The mindful gaze: trait mindful people under an instructed emotion regulation goal selectively attend to positive stimuli

ABSTRACT Trait mindfulness confers emotional benefits and encourages skillful emotion regulation, in part because it helps people more deliberately attend to internal experiences and external surroundings. Such heightened attentional control might help skillfully deploy one’s attention towards certain kinds of stimuli, which may in turn help regulate emotions, but this remains unknown. Testing how trait mindful people deploy attention when regulating their emotions could help uncover the specific mechanisms of mindfulness that confer its emotional benefits. The present study aimed to determine whether high trait mindfulness is associated with sustained attention biases to (i.e. longer gaze at) emotional scenes, when all participants are given the emotion regulation goal of staying in a positive mood. To measure this, we used eye tracking to assess selective attention to positive, neutral, and negative photographs. Higher trait mindfulness was associated with both a stronger attention bias for positive (vs. neutral and vs. negative) images, as well as greater success staying in a positive mood during viewing. Surprisingly, this attention bias towards the positive images did not mediate the relationship between mindfulness and maintenance of positive mood. Future work should compare visual attention to other emotion regulation strategies that may maximise positive affect for mindful people.

Mindfulness is characterised by maintaining focus on one's experiences as they unfold in the present moment and by a curious, nonjudgmental stance in response to what is unfolding (Quaglia et al., 2015).It a state that can be practiced, for example via mindfulness meditation, and as a trait wherein some people are predisposed to be more mindful in daily life (Kiken et al., 2015).This way of livingwhich emphasises a sensitive awareness of what is occurring in the momentis the opposite of moving through the day on "autopilot".
Mindfulness confers many benefits and is theorised to foster emotion regulation.High trait mindful people report greater happiness, lower overall anxiety and sadness, and a more stable mood (Brown & Ryan, 2003;Brown et al., 2007;Cash & Whittingham, 2010).Mindfulness is also relevant to clinical disorders, as it works directly against the automaticity of certain thoughts in psychopathology (Brown et al., 2007).
Despite substantial research on mindfulness, there is still ambiguity surrounding its connection to visual attention to one's environment.This is surprising, as attention is such a core feature of mindfulness.However, little is known about what mindful people visually attend to in their environment, particularly regarding emotional stimuli.Perhaps mindful people because of their attentional predispositionconsume a different "diet" of visual information as they look around the world.And perhaps, since mindful people have better attentional control, they have an easier time using attention-redirection as an emotional regulation strategy.

Attention and emotion
The ability to effectively direct one's attention is an important skill for regulating emotions.Visual attention selects content for further processing; it has correspondingly been referred to as the gateway to conscious awareness (Raila et al., 2015).Attentional biases to visual information are a critical component of emotion, and one can shape their emotional experience by deploying their attention to certain content (MacLeod et al., 2002;Sheppes et al., 2011).This is the core of why attention is regarded as a fundamental, early-stage emotion regulation strategy (Gross, 2008).Emotion regulation refers to the many strategies used to amplify, reduce, or maintain an emotion.It is a vital component of mental health, and difficulties with it are associated with various forms of psychopathology (Gross, 2008;Sheppes et al., 2015).Attention's role as an emotion regulation strategy reflects its importance in determining our daily emotional experience.

Trait mindfulness
The present study investigates correlates of trait mindfulness, or how one naturally attends to the world in a mindful way.It varies naturally across individuals andthough partially innatecan also be improved with mindfulness training (Brown et al., 2007;Quaglia et al., 2015).The trait Mindful Attention Awareness Scale (MAAS; Brown & Ryan, 2003) is frequently used to study mindfulness.Higher scores are correlated with a variety of emotional benefits including less emotional disturbance, lower negative affect and neuroticism, greater life satisfaction, and positive affect (Brown & Ryan, 2003;Quaglia et al., 2015;Wenzel et al., 2015).
Of the two main components of mindfulness, the MAAS is generally believed to capture the presentfocused attention component (i.e.attending to the here and now) in particular, rather than the nonjudgmental stance component (i.e.neutral rather than valenced reactions to what is unfolding; Van Dam et al., 2010).Present focus may relate more closely to heightened attentional control (Malinowski, 2013), and whether present focus relates to the content people visually attend to remains an intriguing question.

Mindfulness and attention
Attentional control is a person's ability to choose what they pay attention to or ignorein other words, the ability to concentrate (Loeffler et al., 2019).This ability, and specifically the ability to control attention to emotional stimuli, is perhaps unsurprisingly associated with the ability to regulate emotions (Feng et al., 2018;Loeffler et al., 2019).That is, people with better attentional control also seem to be better at regulating their emotions.Attentional control has been closely linked to mindfulness (Jha et al., 2007): mindfulness training improves performance of attentional control (Chambers et al., 2008), and higher trait mindfulness predicts greater attentional control (Walsh et al., 2009).
While trait mindfulness is associated with improved attentional control, less is known about whether mindfulness is associated with differences in what one attends to in the worldi.e.attention to different information in the environment.Mindfulness-based practices improve people's ability to direct their attention to stimuli deemed more important in the moment (Jha et al., 2007), and thus when people are aiming to stay in a positive mood, being mindful may enable them to direct attention to information that serves that purpose.Coupled with findings that mindfulness interventions reduce attention's negativity and general emotionality bias (Roca & Vazquez, 2020) and reduce attentional capture by emotional stimuli (Eddy et al., 2015), mindful people may deploy attention to emotional stimuli congruent with their emotion regulation goal.
One way of measuring sustained attention is eye tracking, which captures how long people dwell on certain information.Two eye tracking studies found that experienced mindful meditation practice reduced attentional biases to negatively-valenced faces (Blanco et al., 2020;Pavlov et al., 2015).Most relevant to the present study, two other recent eye tracking studies examined associations with trait mindfulness and found conflicting resultswith one finding it positively correlated with viewing happy and threatening scenes, but not correlated with viewing sad scenes (Ford et al., 2021), and another finding it negatively correlated with viewing sad and threatening scenes, but not correlated with viewing happy scenes (Kraines et al., 2021).It should be noted that participants in these studies were not necessarily trying to regulate their emotions during image viewing.Overall, the state of evidence is far from clear, and better understanding of what mindful people attend to would help understand how mindfulness facilitates emotional benefits.

Maximising positive affect with the motivated gaze
Like any emotion regulation strategy, attention may be deployed differently depending on one's emotion goals at the time (Millgram et al., 2019).People often have a goal of experiencing a positive mood, and we can test how emotion regulation strategies function during this goal (Xing & Isaacowitz, 2006).Without such instruction, individuals' emotional goals can vary widely (Sheppes et al., 2011), and it is difficult to parse apart whether differences in attention deployment are due to different emotion regulation approaches or simply different emotional goals at the time.
Previous work on visual attention has shown that gaze is motivated by goals of maintaining positive affect (the "motivated gaze;" Isaacowitz, 2006).For example, participants instructed to focus on managing their emotions gaze less at negative scenes than those told simply to gather information from the scenes (Xing & Isaacowitz, 2006).And, older people motivated to regulate their emotions in a positive direction show gaze preferences for positive stimuli (Isaacowitz et al., 2006).Indeed, a more negative mood in older adults, and thus ostensibly a greater motivation to regulate emotions, is associated with a stronger positivity bias (Demeyer et al., 2017).Accordingly, given mindful people's awareness of the present moment and strong attentional control, they may be better able to use their gaze in service of maximising positive affect.

The current study
The present study tests whether trait mindful people differentially gaze at emotional content when instructed to maintain a positive mood and if they are in a more positive mood after image viewing.To examine this, we measured selective visual attention to emotional scenes during instructed emotion regulation.Healthy adults viewed image pairs (Positive-Neutral, Positive-Negative, and Negative-Neutral) while their eye movements were tracked.Importantly, all participants were instructed to try to stay in a positive mood during image viewing.Each participant's attention bias for Positive, Neutral, and Negative images on the paired-image trials was computed and correlated with trait mindfulness (measured by the MAAS) while controlling for state affect and the images' low-level visual features.To test whether mindful people more effectively maintained a positive mood during image viewing, post-viewing mood (measured by the mDES) was associated with gaze biases and trait mindfulness.
This study aims to extend previous literature in two important ways.First, to our knowledge, it is the first investigation of the relationship between trait mindfulness and visual attention when all participants were instructed with a shared emotion goal: to maintain a positive mood.By motivating them to maintain a positive mood, participants could then vary in selected strategy, with some using visual attention as a strategy more so than others.Previous research measured how gaze is associated with trait mindfulness during passive viewing and found mixed results (Ford et al., 2021;Kraines et al., 2021).It is important to assess how gaze is allotted under certain emotion regulation goals, because emotion goals impact attention biases (Xing & Isaacowitz, 2006), and studies have thus sometimes used verbal instruction to create shared emotion goals across participants and then measure gaze (van Reekum et al., 2007).Mindfulness practice reduces attentional capture by emotional stimuli (Eddy et al., 2015;Roca & Vazquez, 2020), which possibly makes way for greater influence of the emotion goals of the individual, and accordingly, this study induces a common emotion goal.
Second, this study extends two previous studies of eye tracking to emotional images (Ford et al., 2021;Kraines et al., 2021) by ensuring that low-level visual features (e.g.contrast, luminance) are controlled for.Existing studies (Raila et al., 2015) have found such visual features of emotional images to be confounds, and accounting for them may help uncover why previous studies correlating gaze with trait mindfulness have found conflicting results.Through both of the above, this study aims to clarify existing mixed results about connections between gaze and trait mindfulness.

Participants
103 young adults (primarily from a university in the northeastern United States) participated.Seventy-nine (76.7%) completed the study in exchange for course credit; the remaining 24 were compensated $10.Two participants were excluded from the analysis due to equipment failure, resulting in a final sample of 101 participants.Sample size was based on an average of typical samples in previous eye tracking studies linking visual attention to trait emotion measures (e.g.Quigley et al., 2012;Raila et al., 2015).Of the sample of 101, ages ranged from 18 to 27 (M = 19.49;SD = 1.50), and 62 (61%) were female.Selfreported ethnicities were: 41.6% Caucasian/White, 25.7% Asian-American/Asian, 7.9% African-American/ Black, 10.9% Latino/Hispanic, 8.9% multiracial, and 5.0% other.Nine of these 101 subjects had unreliable eye-tracking data (defined as less than 80% of eye data being successfully recorded in the viewing area on more than one-third of the trials; Raila et al., 2015) and were therefore excluded from analyses that included gaze measures.Two others did not complete all self-report measures, and they were therefore excluded from analyses that included the final affect measure.Thus, due to listwise deletion, analyses involving measures of affect but not gaze include 99 subjects, analyses involving measures of gaze but not affect include 92 subjects, and analyses involving both types of measures include 90 subjects.

Stimuli
Eye gaze was tracked as participants viewed images.For eye tracking apparatus details, see supplemental material.On each of 36 experimental trials, participants viewed two images.Twelve trials consisted of a Positive image (e.g. an Olympic medal) paired with a Neutral image (e.g.file cabinets), 12 consisted of a Negative image (e.g. a pile of human bones) paired with a Neutral image, and 12 consisted of a Positive image paired with a Negative image.Each trial had unique images, for a total of 72 experimental images.Two images at a time were presented in counterbalanced locations on a dark grey background (see Figure 1).For details on image placement in eye tracking trials, see supplemental material.Prior to their experimental administration, all images were validated on valence and arousal ratings; for details on validation, see supplemental material.
A phase-scrambled control image was generated for each experimental image.For details on phase scrambling procedure, see supplemental material.When presented during the experiment (see supplementary Figure S1), each phase-scrambled image was always presented in the same position as its experimental counterpart.Thus, they viewed a full set of 36 matched control trials.Overall, each participant viewed 144 total images (72 experimental; 72 phase-scrambled control) across 72 trials (36 experimental; 36 phase-scrambled control).

Measures
Participants completed two measures: the Trait Mindful Attention Awareness Scale (MAAS; Brown & Ryan, 2003) used to capture trait mindfulness, and the modified Discrete Emotion Scale (mDES; Fredrickson et al., 2003) used to capture positive and negative affect.For details on these measures, see Supplemental material.

Procedure
Participants were given a heart rate monitor chest strap and told that it would monitor their heart rate and breathing during the study.They were instructed to "try to keep as calm, content, and happy as possible" while viewing the images.They were told that physiological measurements would indicate their mood, though no heart rate or breathing data was actually collected.The inclusion of a sham heart rate monitor was to further motivate this instruction, as participants were told that we could use it to estimate whether they had achieved this emotional goal.The baseline mood measurement was given after the emotion goal instruction, because we did not want mood change caused merely by the instruction alone to be captured within this measure.We were interested in the mood change connected to the task, while the instruction was to create a shared emotional goal that may influence cognition and behaviour during the task, attempting to control for peoples' otherwise widely differing emotional goals (Sheppes et al., 2011).
Next, participants completed the mDES (Time 1) at a desktop computer.They were then brought to a separate room and seated in front of the eye tracker.They were instructed to view the screen as if they were at home watching television and asked not to wander their gaze around the room.To avoid demand characteristics, we aimed to disguise the study purpose by including a sham task and stating a false study purpose: participants were told that the study was to determine whether eye movements differ between reading an article, viewing simple shapes, and viewing more complex visual scenes.After these instructions, the experimenter closed the door and turned off the lights.The eye tracker was calibrated to each participant using five-point calibration dots.As part of a concomitant study and the sham task, participants first read a passage (for details see supplemental material), and then they completed the sham shape task in which they were shown a series of five slides, each with a different pair of shapes on it.
Next, during the image viewing, participants were shown a series of image slides lasting about nine minutes total.Each slide included a pair of affective images (e.g.Positive-Neutral) or phase-scrambled control images, each shown for 6 s.Between each image slide, a slide containing only a fixation cross at the center of the grey background was shown for 1 s.Slide order was randomised for each participant.Once finished with the viewing task, participants completed the second round of self-report measures: the mDES (post-viewing) and MAAS.

Data reduction and analyses
For eye-tracking analysis, the two areas of interest (AOIs) on each trial were defined by the boundaries of the two images.Dwell Time was defined as the total time spent fixating and making saccades within a given AOI.
To capture Gaze Biases for one emotional image type relative to the other, the outcome measure of Dwell Time was calculated based on a "difference score" between the two images.For more details on data reduction and analysis, see Supplemental material  images, for a total of three matched pairs of bias scores, or six total scorescomputed for each participant.Reliability of these eye tracking measures was generally good; for details, see Supplemental material.

Overall gaze preferences
Our first step was to calculate the overall Gaze Biases across our sample.We were interested in attention bias driven by the semantic content of the images, rather than by their low-level visual properties (e.g.luminance), so we removed the variance associated with the scrambled images by computing difference scores between the intact and phase-scrambled images for each participant (see Raila et al., 2015).For example, the Positive-Neutral Bias for a participant is computed by subtracting their Positive-Neutral Bias for phase-scrambled images (itself a difference score, explained above) from their Positive-Neutral Bias for intact images.This directly subtracts out effects of the low-level visual properties of the images and determines the amount of baseline Gaze Bias driven by semantic content.
These difference-of-differences measuresshown in Supplementary Table S1 summarise the looking behaviour for the sample as a whole.One-sample ttests were conducted to determine whether the biases were significantly different from zero.The sample overall showed a Positive Bias on both Positive-Neutral Trials and Positive-Negative Trials, and a Neutral Bias on Negative-Neutral Trials.In other words, participants spent much more time viewing Positive images and far less time viewing Negative images across our sample.
To ensure that the groups who had read three different kinds of passages did not differ in our key measures of interest, we conducted one-way, between subjects ANOVAs to compare the groups.There were no significant differences between the groups on either self-report measures of interest or gaze variables, Fs < 2.11, ps > .127.For details and exact statistics, see Supplemental material.

Trait mindfulness is associated with a gaze bias for positive images
Primary analyses below employ the difference scores described in Method, then statistically control for gaze contributed by the low-level visual features in order to isolate the effects of the semantic content per se.
To determine the relationship between trait mindfulness and selective attention to emotional images during instructed emotion regulation, we conducted three multiple linear regressions predicting each of the three Gaze Biases as dependent variables.We tested whether MAAS score was a predictor, as well as baseline mDES PA and NA (controlling for baseline mood aligns with previous studies that aim to isolate the influence of affect-related traits; Raila et al., 2015) and the corresponding Scrambled Image Bias.
In predicting the Positive-Negative Bias, the contribution of MAAS while controlling for all other factors was significant, β = .22,p = .044.The results of this regression are summarised in Table 1.The zeroorder correlation between MAAS and Positive-Negative Bias is also significant without including covariates, r = .22,p = .035.In predicting the Positive-Neutral Bias, the contribution of MAAS while controlling for all other factors was also significant, β = .26,p = .012.The results of this regression are summarised in Supplementary Table S3.The zero-order correlation between MAAS and Positive-Neutral Bias is also significant without including covariates, r = .25,p = .016.The strength of the partial correlations between MAAS and these Gaze Biases is captured in Supplementary Figures S2 and S3.In predicting the Negative-Neutral Bias, however, the contribution of MAAS was not significant, β = -.19 p = .071,although this result was in a direction that would indicate higher trait mindfulness predicting gaze preference for neutral scenes over negative scenes.Taken together, these results demonstrate that trait mindful people at least during instructed emotion regulationgaze longer at positive scenes (over neutral or negative scenes).

Trait mindfulness predicts higher positive affect post-viewing
To test the relationship between trait mindfulness and mood during instructed emotion regulation, we ran multiple linear regressions to predict post-viewing mDES scores (both positive affect [PA] and negative affect [NA]) as dependent variables.First, we looked at MAAS as a predictor of post-viewing mDES PAwith baseline mDES PA and NA also included as predictorswhich revealed a significant independent contribution of MAAS, β = .15,p = .050.This regression model is summarised in Supplementary Table S2, and the strength of the partial correlation between MAAS and post-viewing PA is captured in Supplementary Figure S3.Second, we looked at MAAS as a predictor of post-viewing mDES NAwith the same set of predictorswhich found that the independent contribution of MAAS was not significant, β = -.05,p = .548.Overall results indicate that, during instructed emotion regulation, higher trait mindfulness predicts a more positive mood after viewing emotional images.

No relationship between looking and mood
To assess whether Gaze Biases predicted post-viewing mood, six total partial correlations were conducted between each of the three Gaze Biases (Positive-Neutral, Positive-Negative, and Negative-Neutral) and post-viewing mDES PA and NA, while controlling for baseline mDES PA and NA and the corresponding Scrambled Image Bias.No significant correlations were found, rs < |.18|, ps > .113(for exact statistics, please see Supplemental materials), such that unexpectedly, Gaze Biases did not predict post-viewing mood.
Although this result generally precludes a mediation analysis, to be thorough, we next investigated the possibility that Gaze Biases mediated the relationship between mindfulness and post-viewing PA.For details, see Supplemental material.Overall, results indicate that Gaze Biases towards positive information did not mediate the relationship between mindfulness and post-viewing PA; that is, there was no indirect effect of mindfulness on mood by way of Gaze Bias.Rather, mindfulness separately predicts positive gaze biases and post-viewing PA.

Trait mindfulness interacts with gaze biases in relationship with post-viewing mood
To test whether trait mindfulness interacted with the gaze biases in impacting post-viewing mood, we explored possible interactions between MAAS and each of the three Gaze Biases on the two dependent variables of post-viewing PA and NA.To do so, we conducted six multiple linear regression models.Each model had one of the interaction terms that would be relevant to our theory (e.g.MAAS x Positive-Negative gaze), as well as each of these as separate predictors, the corresponding Scrambled Gaze Bias, and baseline PA and NA.
Of these, there were two significant interactions on the dependent variable of post-viewing PA: between mindfulness and Negative-Neutral Gaze, β = 0.91, t = 2.12, p = .022,and between mindfulness and Positive-Negative Gaze, β = −1.13,t = −2.16,p = .034.No other interactions were significant, βs > 0.44, ts < 1.51, ps > .134.That is, for participants who gazed longer at the relatively negative image in a pair, their mindfulness mattered more for predicting their good mood.Put conversely, for those who are more mindful, their viewing bias mattered less for predicting good mood after viewing.

Discussion
Results indicate that trait mindful peopleat least during instructed emotion regulationattend more to positive (vs.neutral and negative) images and are in a happier mood following image viewing.There was no evidence, however, that their selective attention mediated this relationship between trait mindfulness and happier mood.The lack of a mediation suggests mindful people may maintain their positive mood through another emotion regulation strategy outside of selective visual attention.
Our sample's overall gaze preferences under instructed emotion regulationa bias toward positive images (vs.neutral and negative) and away from negative images (vs.neutral)support previous findings.Experimental manipulations have shown that instructed emotion regulation, compared to passive viewing, decreases attention to negative images (Xing & Isaacowitz, 2006).Here, when all participants were under instructed emotion regulation, they demonstrated an attention bias for positive stimuli (though they were not compared to a control).This makes sense, and it suggests people may accurately intuit selective attention as an effective means of emotion regulation, since increasing attention to positive information does boost positive affect (Grafton et al., 2012;Wadlinger & Isaacowitz, 2011).Those who regularly consume a more positive "diet" of visual information may indeed be happier people (Raila et al., 2015).

Trait mindful people attend to positive info during instructed to emotion regulation
To our knowledge, this is the first study demonstrating a link between trait mindfulness and attention bias to positive stimuli during instructed emotion regulation.Importantly, this relationship was driven by the semantic content of the images rather than their low-level visual features and was independent of state affect.
Previous eye-tracking research has demonstrated links between a positivity bias during naturalistic picture viewing and trait constructs such as happiness, optimism, and old age (Raila et al., 2015;Isaacowitz, 2005;Isaacowitz et al., 2006).The present study adds to this literature by finding that positivity biases were also correlated with trait mindfulness.Studies with higher power may also find that biases in the direction away from negative stimuli (in our study, a trend) are significantly correlated with trait mindfulness, as found in a previous eye tracking study (Kraines et al., 2021).
Though correlational, this finding may reflect the greater capability of mindful people to control their attention to emotional stimuli.People in general alter what they look at when they are trying to feel happy (Xing & Isaacowitz, 2006), and more mindful people may be better equipped to sustain their attention towards certain information as a way to manage their emotions.The when they are trying to feel happy component is important, as people may alter their looking behaviour differentially depending on their emotion regulation goal at the time; this was the primary study design motivation of instructing all of our participants to have the shared emotion regulation goal.The shared emotion regulation goal may partially explain some contradictory findings between our study and the two previous studies to look at trait mindfulness and eye tracking (i.e.Ford et al., 2021;Kraines et al., 2021), as the participants in those previous studies may have varied in their emotion goals in the lab at the time.
This study makes a methodological contribution by testing attentional biases that partial out the influence of images' low-level visual features to isolate influence due to the images' semantic content.Previous work (Raila et al., 2015) has highlighted that neglecting this type of methodological approach could result in misattributing effects to semantic categories that are really driven by confounded low-level visual features.For example, previous findings in the opposite directione.g. that trait mindfulness is connected to biases towards negative information (Ford et al., 2021) may be driven by confounded low-level features.
Research has linked attentional control to both resilience to and treatment for a variety of internalising disorders (e.g.anxiety: Shi et al., 2019) and to emotion regulation ability (Feng et al., 2018;Loeffler et al., 2019), and attentional control may be a key mechanism through which mindfulness-based treatments work (Chambers et al., 2008).Present results add to this literature by indicating that mindfulness predicts what we pay attention to during emotion regulation, which sheds light on the nature of mindfulness' known clinical benefits.

Do trait mindful people use selective attention to maximise positive affect?
Trait mindfulness, in addition to predicting attention towards positive information, also predicted greater positive affect after image viewing.This suggests that mindful people are better at upregulating their mood when instructed to do so while viewing emotional scenes.Surprisingly, there was no relationship between the attention biases and post-viewing positive affect (which opposes previous findings wherein attention biases for positive stimuli have predicted increased positive affect; Sanchez et al., 2014).
This establishes an intriguing scenario: mindful people spent more time gazing at the positive imageswhich could indicate mindful people's ability to exert more purposeful attentional control (Garland et al., 2015) and then were happier after viewing the images, yet the relationship between their mindfulness and their ability to stay in a good mood was not mediated by what they were attending to.These results were unexpected, and they provide initial evidence that the benefits to positive affect conferred by mindfulness may be less related to visual attention, and more to another strategy of emotion regulation, such as cognitive change (Gross, 2008), that then keeps them in a good mood throughout the task.In sum, the mechanism linking trait mindfulness to improved positive affect here likely involves the utilisation of more than just attention to positive stimuli.
This lack of connection between what participants looked at and how they felt afterwards does align with previous findings that gaze biases are consistently linked with trait affect (Raila et al., 2015;Sanchez & Vazquez, 2014), while being indirectly linked with state affect.One study found that while both state affect and trait-level well-being predicted positive gaze bias, the effect of state affect in the model was fully mediated by trait-level well-being, reflecting that long-term affect had a more direct influence than state affect (Blanco & Vazquez, 2021).Accordingly, our failure to find gaze as a mediator influencing state affectalbeit unexpectedis perhaps not surprising.Future studies may test whether, instead, gaze biases mediate a relationship between trait mindfulness and trait happiness.
Interestingly, however, there was an interaction between mindfulness and what participants looked at to predict post-viewing mood.Interpreting these results, perhaps more mindful peopleable to stay focused on the present momentare better able to attend to negative information without impacting their mood (even though they overall spend more time gazing at positive information).Meanwhile, less mindful people, whose minds tend to drift away from the present, may be more in need of the comforting benefits of a positive viewing bias.

Limitations
There are several limitations to this study.First, we do not know the degree of effort each subject put in to regulate their emotions and stay in a positive mood.Participants who wanted to be "good subjects" may have been motivated by the instructions to look at certain stimuli, but not because they were trying to feel internally better.Similarly, we acknowledge that because participants' emotional goals in this study originated from an external source (i.e.task instructions) rather than an internal source (i.e.intrinsically wanting to feel better), this could alter the use of cognitive resources to meet said goals.Future studies may wish to assess the degree of motivation, and self-perceived accomplishment in the attempt to regulate emotions, as other outcomes or covariates.
Second, the present study did not have a control condition to compare instructed emotion regulation vs. a no-instruction condition.Therefore, it remains unknown whether the observed viewing behaviours associated with trait mindfulness occur only under instructed emotion regulation.Future studies may wish to include such a control group.
Third, the MAAS captures one of the fundamental components of mindfulness, present focus, but researchers have argued that it does not capture the second one: nonjudgmental stance (Van Dam et al., 2010).Different kinds of attention biases may correlate with each of these two major components.For example, participants high on the nonjudgmental stance might equally attend to positive, negative, and neutral images, given their less valenced reaction to emotional stimuli.This could explain some differences from the two previous eye tracking studies, which used trait mindfulness questionnaires that also capture the nonjudgmental stance (Ford et al., 2021;Kraines et al., 2021).Future studies may wish to separately examine attention biases associated with present focus and nonjudgmental stance.
Fourth, we naturally cannot control for all potential confounding variables.These include depressive and anxiety symptoms, which may be connected to both mindfulness and mood regulation, and previous experience with mindfulness practice.Participants also likely vary in their use of higher-level emotional regulation strategies such as cognitive reappraisal, which could have been evoked by the instructed emotion goal of staying in a positive mood.As noted earlier, our measure of mindfulness (the MAAS) captures the present-focused attention of mindfulness, while we did not test the nonjudgmental open-minded component.

Future directions
The results of the present study established that trait mindfulness is correlated with an attention bias for positive images, as well as with higher positive affect after viewing, during instructed emotion regulation.Attention biases did not mediate the relationship between mindfulness and mood, and future studies may wish to employ a larger sample size to detect subtler effectsor a clinical sample (e.g.depression) to examine this in people with a wider range in their emotions.
Naturally, this correlational study cannot determine causality.On the one hand, navigating the world focused on the present moment could encourage one to consume more positive information in any given moment.Conversely, attention to positive stimuli could help maintain trait mindfulness if this "diet" of positive information enables focus on the present.Experimental research using brief inductions of mindfulness could explore whether mindfulness causes changes in attention deployment during emotion regulation.Longitudinal data could also evaluate causal relationships between mindfulness and attention to emotional information.
This study is the first look at how trait mindful people allocate their attention during instructed emotion regulation.It contributes to deciphering the mechanisms through which mindful people maximise positive affect, and next steps include determining what other emotion regulatory processes beyond attention allocation may be at play in maintaining this connection.

Disclosure statement
No potential conflict of interest was reported by the author(s).
. To illustrate in plain terms, a Positivity bias of +13 on the Positive-Neutral slides would indicate that the participant spent, on average across the Positive-Neutral slides, 13% more of the trial time gazing at Positive images than at Neutral images.Each of these three Gaze Bias Scores (Positive on Positive-Neutral, Positive on Positive-Negative, and Negative on Negative-Neutral) was computed in the same manner for the matched, phase-scrambled control

Figure 1 .
Figure 1.An example of a positive-neutral experimental trial (i.e. with intact images).

Table 1 .
Regression coefficients for predicting the dependent variable of positive-negative gaze bias.= 90.R 2 = .073,p = .155.B = unstandardized regression coefficient; CI = confidence interval of B; SE = standard error of B; β = standardized regression coefficient.This model includes MAAS score, corresponding Scrambled Positive-Negative Image Bias Score, baseline mDES PA and NA, and Group.* indicates p < .05.