). CERQ adaptive scores were significantly associated with inhibition RT (B = −2.67, p < .001), but none of the interaction terms for CERQ score and condition were statistically significant. This indicates that use of adaptive emotion regulation strategies is associated with worse inhibition (i.e., slower reaction time).

Inhibition Error Rate

Inhibition ErR was not associated with trial, performance was equivalent across conditions (Figure 2, Table S3). Adaptive CERQ scores were significantly associated with inhibition ErR (B = −0.28, p =0.038), but none of the interaction terms were significant. Maladaptive CERQ score was not associated with inhibition error.

Discussion

The primary goal of the present study was to investigate whether engaging in emotion regulation impacts executive functioning; both are crucial to our daily functioning and well-being, and a better understanding of how they influence each other could be valuable. Using mainly a correlational approach, past studies have shown that better attentional control and shifting is associated with more efficient emotion regulation (Hendricks & Buchanan, 2016; Morillas-Romero et al., 2015; Tortella-Feliu et al., 2014). Instead of focusing on the association between emotion regulation and executive function, the current study used an experimental design to test the hypothesized interference effect of emotion regulation on executive function. Our findings extend previous work and provide preliminary evidence that although emotion regulation and executive functioning abilities are associated (Hoorelbeke et al., 2016; Lantrip et al., 2016; Tortella-Feliu et al., 2014), concurrent emotion regulation may affect executive function domains differently.

We hypothesized that participants would demonstrate the best executive function performance in the PositiveMaintain condition, followed by the NegativeReduce and the NegativeMaintain conditions based on previous research showing the differential impact of mood on cognitive abilities (Cohen et al., 2016). Our hypotheses were not consistently supported; although RT was the slowest and number of errors was the highest in the NegativeMaintain condition, differences across ER conditions were not statistically significant. In general, participants demonstrated better set-shifting performance (fewer errors and faster response time) across the emotion regulation conditions relative to baseline. Inhibition performance was generally slower across conditions and the number of errors was higher in the two negative mood conditions. This suggests that engaging in emotion regulation may not interfere with one’s ability to flexibly shift attention, but that more cognitive resources may be required to inhibit responses.

Differences in Executive Function Performance Across Conditions

Set-Shifting Performance

Although there was no significant difference in set-shifting performance between the three regulation conditions, participants demonstrated faster set-shifting response times in the PositiveMaintain and the NegativeReduce conditions relative to baseline trial. Additionally, participants made fewer set-shifting errors in the NegativeMaintain condition relative to baseline, indicating better set-shifting performance (though set-shifting reaction time was equivalent in the NegativeMaintain condition and baseline). These results are contrary to our predictions and may be due to participant motivation and practice effects, rather than concurrent emotion regulation enhancing performance.

Although we found evidence that negative mood was successfully induced in both the NegativeMaintain and the NegativeReduce conditions, the induction of positive mood in the PositiveMaintain condition was only partially successful. This is not surprising; the induction of positive emotions tends to have smaller effect than the induction of negative emotions in experimental settings (Kučera & Haviger, 2012; Zhang, Yu, & Barrett, 2014). Our participants, on average, had relatively high positive mood scores at baseline, which may have made it difficult for them to feel more intense positive emotions by just watching a short video. It is also possible that the positive mood video content (pregnancy announcements) was not relevant or associated with positive feelings among our participants, whereas negative situations (such as a pet’s death) tend to be more universal and effective for inducing negative moods under experimental conditions (Kuijsters, Redi, de Ruyter, & Heynderickx, 2016; Kučera & Haviger, 2012; Siedlecka & Denson, 2019).

Relatedly, our participants’ moods might have an independent effect on their CTST performance. Even though negative mood inductions were successful, our participants still reported higher positive mood than negative mood after the watching the negative mood-inducing videos. This could have influenced their CTST performance as studies have showed that positive emotions can facilitate various cognitive abilities (Cohen et al., 2016; Wang, Chen, & Yue, 2017). They might help explain the unexpected result of statistically equivalent CTST performance across the three emotion regulation conditions.

The improved set-shifting performance in the emotion regulation conditions relative to baseline was unanticipated, but have been related to participants’ motivation and effort. There is evidence indicating that more mentally effortful tasks can be more rewarding, leading to enhanced focus on the task (Inzlicht, Shenhav, & Olivola, 2018; Pessoa, 2009). The effortful process of completing the CTST while regulating one’s emotion may have become intrinsically rewarding over time, leading participants to divert more attentional resources to the task. In fact, Pessoa (2009) reported that reward and motivation can strengthen an individual’s attentional orientation and shifting.

In addition, participants’ worse average CTST reaction time at baseline than in the emotion regulation conditions for set-shifting suggests that their performance was affected by practice. Although we had practice CTST trials at the beginning of the baseline condition, it is likely that, as participants became familiar with the task over repeated trials, their performance improved. Randomization of the conditions ensured that performance was not differentially affected in any single condition.

Inhibition Performance

Participants’ inhibition RT was slower in the PositiveMaintain and NegativeMaintain conditions relative to baseline, but error was not statistically different from baseline. This suggests that more effort was required to avoid errors and is consistent with our expectation and findings of past research showing that individual inhibitory control is impacted by emotion (Cohen et al., 2016; Schmeichel & Tang, 2015; Tabibnia et al., 2011). Past studies have involved emotion-eliciting visual stimuli as part of the executive functioning task (e.g., a go/no-go task where participants need to decide if a face is showing anger or sadness), the CTST presented emotionally neutral visual stimuli (e.g., shapes and dots) to participants in an induced emotional state. It may be that the effect on CTST performance would be more potent if participants had been tasked with making emotionally salient decisions on an emotional go/no-go task while regulating their mood. This approach would also better reflect the complex nature of the interactions between emotion regulation and executive function in real world settings. In addition to facing different emotional (e.g., making a high stakes work decision) and non-emotional tasks (e.g., deciding what to have for dinner), we also experience these tasks either simultaneously (e.g., trying to perform in a job interview while decoding the interviewer’s facial expression) or sequentially (e.g., taking an important exam after receiving the bad news of the death of a loved ones) in our daily lives. The temporal sequence in which we experience these emotional or non-emotional events could potentially have differential impacts on our executive performance. This is an important area for future research.

Association of Emotion Regulation and Executive Function

On average, our participants reported being able to regulate their induced moods only moderately well during each emotion regulation condition. This might be due to a number of factors. First, mood inductions typically only last a few minutes, so by the time the question was asked (after the task), participants’ moods are likely to have shifted toward their baseline mood, even if they initially regulated it successfully. Second, participants’ cognitive resources might have been shared between the CTST and the emotion regulation task, leading to mediocre performance on both tasks. This suggests a bi-directional relationship between emotion regulation and executive function. Thus, studies seeking to disentangle the influence of emotion regulation on executive functioning (and vice versa) are important to better understand the complicated nature of the two systems. Importantly, a participant’s mood outcome is not a proxy for their emotion regulation effort, so even though someone’s mood may not have been maintained or reduced, it does not mean that the manipulation to force concurrent emotion regulation and executive function exertion did not work.

A surprising finding was the observed relationships between participants’ CERQ scores and their CTST performance. Specifically, the CERQ adaptive emotion regulation was associated with fewer set-shifting and inhibition errors and with faster inhibition RT (set-shifting RT was not significantly associated). Conversely, the CERQ maladaptive emotion regulation was not significantly associated with any CTST variable. Past research shows that better executive functioning is not necessarily linked to more adaptive emotion regulation (McRae, Jacobs, Ray, John, & Gross, 2012; Tortella-Feliu et al., 2014). For example, although enhanced attentional shifting allows for adaptive flexible response in the face of a threat, it also facilitates attention to the threat stimuli which could lead to heightened anxiety (Tull, Maack, Viana, & Gratz, 2012). In our study, participants continued to report mildly positive mood after negative mood inductions which is consistent with findings from past studies (Diener, Cha, & Oishi, 2022). It is possible that people who are more likely to use adaptive strategies may have done especially well in the PositiveMaintain and NegativeReduce conditions because these conditions call for adaptive regulation of converging mood. In contrast, there was a significant interaction between ER and the NegativeMaintain condition, which could be that trying to maintain a negative mood was more difficult for those who are generally adept at maintaining a positive mood, especially when asked to be involved in divergent maintenance of opposite moods. This could result in more CTST errors.

Our findings provide preliminary evidence that the regulation of emotion has varying influences on the facets of executive function. Emotion regulation is a complex process and it is not surprising that its impact on an individual’s cognition could be dependent on the types of regulatory strategies implemented and at what point in the emotion regulatory process (Peuters, Kalokerinos, Pe, & Kuppens, 2019; Xiu, Wu, Chang, & Zhou, 2018). Consistent with the Process Model, it is likely that set-shifting capacity is more involved in the “attentional deployment” point of the process whereas inhibitory control capacity might be more involved in the following “cognitive change” point of the process. In order to achieve their emotion regulation goals, participants may have heavily utilized their inhibitory control capacity to control their emotional responses in ways that were consistent with the goals (the “cognitive change” time point), thereby interfering with their inhibitory control performance on the CTST. In contrast, the set-shifting may have been unaffected, as the role of attentional shifting occurs earlier in the emotion regulation process (i.e., the “attentional deployment” time point during the mood induction video), and may have no longer been engaged during the CTST. This could lead to relatively greater set-shifting capacity comparted to inhibitory control capacity on the CTST across all emotion regulation conditions, consistent with our findings. One might expect that this would lead to a drop on the CTST inhibition performance in the emotion regulation conditions relative to baseline, given that inhibitory control ability is more involved in the regulation of the induced mood. However, this pattern was not observed in the present study. A possible explanation is that the expected drop in the inhibition performance was offset by practice, leading to equivalent CTST performance across all conditions including the baseline. This would also help to explain why set shifting performance improved relative to baseline. Our finding suggests that once the emotion regulation process begins, we rely more on our inhibitory control ability, than our set-shifting ability, to monitor our regulation process to make sure that it is aligned with our emotion regulation goal.

Other factors such as individual’s mood, effort and motivation, as well as the nature and the duration of the executive functioning task are likely to contribute to the findings of the current study. For instance, congruent/incongruent trials might be more difficult than switching/non-switching trials, thereby affecting participants’ performance on CTST. Additionally, if an individual is highly motivated to perform well, their drive might enhance focus and limit interference from competing processes, whereas if they were less motivated, their emotional state might have a greater impact on performance. This reflects the complexity of factors – and the relationships between them – that can impact functioning in “real world” settings. Future studies should include a more complete representation of factors likely to influence emotion regulation and executive functioning.

Limitations

This study is one of the first to use an experimental paradigm to investigate the associations between executive function and emotion regulation. As a preliminary investigation, it has several limitations – future work is necessary to replicate and expand upon our findings. First, our small sample size limits the generalizability of the findings. Second, this was the first study to use our selected mood induction videos; future studies using mood induction stimuli that have been previously validated will add support to our findings. Related, although our mood induction videos were chosen to reflect universal human experiences, they may not have elicited responses of the same valence and intensity in all participants, which would have implications for the emotion regulation task, including how easy or difficult regulating one’s emotional state was. In particular, the induction of positive mood in the PositiveMaintain condition was only partially successful. Our selected video, a compilation of people announcing pregnancies, may have only resonated with a subset of individuals, and could have triggered negative emotions in some people. As such, future studies should use positive mood inducing content that has a demonstrated ability to elicit positive moods universally.

Additionally, although our utilization of a spontaneous regulatory approach improves the ecological validity of this study, we were not able to control for individual differences in emotion regulation strategies. While our data showed that participants were equally effective at regulating emotion during the experiment, regardless of whether they reported using one or more than one regulatory strategy, we do not know what type(s) of emotion regulation strategies were employed by participants. There is evidence that regulatory strategies may be differentially associated with the facets of executive functioning (Pruessner et al., 2020). Braunstein, Gross, and Ochsner (2017) reported that emotion regulation strategies can be either explicit or implicit and the amount of cognitive effort used in the implementation of these strategies can differ. Thus, depending on the effectiveness of the mood induction, baseline mood, and participants’ choice of regulatory strategies, the cognitive burden of emotion regulation and, consequently, impact on CTST performance, could vary.

Future Directions

As both mood and the emotion regulation strategies implemented might have independent effects on executive function, we recommend future studies employ a research design that allows the examination of these effects separately. For instance, future studies could compare executive function performance between groups with induced mood but without emotion regulation, emotion regulation without induced mood, and induced mood and emotion regulation to better understand the independent and additive effects of mood and emotion regulation on executive functioning.

Future studies should also consider ways to eliminate practice effects that might influence participants’ performance on an executive functioning task. Falleti, Maruff, Collie, and Darby (2006) showed that practice effect on cognitive testing is more apparent between the first and second round of assessment and individual’s performance tend to stabilize with no further improvement after the second round. Thus, one possible way of eliminating practice effect is to have the participants perform several practice trials of the task before responding the actual task.

Conclusion

The present study examined how simultaneous emotion regulation impacts executive functioning. We provide preliminary evidence that participants’ inhibitory control may be more heavily recruited when they were regulating their emotions, causing some interference, whereas their set-shifting ability may remain relatively intact during the emotion regulation process. When an individual has a specific emotion regulation goal to achieve, his/her inhibitory control capacity might be required to control the emotional responses so that it is consistent with the regulatory goal. Attentional shifting ability might play a more important role as the individual attends to stimuli to generate an emotion. As one of the few experimental investigations, our work adds to the literature describing the impact of emotion regulation on executive functioning. Future studies should experimentally investigate the relationship between emotion regulation and executive function while addressing limitations of the present investigation including small sample size, consistency of the induced mood, and practice effects.

Data Availability Statement

The data that support the findings of this study are openly available in OSF Storage at https://osf.io/w7bc2/