Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 6.
Published in final edited form as: Psychol Health. 2020 Jun 13;36(4):444–460. doi: 10.1080/08870446.2020.1778696

Same-day, cross-day, and upward spiral relations between positive affect and positive health behaviours

Barbara L Fredrickson 1, Cara Arizmendi 2, Patty Van Cappellen 3
PMCID: PMC9536170  NIHMSID: NIHMS1618345  PMID: 32538212

Abstract

Objective:

This project investigated same-day and lagged (i.e., from one day to the next) associations between daily positive affect and three distinct positive health behaviours: physical activity, fruit and vegetable intake, and meditation. Cross-day analyses also examined the role of positive affect felt during the targeted health behaviours.

Design:

Secondary data analyses used a 9-week daily diary study in which midlife adults (N = 217) were randomized to learn one of two contemplative practices (i.e., mindfulness meditation or loving-kindness meditation) while reporting nightly on their emotions and health behaviours.

Results:

Results of same-day analyses revealed positive associations, both between-person and within-person, for the three positive health behaviours with daily positive affect. Results of lagged analyses revealed that positive affect experienced during fruit and vegetable intake on a given day predicted next-day fruit and vegetable intake, and that fruit and vegetable intake on a given day predicted next-day positive affect.

Conclusion:

The observed same-day relations between daily positive affect and engagement in positive health behaviours illuminate one path through which positive affect may contribute to health. The observed cross-day relations reveal a need for interdisciplinary research on mechanisms through which fruit and vegetable intake may shape next-day positive affect.

Keywords: positive psychology, nutrition, physical fitness, contemplative science, upward spiral theory of lifestyle change

Habitual experiences of positive affect, meta-analyses show, are consistently linked to longevity (Chida & Steptoe, 2008; Zhang & Han, 2016). Multiple mechanisms, likely operating in conjunction, have been identified to account for this link. These include (a) salutogenic biological profiles (Kok et al., 2013); Ong, Benson, Zautra, & Ram, 2018); Wilson et al., 2017); (b) stress-buffering (Blevins, Sagui, & Bennett, 2017); Okely, Weiss, & Gale, 2017); Robles, Brooks, & Pressman, 2009); (c) resource-building (Cohn, Fredrickson, Brown, Mikels, & Conway, 2009); Fredrickson, Cohn, Coffey, Pek, & Finkel, 2008); Ouweneel, Blanc, & Schaufeli, 2012); and (d) health behaviours (Ironson, Banerjee, Fitch, & Krause, 2018); Kubzansky & Thurston, 2007), each known to be associated with positive affect. The current investigation targets this last proposed mechanism, health behaviours, and specifically examines the links between positive affect and the subset of health behaviours that, when enacted, enhance physical health, collectively known as “positive health behaviours.” The behaviours we examine include physical activity, fruit and vegetable intake, and meditation. Although each promotes wellness differently, fits into people’s lives differently, requires different skills, and may contribute differently to people’s sense of meaning and identity, we target this set because health-promoting behaviours tend to cluster (Lippke, Nigg, & Maddock, 2012).

Two forms of positive affect are likely to be related to positive health behaviours: enjoyment of the day and enjoyment of the specific health-related activity. We assessed the first form, enjoyment of the day (a.k.a., “daily positive affect”), as an aggregate of ten different positive emotions endorsed in end-of-day reports (e.g., peak experiences of joy, interest, gratitude, or love on that day). Because positive affective valence has been demonstrated to increase agency (Chang, Algoe, & Chen, 2017), ambient positive affect may implicitly energize activity engagement (e.g., Cameron, Bertenshaw, & Sheeran, 2018). An alternative direction of causality may run from positive health behaviours to the enjoyment of life: Those who possess personal resources such as physical health and resilience should be better equipped to manage ups and downs in daily living, which can make days more enjoyable (Cohn et al., 2009). Either direction of causality predicts same-day associations between daily positive affect and positive health behaviours. Examining cross-day associations offers one possibility to disentangle the two causal directions.

The second form of positive affect we investigated is enjoyment of the activity, assessed as positive feelings during specific bouts of a positive health behaviour (e.g., “How positive did you feel, on average, while you were physically active [today]?” reported retrospectively at end-of-day). The affective rewards of positive health behaviours have been found to predict long-term behavioural maintenance (Cohn & Fredrickson, 2010)) Williams, Dunsiger, Jennings, & Marcus, 2012). Experimental studies show that positive affect felt during a given activity creates implicit motives for that activity that augment subsequent behavioural effort (Aarts, Custers, & Marien, 2008; Custers & Aarts, 2005;) Rice & Fredrickson, 2017). Repeated enjoyment of an activity can also build personal resources, such as physical fitness or social integration, which may subsequently amplify later enjoyment of the activity (Rice, Adair, Tepper, & Fredrickson, 2019). These two pathways linking enjoyment of the activity to sustained engagement are central to the Upward Spiral Theory of Lifestyle Change (Fredrickson, 2013); Van Cappellen et al., 2018). Although tests of these mechanisms are beyond the scope of the current paper, this theorizing drew us to examine enjoyment of specific positive health behaviours alongside enjoyment of the day.

Among the positive health behaviours examined here, physical activity has been studied most, with within-day links to positive affect well-established (Reed & Ones, 2006). Physical activity produces transient boosts in positive affect both during physical activity (if below the ventilatory or lactate threshold; Ekkekakis, Parfitt, & Petruzzello, 2011), Emerson & Williams, 2015) as well as over the ensuing 4–5 hours (Ekkekakis et al., 2011;) Wichers et al., 2012;) for a meta-analysis, see Reed & Ones, 2006). Recent evidence also suggests within-day reciprocal associations: In a study using ecological momentary assessment with low-active, overweight adults (N = 59), beyond reporting more positive affect on days on which they exercised, participants were also more likely to exercise on days when they experienced more positive affect earlier in the day (Emerson, Dunsiger, & Williams, 2018). Additionally, a review of 24 studies concluded that positive affect experienced during physical activity forecasts people’s physical activity in the weeks and months ahead, whereas positive affect experienced after physical activity does not (Rhodes & Kates, 2015). It remains unknown, however, whether this conclusion also applies to shorter timespans, such as from one day to the next.

Although links between positive affect and fruit and vegetable intake have been less studied, a recent cross-sectional study of college students (N = 1270) found positive affect to increase linearly as a function of fruit and vegetable intake (Warner, Frye, Morrell, & Carey, 2017) see also Brookie, Best, & Conner, 2018). A 3-week daily diary study of young adults (N = 281) also found same-day associations between positive affect and fruit and vegetable intake, plus cross-day associations, with yesterday’s fruit and vegetable intake predicting today’s positive affect, and not vice versa (White, Horwath, & Conner, 2013, but see Conner, Brookie, Richardson, & Polak, 2015). A subsequent randomized controlled trial with young adults (N = 171) found that increasing fruit and vegetable intake raised participants’ reports of vitality and flourishing, but not positive mood (Conner, Brookie, Carr, Mainvil, & Vissers, 2017). No studies have yet examined whether positive affect felt during fruit and vegetable consumption (vs. that felt any time during the targeted day) also plays a role.

The third positive health behaviour we target is meditation because one of the most common reasons that people report for practicing meditation is to maintain their general health and well-being (Burke, Lam, Stussman, & Yang, 2017). Our team conducted two 9-week daily diary studies of midlife adults (total N = 339) who initiated a regular practice of either mindfulness meditation or loving-kindness meditation (Fredrickson et al., 2017; Study 2 of which reports on the same sample as used here). Both practices produced week-by-week gains in enjoyment of the day, indexed as the aforementioned aggregate of ten positive emotions. Multilevel models also revealed same-day associations with positive affect, both between-persons and within-persons, with the latter being stronger for loving-kindness meditation (Fredrickson et al., 2017). Cross-day effects and links with positive affect felt during meditation (vs. that felt any time during the targeted day) remain unknown.

Here, we conducted secondary data analyses of a 9-week daily diary study in which midlife adults (N = 217) were randomized to learn one of two contemplative practices (i.e., mindfulness meditation or loving-kindness meditation) while reporting nightly on their emotions and health behaviours (Fredrickson et al., 2017, Study 2). If a given health behaviour was reported, participants also reported how much, plus the affect they experienced during the behaviour. We examined within-day and lagged (i.e., over one day) associations between two forms of positive affect (i.e., across the day and during health behaviours) and the three positive health behaviours identified above. For comparison purposes, we also examined negative affect and one negative health behaviour, namely, alcohol consumption, which has also been linked to same-day positive affect and therefore is important to distinguish from behaviours that promote health (Bold et al., 2016).

Based on the research literature reviewed above, we hypothesized that same-day positive associations would exist between the three targeted positive health behaviours (i.e., physical activity, fruit and vegetable intake, and meditation) and enjoyment of the day. Moreover, we hypothesized that these predicted positive associations would emerge both between persons and within persons. To address unanswered research questions about cross-day associations between positive health behaviours and positive affect, we also conducted within-person lagged analyses using these same variables. These allowed us to investigate whether yesterday’s positive health behaviours predict today’s positive affect and vice versa. Lagged analyses also examined the role of enjoyment during the targeted health-related activity. Specifically, we examined the within-person, lagged, cross-day associations between the positive affect people retrospectively reported experiencing during the three positive health behaviours and their next-day engagement in those behaviours. Finally, for comparison purposes, we conducted similar same-day and cross-day analyses with daily negative affect, and also with the negative health behaviour of alcohol use.

Method

Participants

Participants were midlife adults aged 35 to 64 (Inclusion criteria: interest in lifestyle change and meditation; daily internet access; fluent in English. Exclusion criterion: established meditation practice.) Procedures for recruitment and screening have been described elsewhere (Fredrickson et al., 2017; Online Supplementary Material [OSM] lists prior publications from this NIH-supported study [R01CA170128]). Following informed consent, participants (N = 231) were randomized to one of two meditation workshops: mindfulness meditation (MM; n = 113) or loving-kindness meditation (LKM; n = 118). Ultimately, 14 participants were excluded, resulting in a final sample of N = 217 (for MM, n = 106; for LKM, n = 111; for details, see CONSORT diagram in OSM). In the analysis sample, mean age was 48.6 years (SD = 9.0). The majority of participants were female (59.9%) and Caucasian (76.5%), and 18.0% were Black. (For more demographic details, see Table 1 [Study 2] in Fredrickson et al., 2017).

Table 1.

Same-day associations between daily positive affect and health behaviours.

Fixed Effects
b β 95% CI p
Intercept 1.52** 1.82 [1.38,1.67] <.001
Phy_C .16** .07 [.13,.20] <.001
Phy_Avg .65** .17 [.31,.99] <.001
Time .02** .05 [.01,.03] <.001
−2LL 13500.88
b β 95% CI
Intercept 1.23** 1.82 [1.03,1.44] <.001
FVI_C .03** .04 [0.02,0.04] <.001
FVI_Avg .13** .25 [0.08,0.17] <.001
Time .02** .04 [0.01,0.02] <.001
−2LL 13816.73
b β 95% CI
Intercept 1.52** 1.82 [1.32,1.73] <.001
Med_C .02** .02 [.00,.03] .01
Med_Avg .15* .11 [.02,.28] .02
Time .02** .05 [.01,.03] <.001
−2LL 13771.95
b β 95% CI
Intercept 1.76** 1.82 [1.64,1.88] <.001
Alc_C .08** .07 [.06,.10] <.001
Alc_Avg −.02 −.02 [−.12,.08] .75
Time .02** .04 [.01,.03] <.001
−2LL 13970.17
Open in a new tab

Note. Phy = physical activity; FVI = fruit and vegetable intake; Med = meditation; Alc = alcohol consumption.

*

_C indicates person mean-centered variable, or within-person effects.

*

_Avg indicates individual means of variable, or between-person effects.

Given that the present research involved secondary analyses, a priori power analyses to determine optimal sample size were not possible (they were, however, conducted for the unrelated primary analyses). Post-hoc power analyses were conducted using the “sjstats” package in R 3.3.1 for small, medium, and large effect sizes. A very small effect size of 0.1 (Cohen’s D) with 217 subjects requires 50 observations per person to achieve power of 0.8. A medium effect size of 0.5 requires 50 individuals with three observations, and a large effect size of 0.8 requires 30 individuals with two observations each. We conclude that this investigation (with 63 assessment days, and a 79% mean response rate) is adequately powered to detect even very small effect sizes.

Procedure

All procedures were approved by the Institution Review Board of the University of North Carolina at Chapel Hill. Data were collected over five waves (from May 2013 to May 2015), in which participants completed 11 weeks of daily diary reporting. This study includes data from the last nine weeks of data collection, as the initial two weeks were used to get participants accustomed to completing the reports. Of the focal nine weeks, the first six weeks included the weekly meditation workshops, held in small groups with parallel formats for MM or LKM. For MM, practitioners were guided to direct their attention to the contents of consciousness in the present moment, with targets of consciousness progressively expanded over the six weeks, from breathing and hearing (Week 1), the body (Week 2), emotions (Week 3), thoughts (Week 4), and choiceless awareness (Week 5), with Week 6 being review and integration. For LKM, practitioners were guided to self-cultivate warm and friendly feelings, with social targets progressively expanded over the six weeks, from a loved one (Week 1), oneself (Week 2), an acquaintance (Week 3), a difficult person (Week 4), and all beings (Week 5), with Week 6 being review and integration (For more details, see Fredrickson et al., 2017). Participants’ data were aligned at the date of the first workshop session they attended.

Measures

Affect

Daily affect was assessed each evening using the modified Differential Emotions Scale (mDES). The mDES includes 20 items to assess the degrees to which respondents experience different emotions, both positive and negative, within a given time frame (Fredrickson, 2013). Ten positive emotions (i.e., amusement, awe, gratitude, hope, inspiration, interest, joy, love, pride, and serenity) and ten negative emotions (i.e., anger, contempt, disgust, embarrassment, fear, guilt, hate, sadness, shame, and stress) are assessed, each with a trio of adjectives (e.g., “awe, wonder, amazement” and “contemptuous, scornful, disdainful”). For each item, participants are asked to indicate the greatest degree to which they experienced the given feelings over the past 24 hours using a 5-point scale in which 0 = Not at all; 1 = A little bit; 2 = Moderately; 3 = Quite a bit; and 4 = Extremely. Composite scores for daily positive affect (PA-day) and daily negative affect (NA-day) were obtained by calculating the mean of the relevant ten items within each day. Respective reliabilities (omega coefficients) for between-person differences and within-person changes were 0.87 and 0.96 for positive affect and 0.79 and 0.96 for negative affect. (Details on reliability calculations are in Fredrickson et al., 2017).

Health behaviours and affect during health behaviours

Participants reported daily, retrospectively at end-of-day, on physical activity, fruit and vegetable intake, meditation practice, alcohol consumption, and tobacco use during the study. These assessments use the event reconstruction method, a survey approach validated to be comparable to ecological momentary assessment in minimizing retrospection bias (Grube, Schroer, Hentzschel & Hertel, 2008). Questions regarding health behaviours asked whether the participant engaged in the behaviour that day, and if so, how much they engaged in the given behaviour, and how positive and (separately) how negative they felt during the behaviour. For this study, we excluded tobacco use because few participants reported it. We use the terms “PA-behaviour” and “NA-behaviour” to distinguish these from the above-mentioned “PA-day” and “NA-day.”

For physical activity, participants were first asked, “In the last 24 hours, have you engaged in any [vigorous/moderate] physical activity?” Examples of vigorous activity included running, swimming, aerobics, sports, and heavy yard work. Examples of moderate activity included brisk walking, bicycling, vacuuming, and gardening. If participants responded, “yes” to vigorous or moderate physical activity, they were next asked to report how long they were active (in minutes). Next, using a 0–4 scale (devised for this study) and separately for vigorous and moderate physical activity, participants indicated how [positive/negative] they felt, on average, while they were physically active. For analyses, we summed responses across vigorous and moderate physical activity to create a composite variable of total minutes spent engaged in physical activity. Similarly, we computed means across vigorous and moderate physical activity regarding positive and negative affect (separately) felt during activity engagement.

Regarding diet, participants were asked, “Approximately how many cups of FRUIT did you eat in the last 24 hours?” This question was repeated for vegetables. Response options ranged from 0 to 6+ in one-cup intervals. Examples of serving volumes were provided (e.g., 4 large strawberries = ½ cup; 1 large ear of corn = 1 cup). Again using our 0–4 scale, participants then indicated how [positive/negative] they felt, on average, while “consuming the fruits and/or vegetables.” For analyses, we summed the separate fruit and vegetable volume responses to form one variable that assessed the combined cups of fruits and vegetables consumed.

For meditation, participants indicated whether they engaged in the behaviour. If “yes,” they were asked, “How much time (in minutes) did you spend on meditation in the last 24 hours? If there were multiple sessions, make sure to add them all together.” Again using our 0–4 scale, participants then indicated how [positive/negative] they felt, on average, while meditating.

For alcohol use, participants were again asked to endorse whether they engaged in the behaviour. If they responded “yes,” they were then asked, “Approximately how many alcoholic drinks did you consume in the last 24 hours?” Again using our 0–4 scale, participants then indicated how [positive/negative] they felt, on average, while drinking alcohol.

Additional measures, used as covariates in sensitivity analyses, are described in OSM.

Analytic Methods

We conducted analyses using R 3.3.1 and the “lme4” package. We first examined same-day relations between each of the four health behaviours (i.e., the three positive health behaviours plus alcohol consumption) and daily positive affect (i.e., PA-day). Given that Study 2 in Fredrickson et al. (2017) used the same sample as used here, we flag the small subset of results regarding meditation that were reported previously. Next, to explore possible causal order, we then examined lagged relations (e.g., physical activity yesterday predicting positive affect today). To test same-day relations, we fit four separate models, nesting daily reports (Level 1) within persons (Level 2). We asked: Is there a same-day association between PA-day and each of the four health behaviours? To test cross-day relations, we fit twelve separate, two-level models, nesting daily reports (Level 1) within persons (Level 2). We first asked: Does higher engagement in a health behaviour predict next-day PA-day? Then we asked: Does higher endorsement of PA-day predict next-day behaviours? Dependent variables across models were PA-day, and each of the four health behaviours. The behaviours of physical activity and time spent meditating were rescaled to allow for easier computation. Minutes spent engaged in physical activity were divided by 100 and minutes spent engaged in meditation were divided by ten. Finally, we further tested cross-day relations in which positive affect felt during each health behaviour on one day (PA-behaviour) predicts engagement in the health behaviour the next day. For completeness, we tested parallel analyses with negative affect (NA-day and NA-behaviour) as well as sensitivity analyses of moderation by randomized condition (MM vs. LKM) or by a set of covariates selected a priori (i.e., biological sex, age, ethnicity, body mass index, systemic inflammation). For brevity, these analyses are only summarized (with detailed results in OSM). Details on our model building approach, including preliminary growth curve models, are available in OSM. Importantly, each predictor was included both as a person mean-centered variable (Level 1) and as an individual mean over time variable (Level 2), to test for within-person differences and between-person differences, respectively.

Results

Preliminary Analyses

Comparing MM and LKM conditions revealed no between-group differences in PA-day, physical activity engagement, cups of fruits and vegetables consumed, minutes spent meditating, or number of alcoholic drinks consumed. Intraclass correlations (ICCs) indicated that considerable variance in each was attributable to between-person effects (i.e., from 25% to 70%). Regarding PA-behaviour, participants in MM reported higher levels of positive affect during physical activity and fruit and vegetable intake. No between-group differences emerged in PA-behaviour for meditation or alcohol use. Means by condition (MM vs. LKM) and statistical details of these analyses are presented in OSM.

Planned preliminary growth curve models (reported in detail in OSM) revealed linear increases within individuals over the nine weeks for PA-day (reported previously in Fredrickson et al., 2017) and fruit and vegetable intake (β = 0.08, p = .01). A linear decrease in time spent meditating also emerged (β = −0.10, p = .0001), which was not altogether unexpected, given that the meditation workshops ended three weeks before the end of daily reporting. No within-person change over time was evident for physical activity or alcohol use. No differences between MM and LKM groups emerged in any growth curve models. Post-hoc analyses revealed same-day associations—both between-persons (β = .28, p = .03) and within-persons (β = .10, p < .0001)—between time spent meditating and cups of fruits and vegetables consumed (see OSM for details).

Same-day Associations

Fixed effects models that tested same-day relations between PA-day and health behaviours are presented in Table 1. For each behaviour examined, a positive, significant within-person association was evident between the behaviour and PA-day. These within-person results indicate that participants who engaged in the behaviour more frequently than their own daily average reported higher positive affect for that day. Excepting for alcohol use, a positive, significant between-person association was also evident between all health behaviours and PA-day. These between-person results indicate that, on average, those participants who engaged in the given behaviour more frequently than others reported higher intensity daily positive affect.

Cross-day Associations

Tests of cross-day associations can shed light on possible within-person causal directions between positive affect and health behaviours. (We note here that between-person cross-day associations with PA-day, reported in OSM Tables, are virtually identical to between-person same-day associations.) Models in which PA-day predicted the behaviours of meditation and alcohol use did not converge, nor did the model in which PA-behaviour for meditation predicted next-day meditation practice. Thus, of the 12 models we fit, results are available for nine.

Yesterday’s health behaviours predicting today’s daily positive affect

Results for the lagged associations between health behaviours on one day and PA-day on the next are summarized here, with statistical details in Table S1 in OSM. In all models, PA-day demonstrated significant autoregressive effects (β = .09, p < .01). The pattern of these autoregressive effects revealed that participants who experienced higher than their own average intensity of PA-day on one day experienced higher intensity PA-day the next day. Significant, within-person lagged relationships were found for the behaviours of fruit and vegetable intake and alcohol consumption: On average, participants who reported eating higher than their own average amount of fruits and vegetables on one day experienced higher intensity PA-day the next day (β = .01, p < .05); participants who reported drinking a greater number of alcoholic drinks than their own average experienced lower intensity PA-day the next day (β = −.01, p < .01).

Yesterday’s daily positive affect predicting today’s health behaviours

Results for the lagged associations between PA-day on one day and health behaviours on the next are summarized here, with statistical details in Table S2 in OSM. Both physical activity and fruit and vegetable intake demonstrated significant, positive autoregressive effects (Phy: β = .04, p < .001; FVI: (β = .25, p < .001), indicating that participants who were physically active for more time than their own average, or who consumed a greater amount of fruits and vegetables than their own average were likely to be active for longer and eat greater amounts of fruits and vegetables the next day. However, no significant within-person associations emerged for yesterday’s PA-day predicting next-day health behaviours.

Yesterday’s positive affect during health behaviours predicting today’s health behaviours

Results for the lagged associations between positive affect experienced during a given health behaviour (PA-behaviour) on one day and the degree of engagement in that same health behaviour the next day are summarized here, with statistical details in Table S3 in OSM. (The growth curve for fruit and vegetable intake could not be included in the model in order to reach convergence.) Positive, between-person relationships emerged for the behaviours of physical activity (p = .01; β unavailable due to convergence issues) and fruit and vegetable intake (β = .60, p < .001), in which participants who, on average, reported higher levels of PA-behaviour for each of these behaviours reported higher levels of those behaviours, on average, the next day. Additionally, a positive, within-person effect for fruit and vegetable intake emerged (β = .06, p = .002), in which participants who on a given day reported higher PA-behaviour for fruit and vegetable intake (relative to their own average) were likely to consume greater amounts of fruits and vegetables the next day. No other significant effects emerged.

Parallel analyses for negative affect

We repeated each analysis reported above replacing PA-day and PA-behaviour with NA-day and NA-behaviour, respectively, and report the results in the OSM (see Section 5 and Tables S4–S7). The overall pattern of mostly null results suggests that the associations reported herein are specific to positive affect. The only significant same-day association was a negative within-person link suggesting that devoting more time to physical activity on a given day was linked to experiencing less intense negative affect that day (β = −.02, p < .001). The cross-day associations also revealed a between-person effect such that those who reported higher intensity negative affect during alcohol consumption on one day reported more alcohol consumption the next day (β = .32, p < .001), plus a within-person effect such that more alcohol consumption on one day predicting higher intensity negative affect the next day (β = .01, p = .01).

Sensitivity analyses

We also explored whether the results reported here were moderated by randomized condition (MM vs. LKM) or by demographic (i.e., biological sex, age, ethnicity) or health-related variables (i.e., body mass index, systemic inflammation). Table 2 provides a summary of the primary findings of the current investigation. The note that accompanies Table 2 identifies the few cases in which primary findings are qualified by the results of sensitivity analyses (reported in OSM, Section 6). The overall pattern suggests that, among positive health behaviours, the primary findings for physical activity and fruit and vegetable intake were robust to sensitivity analyses and need no qualifications, whereas those regarding meditation need to be qualified. Specifically, the between-person effect for meditation was not sufficiently robust to survive sensitivity analyses and the within-person effect for meditation was stronger for those randomized to LKM versus MM (β = .02, p = .02; reported in Fredrickson et al., 2017). One intriguing result for the negative health behaviour of alcohol consumption was that participants randomized to learn LKM (versus MM) showed a weaker, within-person link between alcohol use and daily positive affect (β = −.02, p = .046).

Table 2.

Summary of Results for Hypotheses and Research Questions on Same-day and Lagged Associations Between Positive Affect and Positive Health Behaviours.

Hypothesis Supported: Significant Same-Day Between-Person Effects for:
  Physical Activity ←→ Daily Positive Affect
  Fruit and Vegetable Intake ←→ Daily Positive Affect
  Meditation ←→ Daily Positive Affect a, b
Hypothesis Supported: Significant Same-day Within-Person Effects for:
  Physical Activity ←→ Daily Positive Affect
  Fruit and Vegetable Intake ←→ Daily Positive Affect
  Meditation ←→ Daily Positive Affecta
Research Questions Addressed: Significant Lagged Within-Person Effects for:
  One Day’s Positive Affect During Fruit and Vegetable Intake (FVI) → Next Day’s FVI
  One Day’s FVI → Next Day’s Daily Positive Affect
Open in a new tab

Note:

a

Sensitivity analyses revealed that this effect was no longer significant when condition interaction effects were added to the model and that results indicated a stronger within-person effect for loving-kindness meditation than for mindfulness meditation.

b

Sensitivity analyses revealed that this effect was no longer significant when covariates were added to the model.

Discussion

Positive affect, when frequently experienced, has been established by meta-analyses to predict longevity (Chida & Steptoe, 2008; Zhang & Han, 2016). One of several mechanisms that may account for this long-term benefit of frequent positive affect hinges on the positive health behaviours that positive affect may motivate (Van Cappellen et al., 2017). Alternatively, to the extent that positive health behaviours produce positive affect, the link between positive affect and longevity might be explained by the specific health-related behaviours that give rise to enjoyment. The current investigation was undertaken to illuminate these various behavioural accounts by assessing the links between two forms of positive affect—enjoyment of the day and enjoyment of specific health-related behaviours—and three types of positive health behaviours—physical activity, fruit and vegetable intake, and meditation. Specifically, we examined their same-day, cross-day, and day-to-day upward spiral relations. As points of contrast, we also conducted analyses with parallel measures of negative affect and with the negative health behaviour of alcohol consumption.

The overall findings are summarized in Table 2. First, we hypothesized that our data would replicate past findings of same-day relations between daily positive affect and each of the three positive health behaviours (e.g., Reed & Ones, 2006; White et al., 2013; Fredrickson et al., 2017). As shown in Table 2, for each distinct behaviour, significant same-day associations were evident, as indicated by both between-person effects and within-person effects. The significant between-person effects indicate that individuals in midlife who (a) are more physically active, (b) eat more fruits and vegetables, and (c) meditate more (relative to other midlife individuals), also report experiencing a range of daily positive emotions to higher degrees. Evidence for these three individual differences is complemented by evidence for day-to-day processes that fluctuate within individuals. That is, the significant within-person effects indicate that on days in which midlife adults (a) are more physically active, (b) eat more fruits and vegetables, and (c) meditate more (relative to other days), they report experiencing positive emotions to higher degrees. (See Table 2 for qualifications to the effects for meditation.) Because within-person effects better match theoretical questions about change over time (Curran & Bauer, 2011), these day-by-day variations underscore the need for cross-day analyses, which can begin to unpack whether positive health behaviours increase enjoyment of the day, whether enjoyment of the day increases positive health behaviours, or whether both directions of causality are viable. Regardless, the observed associations appear to be specific to positive affect: Same-day links between the three positive health behaviours and daily negative affect did not emerge, with the sole exception of evidence that days with more physical activity are marked by lower intensities of negative affect.

In contrast to the hypothesized same-day associations, which all received empirical support, results for our research questions regarding cross-day effects were mixed. We note that two of the three lagged models for meditation did not converge, and in the one that did converge, no significant within-person effect emerged. Likewise, although all three lagged models for physical activity converged, no significant within-person effects emerged for that activity either. By contrast, for the positive health behaviour of eating fruits and vegetables, two of the three lagged within-person effects emerged as significant. First, when individuals ate more fruits and vegetables on one day, they enjoyed their next day more, as indicated by their self-reports of experiencing a range of daily positive emotions to higher degrees. Importantly, the reciprocal lagged effect was not significant, that running from daily positive affect on one day to fruit and vegetable intake on the next. The finding of cross-day affective benefits for fruit and vegetable intake in our midlife adult sample joins prior similar evidence from young adult samples (Conner et al, 2017; White et al., 2013), each of which showed comparable (small) effect sizes. In both age groups, this effect was not mirrored by reciprocal evidence that yesterday’s totality of positive affect predicted today’s fruits and vegetables intake.

To our knowledge, the present investigation is the first to test the association between the positive affect people experienced specifically during fruit and vegetable intake on one day and the amount of fruit and vegetable intake on the next. This is the second of the two lagged associations that received empirical support. Taken together, the two significant cross-day findings regarding fruit and vegetable consumption suggest a day-to-day upward spiral dynamic in which the enjoyment people felt while eating their fruits and vegetables yesterday predicts eating more fruits and vegetables today, which in turn forecasts having a more enjoyable day tomorrow. Again, these associations appear to be specific to positive affect: Cross-day links between the three positive health behaviours and daily negative affect did not emerge, with the sole exception of evidence that people who experienced more negative affect during meditation on one day, meditated less the next day.

Whereas negative affect provides one point of contrast in this investigation, the negative health behaviour of alcohol consumption provides another. Although tests of same-day associations suggest that days on which our midlife participants consumed more alcohol were days with higher positive affect, tests of lagged associations suggest that these days were followed by days with both reduced positive affect and increased negative affect. Further underscoring that excessive alcohol consumption is a negative health behaviour, the more negative affect that participants felt while drinking on one day predicted more alcohol consumption the next day, a finding that may reflect the long-terms risks of drinking to self-medicate negative feelings (Crum et al., 2013). Our findings suggest that although alcohol consumption on a given day is linked with higher positive affect that same day, which may reflect a short-term gain, it also appears to bring long-term pain, in terms of worsened affect the next day, which may prompt greater consumption the next. Intriguingly, sensitivity analyses suggest that the observed short-term gain of alcohol consumption is lessened among participants randomized to learn loving-kindness meditation, a finding that aligns with the idea that natural rewards, like interpersonal warmth, may help people to curb unhealthy pleasures (c.f., Garland, Atchley, Hanley, Zubieta & Froeliger, 2019)

The results of this investigation provide robust evidence for same-day links between daily positive affect and two of the most commonly studied positive health behaviours, namely, physical activity and fruit and vegetable intake. More limited evidence emerged for similar affective links for meditation. When we examined lagged associations from one day to the next, a pictured emerged to suggest that greater fruit and vegetable intake may somehow support enjoyment of the next day as well. We hasten to add that causality cannot be inferred from the results reported here. Randomized controlled trials are called for, such as the one that found well-being-enhancing results when young adults consumed more fresh fruits and vegetables (Conner et al., 2017).

We also underscore that the present investigation only examined lagged associations on the scale of one day to the next. Longer lag times between positive affect and positive health behaviours, such as from one month, season, or year to the next, or shorter lag times, such as within a day may also be present and should be targets of future study. Indeed, past evidence suggests that beyond the affective boosts in the hours that follow physical activity, significant lagged effects from positive affect to greater activity engagement have emerged that are both longer (over months; Williams et al., 2012) and shorter (over hours; Emerson et al., 2018) than the cross-day effects examined here. The observed cross-day effects for fruit and vegetable intake may also reflect that eating, for most people, is a daily imperative, whereas physical activity, meditation, and alcohol consumption are optional activities.

Although not the target of study here, consistent with the observation that health-promoting behaviours tend to cluster (Lippke et al., 2012), our initial descriptive evidence suggests that, in the process of learning to meditate, participants’ daily positive affect increased over time. Additionally, a side benefit appears to have emerged in terms of an increase in participants’ fruit and vegetable consumption. Moreover, post-hoc analyses of same-day associations between meditation and fruit and vegetable intake revealed both between-person and within-person effects that are consistent with the benefits of mindful eating (e.g., Jordan, Wang, Donatoni, & Meier, 2014). This latter pattern of results indicates both that people who meditate more (relative to other people) eat more fruits and vegetables and also that days that include more minutes of meditation (relative to other days) include more servings of fruits and vegetables. Further research on same-day and cross-day associations among distinct positive health behaviours is warranted.

The research design used here has both strengths and weaknesses. Strengths include the sample size and the densely repeated assessments of affect and behaviour over 9 weeks. Weaknesses include that it relies entirely on participants’ retrospective self-reports of affect and behaviour and therefore carries the limitations of retrospective self-reports. To a degree, those limitations are lessened because participants only reported on one day at a time or one event at a time (e.g., a bout of physical activity on that day), rather than on their affect and health behaviours in general. Nevertheless, future work would benefit from implicit or ecological momentary assessments of affect and objective assessments of behaviour. Another weakness is that the study relied on secondary data analyses that were not pre-registered. Additionally, generalization to other age groups and to other ethnicities and cultures may not be warranted.

Also unanswerable from this research is why eating more fruits and vegetables yesterday might lead to more positive affect today. From a psychological perspective, perhaps when people are able to acquire, prepare, and consume more fruits and vegetables they feel more virtuous or efficacious that day and the next. The finding that positive affect felt during fruit and vegetable consumption predicts increased next-day consumption aligns with a recent randomized study that found that when exciting language is used to describe vegetables (e.g., “twisted carrots” and “dynamite beets”) consumption at a large university cafeteria increased (Turnwald, Boles, & Crum, 2017). From a nutrition perspective, perhaps the micro-nutrients, fiber, or water content of fruits and vegetables, or their effects on anti-inflammatory, antioxidant, or microbiome processes support the experience of same-day and next-day positive affect. Other explanations for this link do not assume a causal effect, but instead reflect “third variable” effects. As one example, to the extent that people cycle through healthy and unhealthy eating patterns, they may feel that if they ate healthy yesterday, they deserve to indulge themselves (on sweets or comfort foods) today, and that those unhealthy foods created today’s positive affect. Undermining this alternative explanation, however, is the significant, positive autoregressive effect for fruit and vegetable intake, which revealed that study participants’ intake levels did not cycle between high and low consumption. Rigorous tests of plausible mechanisms that may account for the observed lagged link between eating more fruits and vegetables on one day and greater enjoyment of the next day will require interdisciplinary collaborations between nutrition and affective scientists. We encourage such pursuits and look forward to the fruits of them.

Supplementary Material

Supp 1

Acknowledgements

This investigation was supported by a research grant awarded to Barbara L. Fredrickson by the National Cancer Institute (R01CA170128) of the U.S. National Institutes of Health (NIH). This funding agency played no role in study design, data collection, data analysis, decision to publish, or preparation of the manuscript. The authors wish to thank Ann M. Firestine for overseeing data collection and data management, and members of the Positive Emotions and Psychophysiology Laboratory for offering thoughtful comments on earlier versions of the manuscript. The authors also extend their gratitude to the study participants, each of whom devoted time and energy across months to be involved in this research.

Contributor Information

Barbara L. Fredrickson, Department of Psychology and Neuroscience, University of North Carolina, Chapel Hill, NC, USA

Cara Arizmendi, Department of Psychology and Neuroscience, University of North Carolina, Chapel Hill, NC, USA.

Patty Van Cappellen, Social Science Research Institute, Duke University, Durham, NC, USA.

References

  1. Aarts H, Custers R, & Marien H. (2008). Preparing and Motivating Behavior Outside of Awareness. Science, 319(5870), 1639–1639. 10.1126/science.1150432 [DOI] [PubMed] [Google Scholar]
  2. Blevins CL, Sagui SJ, & Bennett JM (2017). Inflammation and positive affect: Examining the stress-buffering hypothesis with data from the National Longitudinal Study of Adolescent to Adult Health. Brain, Behavior, and Immunity, 61, 21–26. 10.1016/j.bbi.2016.07.149 [DOI] [PubMed] [Google Scholar]
  3. Bold KW, Fucito LM, Corbin WR, DeMartini KS, Leeman RF, Kranzler HR, & O’Malley SS (2016). Daily relations among affect, urge, targeted naltrexone, and alcohol use in young adults. Experimental and Clinical Psychopharmacology, 24(5), 367–375. 10.1037/pha0000090 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brookie KL, Best GI, & Conner TS (2018). Intake of raw fruits and vegetables is associated with better mental health than intake of processed fruits and vegetables. Frontiers in Psychology, 9. 10.3389/fpsyg.2018.00487 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Burke A, Lam CN, Stussman B, & Yang H. (2017). Prevalence and patterns of use of mantra, mindfulness and spiritual meditation among adults in the United States. BMC Complementary and Alternative Medicine, 17. 10.1186/s12906-017-1827-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cameron DS, Bertenshaw EJ, & Sheeran P. (2018). Positive affect and physical activity: Testing effects on goal setting, activation, prioritisation, and attainment. Psychology & Health, 33(2), 258–274. 10.1080/08870446.2017.1314477 [DOI] [PubMed] [Google Scholar]
  7. Chang Y-P, Algoe SB, & Chen LH (2017). Affective valence signals agency within and between individuals. Emotion, 17(2), 296–308. 10.1037/emo0000229 [DOI] [PubMed] [Google Scholar]
  8. Chida Y, & Steptoe A. (2008). Positive Psychological Well-Being and Mortality: A Quantitative Review of Prospective Observational Studies: Psychosomatic Medicine, 70(7), 741–756. 10.1097/PSY.0b013e31818105ba [DOI] [PubMed] [Google Scholar]
  9. Cohn MA, & Fredrickson BL (2010). In search of durable positive psychology interventions: Predictors and consequences of long-term positive behavior change. The Journal of Positive Psychology, 5(5), 355–366. 10.1080/17439760.2010.508883 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cohn MA, Fredrickson BL, Brown SL, Mikels JA, & Conway AM (2009). Happiness unpacked: Positive emotions increase life satisfaction by building resilience. Emotion, 9(3), 361–368. 10.1037/a0015952 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Conner TS, Brookie KL, Carr AC, Mainvil LA, & Vissers MCM (2017). Let them eat fruit! The effect of fruit and vegetable consumption on psychological well-being in young adults: A randomized controlled trial. PLoS ONE, 12(2). 10.1371/journal.pone.0171206 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Conner TS, Brookie KL, Richardson AC, & Polak MA (2015). On carrots and curiosity: Eating fruit and vegetables is associated with greater flourishing in daily life. British Journal of Health Psychology, 20(2), 413–427. 10.1111/bjhp.12113 [DOI] [PubMed] [Google Scholar]
  13. Crum RM, Mojtabai R, Lazareck S, Bolton JM, Robinson J, Sareen J, … Storr CL (2013). A Prospective Assessment of Reports of Drinking to Self-medicate Mood Symptoms with the Incidence and Persistence of Alcohol Dependence. JAMA Psychiatry, 70(7), 718–726. 10.1001/jamapsychiatry.2013.1098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Curran PJ, & Bauer DJ (2011). The Disaggregation of Within-Person and Between-Person Effects in Longitudinal Models of Change. Annual Review of Psychology, 62(1), 583–619. 10.1146/annurev.psych.093008.100356 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Custers R, & Aarts H. (2005). Positive Affect as Implicit Motivator: On the Nonconscious Operation of Behavioral Goals. Journal of Personality and Social Psychology, 89(2), 129–142. 10.1037/0022-3514.89.2.129 [DOI] [PubMed] [Google Scholar]
  16. Ekkekakis P, Parfitt G, & Petruzzello SJ (2011). The Pleasure and Displeasure People Feel When they Exercise at Different Intensities. Sports Medicine, 41(8), 641–671. 10.2165/11590680-000000000-00000 [DOI] [PubMed] [Google Scholar]
  17. Emerson JA, Dunsiger S, & Williams DM (2018). Reciprocal within-day associations between incidental affect and exercise: An EMA study. Psychology & Health, 33(1), 130–143. 10.1080/08870446.2017.1341515 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Emerson JA, & Williams DM (2015). The Multifaceted Relationship Between Physical Activity and Affect: Physical Activity and Affect. Social and Personality Psychology Compass, 9(8), 419–433. 10.1111/spc3.12190 [DOI] [Google Scholar]
  19. Fredrickson BL (2013). Positive Emotions Broaden and Build. In Devine P. & Plant A. (Eds.), Advances in Experimental Social Psychology (Vol. 47, pp. 1–53). 10.1016/B978-0-12-407236-7.00001-2 [DOI] [Google Scholar]
  20. Fredrickson BL, Boulton AJ, Firestine AM, Van Cappellen P, Algoe SB, Brantley MM, … Salzberg S. (2017). Positive Emotion Correlates of Meditation Practice: A Comparison of Mindfulness Meditation and Loving-Kindness Meditation. Mindfulness, 8(6), 1623–1633. 10.1007/s12671-017-0735-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fredrickson BL, Cohn MA, Coffey KA, Pek J, & Finkel SM (2008). Open hearts build lives: Positive emotions, induced through loving-kindness meditation, build consequential personal resources. Journal of Personality and Social Psychology, 95(5), 1045–1062. 10.1037/a0013262 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Garland EL, Atchley RM, Hanley AW, Zubieta J-K & Froeliger B. (2019). Mindfulness-oriented recovery enhancement remediates hedonic dysregulation in opioid users: Neural and affective evidence for target engagement. Science Advances, 5: eaax1 569. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Grube A, Schroer J, Hentzschel C, & Hertel G. (2008). The event reconstruction method: An efficient measure of experience-based job satisfaction. Journal of Occupational and Organizational Psychology, 81, 669–689. [Google Scholar]
  24. Ironson G, Banerjee N, Fitch C, & Krause N. (2018). Positive emotional well-being, health Behaviors, and inflammation measured by C-Reactive protein. Social Science & Medicine, 197, 235–243. 10.1016/j.socscimed.2017.06.020 [DOI] [PubMed] [Google Scholar]
  25. Jordan CH, Wang W, Donatoni L, & Meier BP (2014). Mindful eating: Trait and state mindfulness predict healthier eating behavior. Personality and Individual Differences, 68, 107–111. 10.1016/j.paid.2014.04.013 [DOI] [Google Scholar]
  26. Kok BE, Coffey KA, Cohn MA, Catalino LI, Vacharkulksemsuk T, Algoe SB, … Fredrickson BL (2013). How Positive Emotions Build Physical Health: Perceived Positive Social Connections Account for the Upward Spiral Between Positive Emotions and Vagal Tone. Psychological Science, 24(7), 1123–1132. 10.1177/0956797612470827 [DOI] [PubMed] [Google Scholar]
  27. Kubzansky LD, & Thurston RC (2007). Emotional Vitality and Incident Coronary Heart Disease: Benefits of Healthy Psychological Functioning. Archives of General Psychiatry, 64(12), 1393–1401. 10.1001/archpsyc.64.12.1393 [DOI] [PubMed] [Google Scholar]
  28. Lippke S, Nigg CR, & Maddock JE (2012). Health-Promoting and Health-Risk Behaviors: Theory-Driven Analyses of Multiple Health Behavior Change in Three International Samples. International Journal of Behavioral Medicine, 19(1), 1–13. 10.1007/s12529-010-9135-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Okely JA, Weiss A, & Gale CR (2017). The interaction between stress and positive affect in predicting mortality. Journal of Psychosomatic Research, 100, 53–60. 10.1016/j.jpsychores.2017.07.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ong AD, Benson L, Zautra AJ, & Ram N. (2018). Emodiversity and biomarkers of inflammation. Emotion, 18(1), 3–14. 10.1037/emo0000343 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ouweneel E, Blanc PML, & Schaufeli WB (2012). Don’t leave your heart at home. Career Development International. 10.1108/13620431211280123 [DOI] [Google Scholar]
  32. Reed J, & Ones DS (2006). The effect of acute aerobic exercise on positive activated affect: A meta-analysis. Psychology of Sport and Exercise, 7(5), 477–514. 10.1016/j.psychsport.2005.11.003 [DOI] [Google Scholar]
  33. Rhodes RE, & Kates A. (2015). Can the Affective Response to Exercise Predict Future Motives and Physical Activity Behavior? A Systematic Review of Published Evidence. Annals of Behavioral Medicine, 49(5), 715–731. 10.1007/s12160-015-9704-5 [DOI] [PubMed] [Google Scholar]
  34. Rice EL, Adair KC, Tepper SJ, & Fredrickson BL (2019). Perceived social integration predicts future physical activity through positive affect and spontaneous thoughts. Emotion, No Pagination Specified-No Pagination Specified. 10.1037/emo0000616 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rice EL, & Fredrickson BL (2017). Do positive spontaneous thoughts function as incentive salience? Emotion, 17(5), 840–855. 10.1037/emo0000284 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Robles TF, Brooks KP, & Pressman SD (2009). Trait positive affect buffers the effects of acute stress on skin barrier recovery. Health Psychology, 28(3), 373–378. 10.1037/a0014662 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Turnwald BP, Boles DZ, & Crum AJ (2017). Association Between Indulgent Descriptions and Vegetable Consumption: Twisted Carrots and Dynamite Beets. JAMA Internal Medicine, 177(8), 1216–1218. 10.1001/jamainternmed.2017.1637 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Van Cappellen P, Rice EL, Catalino LI, & Fredrickson BL (2018). Positive affective processes underlie positive health behaviour change. Psychology & Health, 33(1), 77–97. 10.1080/08870446.2017.1320798 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Warner RM, Frye K, Morrell JS, & Carey G. (2017). Fruit and Vegetable Intake Predicts Positive Affect. Journal of Happiness Studies, 18(3), 809–826. 10.1007/s10902-016-9749-6 [DOI] [Google Scholar]
  40. White BA, Horwath CC, & Conner TS (2013). Many apples a day keep the blues away—Daily experiences of negative and positive affect and food consumption in young adults. British Journal of Health Psychology, 18(4), 782–798. 10.1111/bjhp.12021 [DOI] [PubMed] [Google Scholar]
  41. Wichers M, Peeters F, Rutten BPF, Jacobs N, Derom C, Thiery E, … van Os J. (2012). A time-lagged momentary assessment study on daily life physical activity and affect. Health Psychology, 31(2), 135–144. 10.1037/a0025688 [DOI] [PubMed] [Google Scholar]
  42. Williams DM, Dunsiger S, Jennings EG, & Marcus BH (2012). Does Affective Valence During and Immediately Following a 10-Min Walk Predict Concurrent and Future Physical Activity? Annals of Behavioral Medicine : A Publication of the Society of Behavioral Medicine, 44(1), 43–51. 10.1007/s12160-012-9362-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wilson TE, Weedon J, Cohen MH, Golub ET, Milam J, Young MA, … Fredrickson BL (2017). Positive affect and its association with viral control among women with HIV infection. Health Psychology, 36(1), 91–100. 10.1037/hea0000382 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhang Y, & Han B. (2016). Positive affect and mortality risk in older adults: A meta-analysis: Positive affect and mortality meta-analysis. PsyCh Journal, 5(2), 125–138. 10.1002/pchj.129 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp 1
NIHMS1618345-supplement-Supp_1.docx (162.9KB, docx)

RESOURCES