A Matter of the Heart: Daytime Relationship Functioning and Overnight Heart Rate in Young Dating Couples

Abstract

Background

Although past longitudinal research demonstrates that romantic partners affect one another’s health outcomes, considerably less is known about how romantic experiences “get under the skin” in everyday life.

Purpose

The current study investigated whether young couples’ naturally occurring feelings of closeness to and annoyance with each other during waking hours were associated with their overnight cardiovascular activity.

Methods

Participants were 63 heterosexual young adult dating couples (M_age = 23.07). Using ecological momentary assessments, couples reported their hourly feelings of closeness to and annoyance with their partners across 1 day; subsequent overnight heart rate was captured through wearable electrocardiogram biosensors. Actor–partner interdependence models tested whether individuals’ overnight heart rate varied as a function of (a) their own daytime feelings of closeness and annoyance (actor effects) and (b) their partner’s daytime feelings of closeness and annoyance (partner effects) while controlling for daytime heart rate.

Results

Although young adults’ feelings of romantic closeness and annoyance were unrelated to their own overnight heart rate (i.e., no actor effects), gender-specific partner effects emerged. Young men’s nocturnal heart rate was uniquely predicted by their female partner’s daytime relationship feelings. When women felt closer to their partners during the day, men exhibited lower overnight heart rate. When women felt more annoyed with their partners during the day, men exhibited heightened overnight heart rate.

Conclusions

The findings illustrate gender-specific links between couple functioning and physiological arousal in the everyday lives of young dating couples, implicating physiological sensitivity to partner experiences as one potential pathway through which relationships affect health.

Capturing a day in the life of young heterosexual dating couples, women’s daytime feelings about their partners were related to men’s overnight heart rate. When women felt more annoyed and irritated with their partners throughout the day, men subsequently exhibited a higher overnight heart rate; when women felt emotionally closer to their partners throughout the day, men exhibited a lower overnight heart rate

Close relationships serve as vital determinants of long-term physical health outcomes [1–3], including cardiovascular disease, the leading cause of death for men and women worldwide [4]. Individuals experiencing romantic violence, conflict, or divorce are more likely to develop cardiovascular disease in adulthood [5, 6], whereas adults who perceive greater marital quality exhibit reduced cardiovascular stress reactivity [6] and better cardiovascular health [7]. Despite evidence for far-reaching effects of adult romantic relationships on cardiovascular outcomes, considerably less is known about the early stages of intimate relationships and how they might incrementally contribute to health problems over time.

Although extensive research highlights long-term links between couple functioning and disease outcomes, theoretical perspectives have called for a more mechanistic approach, wherein relationship processes are proposed to affect health by altering underlying stress regulatory systems [2, 3, 8]. To develop a fine-grained understanding of links between intimate relationships and cardiovascular disease outcomes, it may, therefore, be important to consider how subtle, everyday couple interactions relate to acute biological changes. The goal of the current study, in turn, is to investigate whether naturally occurring positive and negative romantic experiences relate to couples’ cardiovascular arousal in daily life. In doing so, we are guided by the premise that everyday ups and downs in young people’s romantic relationships, although often appearing trivial or transient, may map onto prolonged changes in physiological arousal; repeated activation of stress-sensitive biological systems could then cumulatively impact health over time [9, 10].

Cardiovascular reactivity is one key physiological pathway through which close relationships affect long-term health outcomes [11, 12]. That is, people who exhibit exaggerated cardiovascular responses to stress or chronically elevated heart rate experience greater long-term risk for hypertension and cardiovascular disease [13, 14]. To better understand interpersonal correlates of such reactivity, a number of studies have investigated associations between romantic relationship dynamics and cardiovascular responses in lab-based settings. For example, couples exhibit heightened heart rate and blood pressure while having an argument [15], whereas they show attenuated cardiovascular stress reactivity in the context of spousal support [16]. In the current study, we take couples outside of the lab and “into the wild” to investigate whether naturally occurring positive (closeness) and negative (annoyance) feelings toward one’s partner predict young people’s overnight cardiovascular activity. We focus specifically on overnight cardiovascular activity because heightened arousal can interfere with the critical restorative function of sleep and its health benefits [17].

Studying couple functioning and cardiovascular activity in naturalistic settings offers an ecologically valid approach [10] that can also provide unique insight into individual differences in physiological recovery processes. For example, although not necessarily focused on couple interactions, experience sampling research documents prolonged effects of stressors on cardiovascular activity such that people who experience higher levels of worry during the day exhibit elevated heart rate during sleep [18, 19]. It is important to understand interpersonal factors that predict prolonged cardiovascular activation—even once social interactions have ended (e.g., during sleep)—because people with poorer stress-related heart rate recovery are more likely to develop cardiovascular health problems [20]. However, to our knowledge, no studies have considered whether couples’ daytime functioning spills over into elevated overnight heart rate—that is, even in the absence of social interaction.

Although the research reviewed thus far primarily focuses on how people’s experiences in and subjective perceptions of romantic relationships affect their own physiological arousal and disease outcomes, there is also empirical evidence that individuals’ health is shaped by their partner’s experience in the relationship. These findings are primarily based on longitudinal studies that track couples over the course of multiple years. For example, among elderly couples, greater spousal life satisfaction prospectively predicts reduced mortality risk 8 years later, over and above one’s own life satisfaction [21]. Similarly, spousal life satisfaction has been identified as a concurrent and prospective predictor of physical well-being such that individuals with happy, compared to unhappy, romantic partners report more exercise, a health-promoting behavior, and overall better health [22].

Some studies have also investigated potential gender differences in links between relationship functioning and health, albeit with mixed results. For example, in a meta-analysis of studies examining associations between adults’ marital quality and health outcomes, there was little evidence for gender differences—that is, higher marital quality (as reported by self or partner or coded through behavioral observations) predicted better physical health for both men and women [6]. However, other research focusing specifically on cross-over effects, where one partner’s experience or feelings are related to changes in the other partner’s outcomes, provide more consistent evidence for gender differences such that the effects are pronounced in men compared to women. That is, men appear particularly physiologically sensitive to their partners’ feelings and experiences. For example, men—but not women—exhibit attenuated cortisol reactivity when their romantic partners report greater relationship satisfaction [23] and when female partners provide support during a lab-based stressor [24]. Men, but not women, also exhibit heightened physiological reactivity when navigating a lab-based relationship conflict discussion with an insecurely attached [25] or depressed partner [22]. Such gender differences may be especially pronounced in the context of early intimate relationships insofar as past research shows that young men, compared to young women, express less confidence in navigating romantic interactions [26]. Together, these findings suggest that, although findings on gender differences have been mixed, differences tend to emerge when examining cross-partner effects of relationship factors on health (i.e., as opposed to individuals’ subjective perceptions of their relationships).

The Present Study

The overarching goal here is to investigate how young adult couples’ daytime relationship feelings predict their overnight cardiovascular activity—a putative restorative process— in a naturalistic setting. Young adulthood is a time when committed, intimate relationships become increasingly common [27] and couples may become locked into patterns of functioning that, for better or for worse, have a cumulative impact on long-term health [2, 8]. However, the majority of past research on relationships and health focus on middle or older adults in committed, long-term relationships [3] and examine patterns of couple functioning and physiological arousal in lab-based settings (for exception, see [10, 28]). Thus, capturing a “day in the life” of young dating couples, the current study capitalizes on intensive self-reported ecological momentary assessments and physiological data to examine how relationship functioning during waking hours may relate to increased or attenuated overnight cardiovascular activity. By focusing on overnight heart rate, the study can evaluate how interpersonal functioning predicts cardiovascular activity during sleep, over and above daytime cardiovascular activity [17, 18].

Three aims were investigated to identify how positive and negative romantic relationship perceptions map onto one’s own and one’s partner’s physiological arousal in everyday life. First, participants’ own feelings of closeness to and annoyance with their partners throughout the day were tested as predictors of their own overnight heart rate. It was hypothesized that young adults who felt closer to their partners during the day would exhibit lower overnight heart rate, whereas those who felt more annoyed with their partners would exhibit higher overnight heart rate. Second, participants’ partners’ feelings of closeness and annoyance throughout the day were tested as predictors of their own overnight heart rate. It was hypothesized that young adults would exhibit lower overnight heart rate when their partners felt closer to them during the day, whereas they would exhibit higher overnight heart rate when their partners felt more annoyed with them during the day. In light of past evidence that men exhibit heightened sensitivity to relationship dynamics [22, 25], gender differences were also tested across all analyses, and it was hypothesized that the aforementioned effects would be stronger in men compared to women.

Method

Participants

Participants were young adult couples involved in a study examining how family-of-origin experiences relate to young adult dating experiences [28]. Most couples were recruited from the greater Los Angeles area through online postings and were included if they were between the ages of 18 to 25 and had been dating for at least 2 months. A subset of participants (n = 14) was rerecruited from a larger longitudinal study of families that began when participants were children and early adolescents and then brought in a dating partner for this study.

In order to be included in the current analytic sample, both members of a couple had to participate in all components of a day-long home data collection procedure, including wearing heart rate monitors overnight. Given the present study’s interest in within-couple gender differences, the current study only focused on data collected from heterosexual couples. As a function of these inclusion criteria, the current analytic sample included 63 couples (M_age = 23.07; standard deviation [SD] = 3.03) from a larger study (N = 109 couples). Specifically, couples were excluded if they were same sex (n = 3), one of the partners was missing overnight heart rate data (n = 28), or both partners were missing overnight heart rate data (n = 18). The relatively large number of participants without overnight heart rate data was due to issues of physical discomfort (e.g., taking heart rate monitor off at night because it was itchy), improper reattachment (e.g., taking off monitor to shower and reporting difficulty reattaching it correctly), and misunderstanding of study instructions (e.g., participants missing or ignoring the instruction to wear their monitors overnight).

To examine whether the current analytic sample significantly differed from those excluded, comparisons were conducted across several individual demographic (age and ethnicity) and couple-level (relationship length and whether they live together) factors. A series of independent sample t-tests indicated that there were no differences between the analytic and excluded samples based on relationship length (t[107] = −1.36, p = .177), age (women: t[107] = −.533, p = .595; men: t[107] = .157, p = .876), or family income (women: t[102] = .371, p = .711; men: t[99] = −.241, p = .810). Results from chi-square tests also indicated no difference between the analytic sample and excluded sample based on whether or not couples lived together (χ ² [1] = .777, p = .378) or participants’ race/ethnicity (women: χ ² [6] = 8.010, p = .237; men: χ ² [5] = 5.303, p = .380).

On average, couples in the current analytic sample had been dating for slightly over 2 years (M = 28.09 months; SD = 22.82; range: 2–104 months) and almost half (48%) of the couples lived together. Additionally, 75% of the couples slept in the same bed the night of data collection. Participants were ethnically diverse, with 11.9% African American, 22.2% Hispanic, 33.3% White/non-Hispanic, 13.5% Asian, and 18.3% Multiracial. More than half of the participants (54.0%) were full- or part-time students; 69.8% were employed full or part time.

Procedure

Informed consent was obtained from all individual participants included in the study prior to the initiation of study procedures. Couples were asked to select a day that they would be spending at least 5 hr together. On the morning (typically at 10:00 am) of their chosen day, couples came to the lab and were equipped with wearable physiological monitors. Each partner also received a smartphone programmed to alert them to take mobile surveys every hour, and couples received surveys on the same hourly schedule. They were instructed to complete the surveys separately and not to discuss any answers with one another. In addition to asking couples about their feelings of annoyance with and closeness to their partners, these surveys also asked about general stress levels, other emotions (e.g., happiness and anger), and contact with partners (e.g., text and whether they were physically together). Physiological input was recorded continuously for the duration of data collection (approximately 24 hr), and surveys were completed every hour for all waking hours that day. Number of surveys completed across the day ranged from 4 to 17 for women (M = 12.70, SD = 2.16) and 3 to 16 for men (M = 12.89, SD = 1.96), and participants, on average, responded to 89% of sent surveys. All procedures were approved by the University of Southern California Institutional Review Board. Additional recruitment and procedural details can be found in recent publications (see [28]).

Equipment

Smartphones

Each member of each couple was lent a Nexus 5 mobile phone to complete hourly electronic surveys.

Actiwave

The Actiwave is an ambulatory monitoring device containing a physiological sensor to record electrocardiogram (ECG) signals, time, and movement [29]. The device was attached to participants’ chests using two stick-on electrodes. Given battery life considerations, the sampling rate was set at 32 Hz [30].

Measures

Daytime closeness and annoyance

Every waking hour, couples completed brief surveys about their feelings and experiences over the past hour (M_{surveyscompleted} = 14.38, SD = 1.41). These surveys were completed starting in the morning (i.e., after their lab visit, typically at 10:00 am) and until participants went to bed in the evening. To capture perceived emotional closeness, participants were asked, “In the last hour, how close or connected did you feel towards your romantic partner?” To capture perceived annoyance, participants were asked, “In the last hour, how irritated or annoyed did you feel towards your romantic partner?” Responses were rated on a scale from 0 (not at all) to 100 (extremely). Each partner’s hourly ratings were separately aggregated across the day to create daytime averages of each participant’s perceived closeness to and annoyance with their partner.

Nocturnal heart rate

Matlab scripts were used to develop an automated algorithm that detected artifacts (e.g., movement) in the heart rate data. The automated algorithm performed an anomaly detection technique to identify parts of the signal with abnormal variations, mostly caused by confounding factors related to movement [31]. Subsequently, human annotators were provided with a detailed set of instructions to inspect the ECG signals superimposed with the detected artifact locations, as well as to identify any artifacts that might have been missed by the algorithm. After artifact removal, nocturnal heart rate was calculated by first averaging the heart rate signals across 5 min periods during the time that participants reported being asleep. These 5 min averages were then aggregated to obtain one score for each person’s overall nocturnal heart rate, which was used as the main outcome variable in analyses.

Covariates

Given that young adults’ nocturnal heart rate may be sensitive to a number of other factors, participants provided hourly reports of their physical activity (e.g., exercise), drug and alcohol use, and caffeine intake (all rated on dichotomous “yes/no” scale). They also provided hourly ratings of their stress level (ranging from 0 to 100). Heart rate was measured continuously throughout the day and cleaned using the same procedures as described above with nocturnal heart rate data. Additionally, in a lab session prior to the 24 hr data collection, participants reported their age, relationship length, and cohabitation status. During a poststudy interview, participants reported whether or not they slept in the same bed the evening of data collection.

Analysis Plan

Preliminary analyses tested for differences in nocturnal heart rate as a function of potentially influential daytime behaviors (e.g., drug use) and relationship factors (e.g., relationship length) variables. Next, correlations and descriptive statistics (means and SDs) are presented for the main variables of interest.

The main analyses use actor–partner interdependence models (APIMs) to test how participants’ own daytime relationship feelings relate to their own overnight heart rate (actor effects) and how participants’ partners’ daytime relationship feelings relate to their own overnight heart rate (partner effects) while controlling for both actor and partner effects of daytime heart rate. Because data from romantic partners tend to be correlated, the APIM adjusts for this nonindependence, modeling actor and partner effects simultaneously while accounting for shared within-couple variance [32]. Two APIMs were tested corresponding to the two daytime relationship variables, closeness and annoyance, that were hypothesized to predict nighttime heart rate. Tests of model constraints were used to determine whether female and male participants’ actor and partner paths significantly differed from one another. In order to retain the most parsimonious model, paths that did not significantly vary across partners were constrained to equality.

Finally, we conducted several follow-up analyses that were suggested by reviewers. These analyses considered whether results held when accounting for couple relationship characteristics and daytime stress levels, examined the independent versus overlapping effects of daytime closeness and annoyance on overnight heart rate, and tested whether perceived closeness and annoyance may interactively predict overnight heart rate. All the aforementioned analyses were conducted in Mplus using full information maximum likelihood estimation.

Results

Preliminary Analyses

Given potential confounding factors that may predict nocturnal heart rate, preliminary analyses were conducted to test gender-specific heart rate differences as a function of a range of covariates. There were no significant differences in nocturnal heart rate as a function of daytime physical activity, drug and alcohol use, caffeine consumption, or perceived stress. Additionally, nocturnal heart rate did not significantly differ as a function of individuals’ age, whether couples live together, and whether couples slept in the same bed. Daytime heart rate was significantly associated with nocturnal heart rate for both women (r = .713, p < .001) and men (r = .685, p < .001), and it was, therefore, included as a covariate in all subsequent analyses.

In supplemental analyses suggested by a reviewer, we also examined associations between potential covariates and nocturnal heart rate across the full sample (i.e., not stratified by gender). The only new significant correlation that emerged in the full sample was between daytime caffeine intake and nocturnal heart rate. However, because the results of the actor–partner models were consistent regardless of whether caffeine intake was included as a covariate, we present the most parsimonious models that do not include caffeine intake as a covariate.

Descriptive statistics and bivariate correlations for the main study variables are presented in Table 1. Tests of nonnormality indicated that all variables fell within an acceptable range of skewness (<2) and kurtosis (<7) [33]. Women and men who felt closer to their partners also felt less annoyed with their partners during the day. Female and male partners’ average perceptions of annoyance, but not closeness, were also correlated. Women’s closeness was related to their male partners’ daytime heart rate, and both women’s annoyance and closeness were related to their male partners’ nighttime heart rate. However, men’s daytime closeness and annoyance were unrelated to women’s daytime or nighttime heart rate. There were also several gender differences in the main variable means. Although women and men reported similar levels of closeness across the day, women reported feeling more annoyed with their partners throughout the day and also exhibited higher overnight heart rate compared to men.

Table 1.

Bivariate correlations and descriptive statistics for main study variables

	1.	2.	3.	4.	5.	6.	7.	8.
1. Women’s daytime closeness	–
2. Men’s daytime closeness	.230	–
3. Women’s daytime annoyance	−.286*	−.149	–
4. Men’s daytime annoyance	−.181	−.337**	.620***	–
5. Women’s daytime HR	.122	.040	−.077	−.197	–
6. Men’s daytime HR	−.274*	−.063	.181	−.007	.205	–
7. Women’s nocturnal HR	.079	.123	.124	−.099	.713***	.299*	–
8. Men’s nocturnal HR	−.410**	−.080	.361**	.104	.189	.685***	.279*	–
M (SD)	67.95 (20.96)	70.27 (21.12)	8.58 (8.95)	6.04 (6.65)	71.26 (7.95)	68.68 (8.43)	57.89 (7.13)	53.05 (7.38)

Bolded means indicate a significant (p < .05) gender difference.

HR, heart rate; SD, standard deviation.

*p < .05; **p < .01, ***p < .001.

Actor–Partner Interdependence Models

To determine appropriate model constraints for the APIMs, Wald tests were used to examine whether actor and partner effects significantly differed across men and women. For the model examining daytime closeness as a predictor of overnight heart rate, there was not a significant difference between women’s and men’s actor effects (χ ² [1] = 0.00, p = .974), but there was a significant difference between women’s and men’s partner effects (χ ² [1] = 8.27, p = .004). For the model examining daytime annoyance as a predictor of overnight heart rate, there was not a significant difference between women’s and men’s actor effects (χ ² [1] = 2.24, p = .135), but there was a significant difference between women’s and men’s partner effects (χ ² [1] = 5.39, p = .020). Therefore, actor effects of closeness and annoyance on overnight heart rate were constrained to be equal across men and women, whereas partner effects of closeness and annoyance on overnight heart rate were allowed to freely vary across gender.

Results from the two APIMs are depicted in Fig. 1. Following recommendations for interpreting results in distinguishable dyad models, all effects are represented as unstandardized coefficients [34]. As seen in Fig. 1A, participants’ own daytime feelings of emotional closeness to their partners were unrelated to their own nocturnal heart rate (b = 0.01, p = .785, 95% confidence interval [CI]: −0.04 to 0.05). However, there was a significant partner effect for men: women’s daytime closeness was negatively related to their male partners’ nocturnal heart rate (b = −0.09, p = .003, 95% CI: −0.16 to −0.03) over and above men’s own daytime closeness and daytime heart rate. In other words, men exhibited lower overnight heart rate when their female partners felt closer and more connected to them during the day. Specifically, for every 10-point increase (out of 100) in women’s perceived closeness during the day, men are expected to show a 0.9 beats per minute (BPM) decrease in overnight heart rate, and for every SD increase in women’s closeness, men are expected to show a 1.9 BPM decrease. The APIM exhibited excellent overall fit: χ ² (1) = 0.001, p = .974, Comparative Fit Index (CFI) = 1.00, Root Mean Square Error of Approximation (RMSEA) < .001, Standardized Root Mean Residual (SRMR) = −.001.

Fig. 1.

Unstandardized estimates from actor–partner interdependence models predicting nocturnal heart rate from daytime (A) closeness and (B) annoyance. Solid lines indicate statistically significant pathways and dotted lines indicated statistically nonsignificant pathways. Both models control for men’s and women’s daytime heart rate. *p < .05; **p < .01; ***p < .001. HR, heart rate; r, correlation; e, residual.

A similar pattern was documented when daytime annoyance was the focal predictor. As seen in Fig. 1B, women’s and men’s own daytime feelings of annoyance with their partners were unrelated to their own nocturnal heart rate (b = 0.10, p = .179, 95% CI: −0.05 to 0.24). However, there was a significant partner effect for men such that women’s daytime annoyance was positively related to their male partners’ nocturnal heart rate (b = 0.17, p = .032, 95% CI: 0.02 to 0.33) over and above men’s own daytime annoyance and daytime heart rate. In other words, men exhibited heightened overnight heart rate when their female partners felt more annoyed and irritated with them during the day. Specifically, for every 10-point increase (out of 100) in women’s perceived annoyance during the day, men are expected to show a 1.7 BPM increase in overnight heart rate, and for every SD increase in women’s annoyance, men are expected to show a 1.5 BPM increase. The APIM exhibited adequate overall fit: χ ² [1] = 2.21, p = .137, CFI = .99, RMSEA = .14, SRMR = .02.

Follow-Up Analyses

A series of exploratory follow-up analyses were conducted to examine (a) whether the results from APIMs held when accounting for couple relationship characteristics; (b) whether the results from APIMs held when accounting for each partner’s level of daytime stress; (c) whether women’s daytime closeness and annoyance were independently predictive of men’s overnight heart rate; and (d) whether women’s daytime closeness and annoyance interactively predicted men’s overnight heart rate.

The two APIMs were rerun while accounting for relationship characteristics, specifically whether partners cohabitated and whether partners slept in the same bed the night of data collection. All results presented above held even after accounting for relationship characteristics. That is, higher levels of women’s daytime closeness predicted lower levels of men’s overnight heart rate (b = −0.09, p = .004) and higher levels of women’s daytime annoyance predicted higher levels of men’s overnight heart rate (b = 0.20, p = .023). Additionally, there was evidence that women who slept in the same bed as their partner had higher overnight heart rate (b = 3.47, p = .026 for the closeness model and b = 3.10, p = .052 for the annoyance model). Neither cosleeping nor cohabitating moderated the documented partner effects.

The two APIMs were also rerun while accounting for each partner’s daytime stress, which was computed as the mean of each participant’s hourly reports of stress level throughout the day. All results presented above held even after accounting for actor and partner effects of daytime stress level. That is, higher levels of women’s daytime closeness predicted lower levels of men’s overnight heart rate (b = −0.11, p = .001) and higher levels of women’s daytime annoyance predicted higher levels of men’s overnight heart rate (b = 0.22, p = .022). There were no significant actor or partner effects of daytime stress on overnight heart rate for men or women. Thus, the effects of women’s closeness and annoyance on men’s overnight heart rate persisted even when accounting for both partners’ stress levels throughout the day.

When modeling women’s closeness and annoyance as simultaneous predictors (i.e., in the same model) of men’s overnight heart rate, women’s perceived closeness continued to predict men’s overnight heart rate such that higher levels of female closeness predicted lower levels of male overnight heart rate (b = −0.08, p = .013). However, the effect of women’s daytime annoyance on men’s overnight heart rate was no longer significant (b = 0.13, p = .092).

When modeling women’s closeness and annoyance as interactive predictors, we found a significant two-way interaction (b = −0.01, p = .002). Specifically, higher levels of women’s daytime annoyance predict higher levels of men’s overnight heart rate when women also report lower closeness to their partners (−1 SD, b = 0.219, p = .005) but not when women report higher closeness to their partners (+1 SD, b = −0.155, p = .190).

Discussion

The current study captured 24 hr in the lives of young dating couples to examine short-term associations between relationship functioning and cardiovascular activity. Findings revealed that men’s overnight heart rate varied as a function of their partner’s, but not their own, daytime feelings of closeness and annoyance. Specifically, women’s lower perceived closeness to and greater feelings of annoyance with their partners were each related to higher subsequent heart rate in male partners even after accounting for daytime heart rate. Moreover, supplemental analyses suggested that these relational factors predicted men’s overnight physiological arousal over and above the effects of general perceived stress and that women’s annoyance only predicted men’s elevated heart rate in the absence of perceived closeness. These findings highlight one potential pathway from romantic relationships to health and suggest particular physiological sensitivity to partners among men. The findings also extend prior research typically focusing on married adults by underscoring the interplay of emotions and physiology in young dating couples’ everyday lives. Although several studies have examined patterns of relationship functioning and physiological reactivity among young dating couples (e.g., 35, 36), to our knowledge, this is one of the first studies to investigate such processes using intensive sampling methods within a naturalistic setting.

Moreover, these findings extend past research suggesting that romantic partners can influence one another’s physiology, even over a short period of time [22, 23]. When women felt closer and more connected to their partners throughout the day, men showed attenuated cardiovascular activity overnight. In contrast, when women felt more irritated with their partners, men showed elevated cardiovascular activity overnight. This is consistent with recent research among adult couples in laboratory settings: one study with married couples found that men had greater cortisol (i.e., stress hormone) output when their female partners were generally more depressed and less satisfied with the relationship [22]. The fact that similar patterns are documented here, even over the course of a single “day in the life” of young couples, suggests a subtle process through which women’s relationship feelings may spill over into men’s physiological reactivity.

Although underlying mechanisms accounting for these gender-specific effects were not directly tested, there are several potential explanations. One possibility is that these patterns reflect gendered socialization processes, wherein women typically learn to be more emotionally expressive in relationships [37] and, thus, may exert a greater influence on their male partner’s feelings, behaviors, and even physiological patterns. In terms of closeness, women may more readily express feelings of intimacy through initiating physical touch (e.g., handholding; 38), which, in turn, dampens men’s physiological reactivity [39, 40]. Indeed, intimate physical touch, particularly in the context of stress, has been shown to reduce couples’ alpha amylase (i.e., stress hormone) levels and cardiovascular activity (e.g., blood pressure) [41]. Women may communicate annoyance, on the other hand, through direct (e.g., insults) or indirect (e.g., cold shoulder) expressions that elicit men’s distress and corresponding physiological arousal. For example, women are more likely to use relationally aggressive tactics, such as giving the silent treatment or flirting with others, when they feel upset with their romantic partners [42]. Relatedly, our follow-up analyses suggest that the effect of women’s annoyance on men’s overnight heart rate varied as a function of women’s closeness—it was specifically the combination of women feeling irritated with their partners and a lack of closeness to their partners that predicted men’s elevated overnight heart rate.

Alternatively, men and women may express their feelings of closeness and annoyance similarly but vary in their reactions to, or interpretations of, their partners’ feelings. Past research with married couples found that husbands are more accurate than wives at “decoding” their partner’s positive messages [43] and that men are more likely to feel flooded (i.e., overwhelmed) by their partner’s negative affect [44]. Thus, men may be especially attuned to even subtle expressions of their female partners’ feelings. Future research that incorporates additional physiological indicators (e.g., salivary biomarkers of acute stress) and gathers more detailed emotional feedback from participants (e.g., feelings of being overwhelmed or hurt by partner) will be important for shedding light on these hypothesized mechanisms.

Contrary to hypotheses, neither men’s nor women’s own feelings of annoyance or closeness during the day predicted their own overnight heart rate. This finding is inconsistent with prior longitudinal research demonstrating that people’s perceptions of their relationship quality predict their own health [5–7]. However, very little past work considers if and how relationship dynamics unfold within a single day or among young adults in dating relationships. Recent research examining concurrent associations between relationship satisfaction and physiological reactivity similarly found an absence of actor effects [22]. From a developmental perspective, the intimacy and commitment of romantic relationships are still relatively novel for young adults [26], and there may be heightened sensitivity to the other person’s emotions and experience over and above one’s own ongoing feelings. Indeed, many adolescents and young adults express concerns about relationship dissolution [45] and may closely monitor their partner’s feelings in the service of relationship maintenance.

The current results should be interpreted while keeping in mind several limitations. First, given the intensive nature of data collection and some challenges with compliance for the overnight physiological data collection, the sample size was relatively small. Second, although the study’s focus on 24 hr “in the wild” was a unique methodological strength, the restricted time frame may also limit generalizability to couples’ typical everyday lives and did not allow us to control for prior-night heart rate. Third, although we documented links between women’s daytime annoyance and men’s overnight heart rate, it should be noted that participants’ average levels of annoyance were quite low. Nonetheless, these results show that even women’s relatively mild, small-scale feelings of annoyance with their partners, rather than intense irritation, are linked to men’s nighttime heart rate. Fourth, aside from overnight heart rate, we did not collect data relating to participants’ sleep quality, such as nighttime awakenings or difficulty falling/staying asleep. Insofar as these factors may affect overnight physiological arousal, it will be important for future research to pair physiological monitoring with actigraphy approaches to understand synergies between sleep and cardiovascular functioning. Finally, despite the aforementioned speculations about why women’s relationship feelings predicted men’s overnight heart rate, potential mechanisms were not directly tested. Understanding the day-to-day or moment-to-moment pathways through which early-stage relationships affect physiological functioning is an area ripe for future research on mechanistic processes.

Despite these limitations, the current study incorporated a novel, multimethod approach to capture links between relationship functioning and physiological arousal in everyday life. The findings highlight how young adults’ in-the-moment romantic feelings can acutely map onto their partners’ cardiovascular functioning, even in the absence of concurrent interaction (i.e., during sleep). Although we only focused on couples’ experiences and physiology across one day and night, the patterns we document here, if maintained over time, could cumulatively contribute to young people’s health in meaningful ways. Whereas negative patterns of couple functioning may amplify physiological arousal and increase risk for cardiovascular problems over time, high levels of closeness and intimacy may serve a health-promoting function by regularly attenuating arousal and promoting restorative processes overnight. By examining links between couple functioning and physiological arousal in a young adult sample, this study underscores that connections between relationships and health may begin early and in seemingly subtle ways. Identifying couple processes, such as closeness, that can temper rather than amplify physiological arousal may offer one avenue for capitalizing on intimate relationships as a context for health promotion.

Compliance with Ethical Standards

Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards

The authors declare that they have no conflict of interest.

Authors’ Contributions

H.L.S. conceived of the research questions; A.C.T., S.C.H., and G.M. designed the study; C.P., Y.K., S.S., A.C.T., T.C., and S.C.H. processed and cleaned the data; H.L.S. and C.P. analyzed the data; H.L.S. wrote the article; C.P., Y.K., S.S., A.C.T., T.C., S.C.H., and G.M. reviewed and edited the article.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.