Evidence for a role of the oxytocin system, indexed by
genetic variation in CD38, in the social bonding
effects of expressed gratitude
Sara B. Algoe
* and Baldwin M. Way
Department of Psychology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA, and
Department of Psychology and the
Ohio State Institute for Behavioral Medicine Research, The Ohio State University, Columbus, Ohio 43210
Oxytocin is thought to play a central role in promoting close social bonds via influence on social interactions. The current investigation targeted
interactions involving expressed gratitude between members of romantic relationships because recent evidence suggests gratitude and its expression
provides behavioral and psychological glue to bind individuals closer together. Specifically, we took a genetic approach to test the hypothesis that
social interactions involving expressed gratitude would be associated with variation in a gene, CD38, which has been shown to affect oxytocin secretion.
A polymorphism (rs6449182) that affects CD38 expression was significantly associated with global relationship satisfaction, perceived partner respon-
siveness and positive emotions (particularly love) after lab-based interactions, observed behavioral expression of gratitude toward a romantic partner in
the lab, and frequency of expressed gratitude in daily life. A separate polymorphism in CD38 (rs3796863) previously associated with plasma oxytocin
levels and social engagement was also associated with perceived responsiveness in the benefactor after an expression of gratitude. The combined
influence of the two polymorphisms was associated with a broad range of gratitude-related behaviors and feelings. The consistent pattern of findings
suggests that the oxytocin system is associated with solidifying the glue that binds adults into meaningful and important relationships.
Keywords: close relationships; gratitude; emotion expression; CD38; oxytocin
For centuries, scholars have pondered the nature of the ties that bind
people together. In the 1950s, psychologists Harlow (1958) and Bowlby
(1958) emphasized the necessity of maternal emotional warmth and
comfort for healthy child development. More recently, theorists have
posited that the biological, behavioral and psychological systems that
co-evolved in support of parent–child attachments have come to sup-
port close adult relationships as well (e.g. Hazan and Diamond, 2000;
Diamond, 2004). One approach to illuminating the mechanisms
through which people form and maintain these important relation-
ships has been to shine the empirical spotlight on common social
interactions that occur within them. Recent evidence strongly impli-
cates the emotion and expression of gratitude in the adult bonding
process (Algoe, 2012;Algoe et al., 2013). In this study, we test whether
oxytocin may support the behavior and acute psychological response
surrounding the expression of gratitude in real time. We do this by
using a genetic indicator of oxytocin secretion, variation in the CD38
gene, and by creating a meaningful psychological context in which
oxytocin is likely to reveal its influence: live interactions between mem-
bers of ongoing romantic relationships.
THE BONDING EFFECTS OF GRATITUDE
Recent work regarding the emotion of gratitude supports the theory
that this momentary emotional response to someone else’s kind
gesture has evolved to connect people more closely to others who
would make high-quality relationship partners (see review in Algoe,
2012). Earlier models of gratitude’s role in social life focused on eco-
nomic factors like costs to the benefit-provider (i.e. benefactor), value
to the recipient and repayment of the benefactor (e.g. Trivers, 1971;
McCullough et al., 2001). However, the most recent evidence clarifies
that, rather than merely facilitating interpersonal accounting of re-
sources, gratitude works through a particular relational currency that
is associated with intimacy, perceived partner responsiveness (Algoe
et al., 2008). Perceived partner responsiveness underlies relational in-
timacy and is associated with feeling understood, accepted and cared
for (Reis et al., 2004); it is an important situational trigger for feeling
gratitude upon receipt of a benefit (Algoe et al., 2008;Algoe and
Stanton, 2012). In turn, the grateful individual is likely to demonstrate
responsiveness for the benefactor’s own needs and preferences in the
future (e.g. Kubacka et al., 2011). In short, the positive emotion of
gratitude provides fuel for upward spirals of mutual responsiveness
between dyad members, thereby promoting the quality of the relation-
ship over time.
In light of these strong connections between experienced emotion
and relational outcomes, more recent work has begun to focus in on
how one person’s experienced gratitude might translate to relational
growth for each dyad member. One candidate behavior that follows
from emotion theory is the expression of gratitude; indeed, in situations
that would cause gratitude, acknowledging a benefit is the most fre-
quently reported motive of the benefit recipient (Algoe and Haidt,
2009). Recent studies provide suggestive evidence that, for the grateful
individual, expressing gratitude is necessary to reap the relational bene-
fits (e.g. Lambert et al., 2010;Algoe and Stanton, 2012). For example,
participants who were randomly assigned to ‘go the extra mile’ to
express gratitude to a friend over the course of 3 weeks reported greater
communal relationship strength with that friend by the end of the
study, even compared with a condition in which people were assigned
to focus on thinking about things they appreciate about the friend over
the 3-week period (Lambert et al., 2010).
Received 20 May 2013; Revised 17 September 2013; Accepted 28 December 2013
Advance Access publication 5 January 2014
*These authors contributed equally to this work.
This work was supported by a grant from the National Institute of Mental Health (MH59615) and a grant from
the National Center for Advancing Translational Sciences (8KL2TR000112-05). The content is solely the responsibility
of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors
would like to thank Ms Sara DeMaria for her excellent technical assistance, as well as the exceptional team of 30
research assistants who helped with the Carolina Couples Study, lead by Jenny Bridgers (full list available on first
author’s website). Barbara Fredrickson, B. Keith Payne and Crystal Schiller provided insightful feedback on an earlier
Correspondence should be addressed to Sara B. Algoe, Department of Psychology, University of North Carolina at
Chapel Hill, CB #3270, Chapel Hill, NC 27599, USA. E-mail: email@example.com
doi:10.1093/scan/nst182 SCAN (2014) 9, 1855 ^1861
ßTheAuthor (2014). Published by Oxford University Press. For Permissions, please email: firstname.lastname@example.org
Beyond the grateful person, though, because of gratitude’s ‘other-
praising’ nature (Algoe and Haidt, 2009), expressed gratitude may
provide psychological rewards to the target of the expression (i.e. the
original benefactor) that act as a hook to keep the benefactor engaged
in and satisfied with the relationship with the grateful individual.
Critically, this is most likely to be true if the expresser is perceived
to be responsive when thanking the original benefactor (Algoe et al.,
2013). Specifically, a recent study used a naturalistic interaction in
which one romantic partner expressed gratitude to the other in the
lab; for targets of the expressions who perceived their romantic partner
as being particularly responsivethat is, understanding, validating and
caring (Reis et al., 2004)when expressing gratitude to them, the tar-
get’s satisfaction with the relationship improved over the subsequent
6 months (Algoe et al., 2013). In fact, effects of perceived partner
responsiveness after the partner expressed gratitude to the target in
the lab even held when statistically controlling for perceptions of the
partner’s responsiveness after participating in two different types of
common couple interactions that have been documented to be im-
portant for relationship well-being (i.e. providing support for positive
or negative events; Gable et al., 2006). These findings highlight the
unique and powerful role that expressions of gratitude can play in
social bonding, beyond the other types of behaviors that may foster
relationship quality. Identifying potential biological contributors to
these effects stands to open doors for improved understanding of
social bonding processes.
CLOSE RELATIONSHIPS AND OXYTOCIN
Since the seminal discovery that oxytocin facilitates monogamous
bonds in rodents (Williams et al., 1994), a rapidly growing body of
research in humans has sought to determine if oxytocin is involved in
human social connections, broadly defined. Accordingly, oxytocin has
been associated with everything from parenting behavior (e.g. Feldman
et al., 2012) to behavior toward a new acquaintance (e.g. Kosfeld et al.,
2005). However, for the question of close social bonds, it is notable
that the adult romantic relationship context is unique in its qualitative
experience as well as its impact on health and well-being (e.g.
Berscheid, 1999;Hazan and Diamond, 2000). Because the connection
between oxytocin and psychosocial processes appears to be highly de-
pendent on the behavioral and relational context (see also Bartz et al.,
2011), it is important to study psychosocial correlates of oxytocin
directly in the romantic relationship context.
Within this context, the oxytocin system has been hypothesized to
both facilitate bonding processes and change in response to social
interactions (see Carter and Porges, 2013). In this work, we focus on
the former type of research question, regarding the facilitation of
bonding (for examples of the latter, see Grewen et al., 2005;Holt-
Lunstad et al., 2008;Smith et al., 2013). The few studies that have
addressed whether or how oxytocin might facilitate bonding processes
in the context of romantic relationships do show concurrent associ-
ations between the oxytocin system and a handful of relationship-
relevant behaviors, as observed in well-defined laboratory interactions
between couple-members. Specifically, intranasal administration of
oxytocin led to an increase in positive communication during a con-
flict discussion (Ditzen et al., 2009); similarly, higher levels of plasma
oxytocin have been associated with greater positive behaviors during a
conversation involving one person’s expression of worry or concern
(Gouin et al., 2010), as well as while discussing a shared positive ex-
perience (Schneiderman et al., 2012). Integrating across these diverse
methodologies, these studies suggest that higher synaptic levels of oxy-
tocin promote relationship building behaviors. Finally, variation in the
oxytocin receptor gene (OXTR), which theoretically alters the impact
of oxytocin signaling, was associated with greater use of affiliative cues
during a partner’s expression of worry (Kogan et al., 2011). These
studies suggest that in general, the oxytocin system is involved in re-
lationship-relevant processes. However, the mixture of methods used
to measure or manipulate the oxytocin system combined with theory
and evidence from relationship science suggesting that the various re-
lationship behaviors investigated across these studies play different or
unique roles in the promotion, maintenance and prevention-of-deteri-
oration of the relationship (see Gable and Reis, 2001;Gable et al., 2012;
McNulty and Fincham, 2012), mean it is too early to draw broad
conclusions about oxytocin’s role in facilitating processes central to
the promotion of close relationships.
In short, there is much more to
As reviewed, within the adult close relationship context, there has
been surprisingly little examination of early assertions (Uvnas-Moberg,
1996,1998;Carter, 1998) regarding oxytocin’s role in facilitating
human social bondsassertions which have been recently recapitulated
(Carter and Porges, 2013). We believe a focus on social interactions
involving expressed gratitude provides a prime opportunity to test
associations between the oxytocin system and social processes involved
in promoting adult human pair bonds because of the unique role of
expressed gratitude in this process (Algoe et al., 2013). Moreover, we
go beyond the behavior of expressed gratitude to specify some of the
acute psychological impacts of such interactions that should be
particularly instrumental in the bonding process, namely positive
emotions and perceived partner responsiveness.
GENETIC ASSOCIATIONS WITH RELATIONSHIP PROCESSES
As mentioned earlier, prior research suggests that genetic association
methodology is a viable method for gaining clues to potential neuro-
chemical underpinnings of social bonding related processes in roman-
tic couples (Kogan et al., 2011). We adopt a similar approach here.
However, the cellular mechanism explaining how the previously dis-
cussed OXTR genetic variant could impact cellular signaling has not
yet been identified, and a recent meta-analysis reported no association
with a broad spectrum of socially relevant behaviors (Bakermans-
Kranenburg and van IJzendoorn, 2014). Instead, we studied functional
variation in a different gene, CD38, which is a key regulator of oxytocin
release. Mice with deletion of the CD38 gene exhibit marked reduc-
tions of oxytocin within the cerebrospinal fluid and the plasma (Jin
et al., 2007), indicating that CD38 is necessary for oxytocin secretion.
Unable to release oxytocin, these mice exhibit profound deficits in
basic social processes such as social recognition and memory (Jin
et al., 2007).
In humans, CD38 appears to have a role in oyxtocin signaling and
social processes as well. Levels of CD38 gene expression in peripheral
cells are positively correlated with plasma levels of oxytocin (Kiss et al.,
2011), suggesting that CD38 facilitates oxytocin release. Commonly
occurring variation in CD38 (rs3796863) has also been associated
with plasma oyxtocin levels (Feldman et al., 2012). Therefore, CD38
is an important regulator of oxytocin signaling.
Consistent with a role in regulating oxytocin release, CD38 is also
associated with social processes. Accordingly, levels of peripheral CD38
gene expression are related to clinician-observed social skills in autistic
patients (Kiss et al., 2011;Riebold et al., 2011). The previously men-
tioned single nucleotide polymorphism (SNP) (rs3796863) has also
To add complexity, another body of work tests concurrent associations between plasma oxytocin and self-reported
assessment of global relationship evaluation rather than specific social behavior or its immediate impact in a live
and meaningful context. Presumably, the quality of specific interpersonal interactions adds up to comprise the
overall evaluation of the relationship. Yet in this line of work, researchers sometimes find negative associations
between oxytocin and global relationship evaluation (e.g. Taylor, Saphire-Bernstein, & Seeman, 2010;Smith et al.,
2013). Though not the primary focus of the current investigation, we include a measure of global relationship
1856 SCAN (2014) S. B. Algoe and B. M.Way
been associated with either low functioning autism (Lerer et al., 2010)
or high functioning autism (Munesue et al., 2010). In the non-clinical
context, a study of parenting behavior showed that rs3796863 was
associated with reduced parental touch of infants (Feldman et al.,
2012), suggesting that CD38 affects social engagement. Thus, there is
an emerging literature tying rs3796863 to social processes. However,
the molecular pathway by which this SNP could affect CD38 expres-
sion or function has not been identified.
A different polymorphism in CD38 that is located at the opposite
end (50) of the CD38 gene from rs3796863 has a more clearly deli-
neated molecular pathway by which it can affect CD38 expression.
This SNP, rs6449182, is located in intron 1, which is a regulatory
control region of the CD38 gene (Ferrero et al., 1999) and differentially
affects transcription factor binding (Saborit-Villarroya et al., 2011).
Accordingly, this polymorphism has been associated with differences
in relative levels of CD38 messenger RNA (Jamroziak et al., 2009),
protein (Jamroziak et al., 2009) and enzymatic activity (Polzonetti
et al., 2012). Thus, rs6449182 has functional effects on expression of
CD38 and, by extension, extracellular oxytocin levels in the brain and
plasma. Although the mechanism by which rs6449182 affects CD38
function is better delineated than for rs3796863, it has been less studied
in a psychological context. Therefore in this study, we examine the
association of rs6449182 as well as rs3796863 with gratitude and its
social bonding effects. Because these two SNPs are located in different
portions of the CD38 gene, and genotypes at each locus are potentially
independent (meaning that knowledge of one genotype does not ne-
cessarily predict the other), we also studied their combined effect by
summing genotypes to create a putative index of CD38 gene expression
(see Chapman et al., 2003 for this approach).
The current investigation focuses on whether variation in CD38 is
associated with social interactions involving the expression of grati-
tude. Evidence for the potential relational benefits from expressing
gratitude (e.g. Lambert et al., 2010) as well as receiving an expression
of gratitude (Grant and Gino, 2010;Algoe et al., 2013) suggests that the
oxytocin system may influence both social roles: the expresser of grati-
tude and the recipient of gratitude expression. Moreover, we specify
the relational and affective consequences of such interactions that are
most closely related with high-quality, close, social connections: per-
ceived partner responsiveness and positive emotions (see Fredrickson,
1998 for theory and Kok et al., 2013 for the role of positive emotions in
building social resources, over time). In addition, work that differen-
tiates among positive emotions in momentary social function (e.g.
Algoe and Haidt, 2009;Shiota et al., 2011) suggests that it is worth-
while to further explore whether a specific active emotional ingredient
in the romantic bonding contextloveis relevant (and see Gonzaga
et al., 2006 for links between romantic love and oxytocin). Thus, after
receiving an expression of gratitude and after expressing gratitude, we
examined perceptions of partner responsiveness as well as positive and
negative emotions experienced; we also examined grateful behavior
itself, observed in the lab and as gathered from daily self-reports.
Each member of 77 heterosexual romantic couples participated in a
study on ‘Everyday experiences and feelings of people in romantic
relationships’ (aka the Carolina Couples Study; see additional infor-
mation in Algoe et al., 2013). Couples were recruited from the region
surrounding Chapel Hill, North Carolina, and must have been roman-
tically involved for a minimum of 6 months. The current sample con-
sists of 128 individuals who provided saliva for genotyping (83.1% of
sample; 63 women, 65 men across 69 heterosexual couples; aged 18–57,
M¼28.5, s.d.¼8.65). The majority of participants self-identified as
White/Caucasian (76.2%), and the remaining participants were Black/
African-American (11.9%), East or South Asian (4%), and multi-racial
or of an unidentified racial category (7.9%); 3.2% identified as
Hispanic. Couples had been together for about 4 years (M¼4.2
years), and were either dating exclusively (55.8%) or committed for
life (i.e. married, engaged to be married or living as married; 44.2%).
Overview of design and procedure
The larger Carolina Couples Study was an observational investigation
consisting of two visits to the lab, 2 weeks apart, with brief nightly
questionnaires completed for each of 14 nights between lab visits. The
current investigation focuses on one pair of 5 min interactions between
couple-members at the second lab visit, in which participants were
each given the opportunity to express gratitude to the partner, as
well as nightly self-reported gratitude expression.
Expressed gratitude behavioral task
Each participant was asked to choose something specific, big or small,
that the partner did for him or her, for which the participant felt
grateful (see Algoe et al., 2013). After noting the event, one participant
took up to 5 min to thank the partner, after which they both self-
reported their reactions to the interaction via private online question-
naires. The interaction and self-report were repeated with the second
partner. Whether the male or female participant thanked the partner
first was counterbalanced. Note that each participant in the study was
part of two interactions: one in which he or she expressed gratitude, and
one in which he or she received an expression of gratitude. Responses to
each are important to our investigation because they reflect giving and
receiving of positively valenced, other-focused attention, respectively.
At study entry, each participant reported on global relationship satis-
faction using a well-validated seven-item measure (Hendrick, 1988;
After each interaction, participants rated their agreement with sev-
eral items on scales ranging from ‘not at all true/never true’ (0) to ‘very
true/true all of the time’ (6). First, each participant responded to 24
items to measure emotional response. To reduce demand effects, we
included a broad range of items and information that any given situ-
ation may produce its own set of emotions. The 11 positive emotion
terms (e.g. peaceful, loving, amused, proud) were averaged into a posi-
tive emotions composite (¼0.90, 0.86 after receiving an expression
of gratitude and providing an expression, respectively). Note that the
composite scores of the 13 negative emotion terms had extremely low
ratings after each interaction (M< 0.31), consistent with the nature of
the interactions, and we did not consider them further. Next, each
participant responded to 10 items to measure the core components
of perceived responsiveness (Gable et al., 2006), which include under-
standing, validating and caring (10-item composite ¼0.93, 0.95 after
receiving an expression of gratitude and providing an expression, re-
spectively). Data analysis included three measures both after expressing
gratitude and after receiving an expression of gratitude: composite
scores of perceived partner responsiveness and positive emotions,
and, to probe specificity regarding theory, one targeted positive emo-
tion term, ‘loving’.
Grateful behavior was indexed by a participant’s degree of praising
of the partner’s actions within the expression of gratitude (for theory,
see Algoe and Haidt, 2009), and encompassed both verbal and non-
verbal behaviors to guard against counting insincere praise. The be-
havior was reliably observed by four judges who rated the degree of
Expressed gratitude and CD38 SCAN (2014) 1857
praising on a scale ranging from 1 (no or little praise) to 5 (excellent
degree of benefactor praiseworthiness; ¼0.78), and participants
varied in their spontaneous use of praising behavior when expressing
gratitude, with the average of judges’ scores ranging from 1.25 to 5
Finally, one question from the 14 nightly reports completed prior to
the lab visit measured grateful behavior in a different way: each night,
participants responded to the item, ‘I thanked my partner for some-
thing he/she did that I appreciated’ with a yes or no (coded as 1 and 0
for data analysis). This second measure provides the opportunity to
test convergent and ecological validity of our hypothesis.
DNA was extracted from saliva collected using Oragene kits (DNA
Genotek; www.dnagenotek.com) according to the manufacturer’s in-
structions. Both SNPs were genotyped using Taqman SNP Genotyping
Assays (rs6449182: C___1216863_10; rs3796863: C___1216944_10)
from Applied Biosystems using an ABI 7300 Sequence Detection
System. All samples were successfully genotyped on two separate oc-
casions with complete concordance. Haploview (version 4.2; Barrett
et al., 2005) was used to calculate Hardy–Weinberg Equilibrium values
using the exact test (Wigginton et al., 2005) as well as the normalized
measure of allelic association (D0;Lewontin, 1964) and coefficient of
;Hill and Robertson, 1968). The rs3796863 SNP was
coded in a dominant manner [CC genotype ¼0; A allele carriers (AC
and AA) ¼1] as in prior work (Feldman et al., 2012,2013; Sauer et al.,
2012). The CC genotype, which is present in approximately 40% of the
Caucasian and East Asian populations, has been associated with lower
CD38 expression (Lerer et al., 2010) and lower plasma oxytocin levels
(Feldman et al., 2012). With respect to rs6449182, the G allele is gen-
erally associated with greater CD38 expression, which is reflected
in higher levels of CD38 mRNA, protein, and enzymatic activity
(Jamroziak et al., 2009;Polzonetti et al., 2012; but see Riebold et al.,
2011). This evidence on CD38 expression suggests that the G allele
functions in an additive, not dominant manner, with the GG genotype
(5–10% of the Caucasian population) associated with the greatest
expression. Therefore, the rs6449182 polymorphism was coded in an
additive manner (CC ¼0, CG ¼1; GG ¼2). To create a cumulative
index of putative CD38 expression, the two SNPs were summed
according to their effects on gene expression, as coded earlier.
Table 1 contains descriptive statistics for all dependent measures.
Dependent measures violate assumptions of independence required
for typical ANOVA (analysis of variance) because each member of
the couple provided reports on the same outcomes. Therefore, we
use multilevel modeling (using HLM; Raudenbush et al., 1996) for
all analyses. Dependent measures from lab-based interactions were
tested with two-level models (i.e. person nested within couple), and
daily grateful behavior was tested with a three-level model (i.e. day
within person within couple). (See Supplementary Table S1 containing
correlations among dependent measures as well as intraclass correl-
ations between dyads on each variable.) The distribution of genotypes
for the two SNPs did not deviate from Hardy–Weinberg equilibrium
(Ps > 0.19; for rs6449182- CC n¼83, CG n¼37, GG n¼8; for
rs3796863, CC n¼62, A carriers n¼66). Models were run with
rs6449182, rs3796863, or their combination as a continuous predictor
variable for that person’s own behavior or response to an interaction.
None of the effects reported below was moderated by participant sex,
nor were conclusions altered when controlling for ethnicity.
Similarity of rs6449182 and rs3796863
As expected due to their distinct locations in the CD38 gene, there was
a relatively low degree of association between the two SNPs (in the pri-
mary ethnic group: D0¼0.50, 95% CI [0.09, 0.79]; overall: D0¼0.45,
95% CI [0.08, 0.75]; r
¼0.026). In other words, rs6449182 status is
not correlated with rs3796863 status [r(Spearman’s rho) ¼0.004,
The first section of columns in Table 2 presents the results of analyses
for the nine targeted outcomes, showing that rs6449182 was signifi-
cantly associated with eight of the nine. All associations were in the
same direction, indicating that CC individuals had more positive psy-
chological outcomes or behaviors. Specifically, rs6449182 was signifi-
cantly associated with global evaluation of relationship satisfaction
assessed at study entry; after receiving an expression of gratitude,
rs6449182 was associated with the perception the expresser was re-
sponsive as well as with the specific emotion, loving; after providing
an expression of gratitude rs6449182 was associated with the percep-
tion that the benefactor was responsive as well as experienced positive
emotions (generally) and the specific emotion, loving. rs6449182 was
also associated with the behavioral expression of gratitude, either as
observed in the videorecorded lab interactions by external judges or as
reported by the participant across 14 days: the odds ratio demonstrates
that people with the CC genotype were most likely to report thanking
the partner for an appreciated event on any given day (see Figure 1 for
The second section of columns in Table 2 presents the results of ana-
lyses for the nine targeted outcomes, showing that rs3796863 was sig-
nificantly associated with one of the nine. Specifically, after providing
an expression of gratitude, rs3796863 was significantly associated with
the perception that the benefactor was responsive, in the direction
indicating that those with the CC genotype were most likely to perceive
responsiveness in the benefactor after having expressed gratitude to
him or her. In addition, a marginally significant effect emerged for
self-reported thanking behavior across 14 days in the same direction,
such that those with the CC genotype tended to be most likely to
report thanking the partner on a given day. From these rs3796863
analyses, though all coefficients were in the same direction, no other
associations exceeded the 0.05 significance threshold.
Table 1 Descriptive statistics for all dependent measures
Global relationship satisfaction 6.07 0.48
Receiving a gratitude expression
Perceived partner responsiveness 5.06 0.75
Positive emotions 3.78 1.08
Loving 4.92 0.97
Perceived partner responsiveness 5.14 0.79
Positive emotions 4.11 0.91
Loving 5.00 1.01
Praising behavior 3.25 0.54
Daily expressed gratitude 0.79 0.12
Mean and s.d. reflect intercept and standard deviation of r, respectively, in unconditional HLM
models where the listed variable was the dependent variable. For daily expressed gratitude, the
intercept accounts for Bernoulli distribution of the outcome (0 or 1) and standard error (not
deviation) is reported.
18 5 8 S C A N ( 2 014) S. B. Algoe and B. M.Way
Combined CD38 markers as an index of likelihood of
The third section of columns in Table 2 presents results of analyses for
the nine targeted outcomes, showing that putative CD38 expression
was significantly associated with five of the nine (and an additional two
approach statistical significance: relationship satisfaction and percep-
tion that the partner was loving when receiving an expression of grati-
tude). All significant (and non-significant) associations were in the
same direction, indicating that individuals with presumptive lower
CD38 expression had more positive psychological outcomes or behav-
iors. Specifically, after receiving an expression of gratitude, CD38
expression was significantly associated with the perception that the
benefactor was responsive; after providing an expression of gratitude,
the CD38 expression variable was associated with the perception that
the benefactor was responsive as well as experienced positive emotions
(generally) and the specific emotion, loving. Finally, adding ecological
validity, from measures across 14 days, lower CD38 expression was
associated with a greater likelihood of reporting having thanked the
partner for an appreciated event on any given day.
We tested the same pattern of analyses for relational and affective
responses to a different conversation topic: disclosure of a personal
positive event. Such interactions are positive in valence and responses
to them have been associated with global relationship quality (see
Gable et al., 2006 for task description and evidence). However, the
primary positive emotional theme is joy/enthusiasm rather than grati-
tude; research demonstrates socially functional distinctions between
these emotions (Algoe and Haidt, 2009), and that perceived respon-
siveness after a partner expressed gratitude better captured the variance
in growth in the listener’s relationship satisfaction (i.e. promoting
bonds) than did perceived responsiveness of the partner after respond-
ing to a positive event disclosure (Algoe et al., 2013). No self-report
measures after the positive event disclosure interaction (i.e. positive
emotionsincluding lovingor perceived partner responsiveness) was
significantly predicted by CD38 status, either when sharing a positive
event or when hearing a partner share an event (rs6449181
Ps¼0.21–0.98, MP-value ¼0.55; rs3796863 Ps¼0.06–0.87, M
P-value ¼0.60; CD38 expression Ps¼0.09–0.98, MP-value ¼0.60).
This suggests that the results for the gratitude expression are not
solely due to the positivity of the interaction, and is consistent with
theorizing that oxytocin has context-specific effects (Bartz et al., 2011;
Feldman, 2012), with the focus of the current investigation on one
meaningful context thought to actively promote social bonding.
Adult human pair bonds have been posited to be supported by the
co-evolution of biological, behavioral and psychological processes
(e.g. Hazan and Diamond, 2000;Diamond, 2004;Feldman, 2012).
Resting on this assumption, our investigation focused on a putative
marker of oxytocin secretion, the behavioral expression of gratitude
and psychological responses associated with interactions involving
expressions of gratitude, all of which have been closely implicated in
promoting the quality of social bonds.In laboratory interactions be-
tween people in romantic relationships, and from self-reported behav-
ior in daily life, we found that functional genetic variation in CD38 was
associated with the quality and frequency of grateful behavior toward
the partner. In addition, this variation also predicted the psychological
impact of providing or receiving an expression of gratitude, as well as
global relationship satisfaction. Of the two polymorphisms, rs6449182
was associated with a broader spectrum of the building blocks by
which gratitude facilitates social bonds. Moreover, when these
Table 2 Regression coefficients, significance levels and effect sizes from tests of rs6449182, rs3796863 and putative CD38 expression on the nine targeted outcomes
CD38 CD38 CD38
rs6449182 rs3796863 expression
B95% CI Pd B 95% CI Pd B 95% CI Pd
Relationship satisfaction 0.15 0.29, 0.01 0.03 0.39 0.03 0.24, 0.17 0.77 0.05 0.11 0.23, 0.01 0.07 0.33
Receiving a gratitude expression
Perceived responsiveness 0.24 0.45, 0.02 0.03 0.40 0.14 0.45, 0.18 0.39 0.16 0.20 0.37, 0.04 0.02 0.44
Positive emotions 0.22 0.55, 0.11 0.19 0.24 0.19 0.66, 0.28 0.43 0.15 0.20 0.48, 0.08 0.16 0.26
Loving 0.37 0.69, 0.05 0.02 0.42 0.09 0.58, 0.40 0.71 0.07 0.26 0.55, 0.02 0.07 0.33
Perceived responsiveness 0.19 0.35, 0.03 0.02 0.42 0.37 0.68, 0.06 0.02 0.43 0.26 0.41, 0.12 0.001 0.65
Positive emotions 0.31 0.62, 0.001 0.05 0.36 0.30 0.69, 0.09 0.13 0.28 0.31 0.54, 0.07 0.01 0.47
Loving 0.58 0.92, 0.24 0.001 0.62 0.25 0.62, 0.12 0.18 0.24 0.46 0.71, 0.22 0.000 0.69
Observed praising behavior 0.21 0.42, 0.001 0.05 0.39 0.03 0.20, 0.26 0.79 0.05 0.12 0.31, 0.06 0.19 0.26
Daily expressed gratitude 0.70 0.53, 0.91 0.009 n/a 0.73 0.96, 1.94 0.09 n/a 0.70 0.57, 0.88 0.002 n/a
Bis the unstandardized regression coefficient from multilevel models with CD38 status as a predictor of the outcome of interest; 95% CI is the 95% confidence interval; Pis statistical significance level, and the
effect size (absolute value) is Cohen’s d. For daily expressed gratitude, the model accounts for Bernoulli distribution of the outcome (0 or 1), and we therefore present the odds ratio rather than B.
GG CG CC
Percent of Days Thanked Partner
Fig. 1 Illustration of linear trend in the percent of days, across 2 weeks, participants reported
thanking the partner, by rs6449182 polymorphism. Note that raw data are used in this figure for
illustration, but analyses reported in text control for dependence in data across couple-members.
Four dependent measures contained outliers (i.e. > 3 s.d. beyond mean). These analyses were re-run with outliers
removed, using each index of CD38 expression as predictor; perceived responsiveness after receiving an expression
of gratitude became non-significant as predicted by rs6449182 (P¼0.11); all other conclusions remained the same.
Expressed gratitude and CD38 SCAN (2014) 1859
uncorrelated markers of oxytocin release were combined into a puta-
tive index of CD38 expression, five of nine targeted effects remained,
including thanking behavior in everyday life.
This consistent pattern of findings indirectly linking the oxytocin
system with the quality and quantity of expressed gratitude adds to the
small but growing body of evidence regarding the behavioral and psy-
chological mechanisms by which oxytocin may influence human pair
bonds. Recent theory in relationship science postulates that behaviors
associated with prevention of negative outcomes (e.g. fighting well) or
alleviation of negative emotions (e.g. receiving social support) are not
tapping into the same psychological processes as behaviors associated
with the promotion of positive outcomes or emotions (Gable and Reis,
2001). Empirically, negative emotions were very mild in these inter-
actions, whereas positive emotions were relatively high. Consistent
with theory, variations in rs6449182 in particular predicted differential
positive evaluations of the partner and interaction, whether after
providing an expression of gratitude or receiving an expression.
Perceived partner responsiveness and positive emotions (e.g. love)
after either interaction may have been greater due to consideration
of the partner’s positive qualities or the partner’s actual behaviors in
the interaction. Regardless, in so far as CD38 gene expression affects
oxytocin signaling, our results implicate the oxytocin system in the
psychological reactions to expressions of gratitude which serve to
reward the person for remaining in the relationship. As recent work
highlights, it is the acute psychological impact of such interactions that
can forecast change in relationship quality (Algoe et al., 2013).
More broadly, our findings are important because the social roles of
receiving and providing other-directed positive attention, here enacted
through receiving and providing an expression of gratitude, have each
recently been linked with consequential downstream outcomes. First,
perceiving a partner’s responsiveness upon receipt of an expression of
gratitude has been associated with downstream relational health bene-
fits (Algoe et al., 2013). The findings regarding the impact of providing
an expression of gratitude are interesting in light of the growing evi-
dence for the importance of other-focus for mental health as well as
longevity (e.g. Crocker et al., 2009;Konrath et al., 2012). We see our
data, examined at the level of a live social interaction, as providing
important basic research on potential biological, behavioral and psy-
chological mechanisms for such effects.
Methodologically, integrating live social interactions with measures
from daily life, as was done here, represents a valuable path forward for
improving understanding of how oxytocin is related to creating social
bonds. As reviewed in the introduction, a few studies of structured
social interactions in the laboratory generally indicate that oxytocin is
associated with beneficial relationship-relevant behaviors. Yet other
studies that assess a longer relationship time frame have suggested
that oxytocin is associated with distress in the pair-bond relationship
(see Taylor et al., 2010). Resolving such controversies will require not
only studying the same construct, such as gratitude, across multiple
contexts, but also employing different methodologies and manipula-
tions of the oxytocin system to provide convergent validity. In this
vein, the current investigation draws attention to the utility of focusing
on the gene CD38. There is a well-delineated molecular pathway by
which rs6449182 can impact cellular signaling, which stands in con-
trast to commonly studied polymorphisms in the oxytocin receptor
gene, such as rs53576, for which there has not been a clearly identified
molecular mechanism that could explain how the polymorphism in-
fluences signaling. Though the small sample size requires interpret-
ation of these findings as preliminary until replicated, the pattern of
associations across a variety of measures that have long been posited to
be associated with oxytocinother-directed positive behavior, intimacy
and warm positive feelings (cf. Carter, 1998)suggests that CD38 is a
useful marker for future research investigating the role of oxytocin
system in the promotion of social bonds.
In addition to being involved in processes other than oxytocin se-
cretion, a limitation of the genetic approach using CD38 is that because
genetic variants are indirect measures of neurotransmitter signaling, it
is not yet possible to draw conclusions regarding whether the reported
associations with gratitude are due to greater or lesser release of oxy-
tocin in the brain. Mice that do not express CD38 have low extracel-
lular oxytocin levels (Jin et al., 2007). Similarly, humans with low levels
of expression in lymphocytes have lower plasma oxytocin levels (Kiss
et al., 2011). In this study, the genotypes were coded based on previ-
ously reported effects on CD38 expression, suggesting that lower CD38
expression is associated with greater expression of gratitude and
greater resulting psychological effects. However, taking the next step
to infer that the measures in this study are associated with lower levels
of oxytocin release should be made cautiously for several reasons. First,
the studies associating rs6449182 and rs3796863 with CD38 expression
(e.g. Jamroziak et al., 2009) were conducted in cells derived from
peripheral tissues rather than brain tissue. Yet, it is presumably alter-
ations in CD38 expression within the brain that are leading to the
behavioral and psychological effects seen here. Second, the effects of
rs6449182 on gene expression could be influenced by nutritional and
cellular factors that were not measured in this study. Third, there are
likely to be complex compensatory mechanisms that might counteract
the lifelong effects of these polymorphisms. For example, even though
mice with deletion of the CD38 gene have very low levels of oxytocin in
the plasma, they actually have very high levels of oxytocin inside the
cell because the oxytocin is unable to be released without CD38 (Jin
et al., 2007). Therefore, it will be important in future studies to link
these polymorphisms to levels of oxytocin in the cerebrospinal fluid of
humans to gain a better understanding of how they might be related to
oxytocin signaling and how the aforementioned moderators might
impact the effect of rs6449182 on CD38 expression.
In sum, this research presents the first evidence regarding the bio-
logical underpinnings of expressed gratitude in the formation of adult
human bonds. Moreover, CD38 status was associated with behavior
and psychological reactions within live interactions, as well as global
relationship quality. As an index of oxytocin release, the reliable pat-
tern of results for CD38 variants across several distinct and theoretic-
ally consistent measures provides intriguing evidence for the oxytocin
system in solidifying the ‘glue’ that brings close adult relationship
partners closer together.
Supplementary data are available at SCAN online.
Algoe, S.B. (2012). Find, remind, and bind: the functions of gratitude in everyday rela-
tionships. Social and Personality Psychology Compass,6(6), 455–69.
Algoe, S.B., Fredrickson, B.L., Gable, S.L. (2013). The social functions of the emotion of
gratitude via expression. Emotion,13(4), 605–9.
Algoe, S.B., Haidt, J. (2009). Witnessing excellence in action: the “other-praising” emotions
of elevation, gratitude, and admiration. Journal of Positive Psychology,4(2), 105–27.
Algoe, S.B., Haidt, J., Gable, S.L. (2008). Beyond reciprocity: gratitude and relationships in
everyday life. Emotion,8, 425–9.
Algoe, S.B., Stanton, A.L. (2012). Gratitude when it is needed most: social functions of
gratitude in women with metastatic breast cancer. Emotion,12(1), 163–8.
Bakermans-Kranenburg, M.J., van IJzendoorn, M.H. (2014). A sociability gene?
Meta-analysis of oxytocin receptor genotype effects in humans. Psychiatric Genetics,
Barrett, J.C., Fry, B., Maller, J., Daly, M.J. (2005). Haploview: analysis and visualization of
LD and haplotype maps. Bioinformatics,21,263–5.
Bartz, J.A., Zaki, J., Bolger, N., Ochsner, K.N. (2011). Social effects of oxytocin in humans:
context and person matter. Trends in Cognitive Sciences,15(7), 301–9.
18 6 0 S C A N ( 2 014) S. B. Algoe and B. M.Way
Berscheid, E. (1999). The greening of relationship science. American Psychologist,54(4),
Bowlby, J. (1958). The nature of the child’s ties to his mother. International Journal of
Carter, C.S. (1998). Neuroendocrine perspectives on social attachment and love.
Carter, C.S., Porges, S.W. (2103). The biochemistry of love: an oxytocin hypothesis. EMBO
Chapman, J.M., Cooper, J.D., Todd, J.A., Clayton, D.G. (2003). Detecting disease associ-
ations due to linkage disequilibrium using haplotype tags: a class of tests and the
determinants of statistical power. Human Heredity,56,18–31.
Crocker, J., Olivier, M.-A., Nuer, N. (2009). Self-image goals and compassionate goals:
costs and benefits. Self & Identity,8, 251–69.
Diamond, L.M. (2004). Emerging perspectives on distinctions between romantic love and
sexual desire. Current Directions in Psychological Science,13(3), 116–9.
Ditzen, B., Schaer, M., Gabriel, B., Bodenmann, G., Ehlert, U., Heinrichs, M. (2009).
Intranasal oxytocin increases positive communication and reduces cortisol levels
during couple conflict. Biological Psychiatry,65(9), 728–31.
Feldman, R. (2012). Oxytocin and social affiliation in humans. Hormones and Behavior,61,
Feldman, R., Gordon, I., Influs, M., Gutbir, T., Ebstein, R.P. (2013). Parental oxytocin and
early caregiving jointly shape children’s oxytocin response and social reciprocity.
Feldman, R., Zagoory-Sharon, O., Weisman, O., et al. (2012). Sensitive parenting is asso-
ciated with plasma oxytocin and polymorphisms in the OXTR and CD38 genes.
Biological Psychiatry,72, 175–81.
Ferrero, E., Saccucci, F., Malavasi, F. (1999). The human CD38 gene: polymorphism, CpG
island, and linkage to the CD157 (BST-1) gene. Immunogenetics,49(7–8), 597–604.
Fredrickson, B.L. (1998). What good are positive emotions? Review of General Psychology,2,
Gable, S.L., Gonzaga, G.C., Strachman, A. (2006). Will you be there for me when things go
right? Supportive responses to positive event disclosures. Journal of Personality and
Social Psychology,91, 904–17.
Gable, S.L., Gosnell, C.L., Maisel, N.C., Strachman, A. (2012). Safely testing the fire alarm:
close others’ responses to personal positive events. Journal of Personality and Social
Gable, S.L., Reis, H.T. (2001). Appetitive and aversive social interaction. In: Harvey, J.,
Wenzel, A., editors. Close Romantic Relationships: Maintenance and Enhancement.
Mahwah, NJ: Lawrence Erlbaum, pp. 169–94.
Gonzaga, G.C., Turner, R.A., Keltner, D., Campos, B., Altemus, M. (2006). Romantic love
and sexual desire in close relationships. Emotion,6(2), 163–79.
Gouin, J.-P., Carter, C.S., Pournajafi-Nazarloo, H., et al. (2010). Marital behavior, oxyto-
cin, vasopressin, and wound healing. Psychoneuroendocrinology,35, 1082–90.
Grant, A.M., Gino, F. (2010). A little thanks goes a long way. Explaining why gratitude
expression motivate prosocial behavior. Journal of Personality and Social Psychology,
Grewen, K.M., Girdler, S.S., Amico, J., Light, K.C. (2005). Effects of partner support on
resting oxytocin, cortisol, norepinephrine, and blood pressure before and after warm
partner contact. Psychosomatic Medicine,67(4), 531–8.
Harlow, H.F. (1958). The nature of love. American Psychologist,13, 673–85.
Hazan, C., Diamond, L.M. (2000). The place of attachment in human mating. Review of
General Psychology,4(2), 186–204.
Hendrick, S.S. (1988). A generic measure of relationship satisfaction. Journal of Marriage
and the Family,50,93–8.
Hill, W., Robertson, A. (1968). Linkage disequilibrium in finite populations. Theoretical
and Applied Genetics,38(6), 226–31.
Holt-Lunstad, J., Birmingham, W.A., Light, K.C. (2008). Influence of a “warm touch”
support enhancement intervention among married couples on ambulatory blood pres-
sure, oxytocin, alpha amylase, and cortisol. Psychosomatic Medicine,70(9), 976–85.
Jamroziak, K., Szemraj, Z., Grzybowska-Izydorczyk, O., et al. (2009). CD38 gene poly-
morphisms contribute to genetic susceptibility to B-cell chronic lymphocytic leukemia:
evidence from two case-control studies in polish caucasians. Cancer Epidemiology,
Biomarkers & Prevention: A Publication of the American Association for Cancer
Research, Cosponsored by the American Society of Preventive Oncology,18(3), 945–53.
Jin, D., Liu, H.X., Hirai, H., et al. (2007). CD38 is critical for social behaviour by regulating
oxytocin secretion. Nature,446(7131), 41–5.
Kiss, I., Levy-Gigi, E., Keri, S. (2011). CD 38 expression, attachment style and habituation
of arousal in relation to trust-related oxytocin release. Biological Psychology,88(2–3),
Kogan, A., Saslow, L.R., Impett, E.A., Oveis, C., Keltner, D., Saturn, S.R. (2011). Thin-
slicing study of the oxytocin receptor (OXTR) gene and the evaluation and expression of
the prosocial disposition. Proceedings of the National Academy of Sciences of the United
States of America,108(48), 19189–92.
Kok, B.E., Coffey, K.A., Cohn, M.A., et al. (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), 1123–32.
Konrath, S., Fuhrel-Forbis, A., Lou, A., Brown, S. (2012). Motives for volunteering are
associated with mortality risk in older adults. Health Psychology,31(1), 87–96.
Kosfeld, M., Heinrichs, M., Zak, P.J., Fischbacher, U., Fehr, E. (2005). Oxytocin increases
trust in humans. Nature,435, 673–6.
Kubacka, K.E., Finkenauer, C., Rusbult, C.E., Keijsers, L. (2011). Maintaining close rela-
tionships: gratitude as a motivator and a detector of maintenance behavior. Personality
and Social Psychology Bulletin,37, 1362–75.
Lambert, N.M., Clark, M.S., Durtschi, J., Fincham, F.D., Graham, S.M. (2010). Benefits of
expressing gratitude: expressing gratitude to a partner changes one’s view of the rela-
tionship. Psychological Science,21, 574–80.
Lerer, E., Levi, S., Israel, S., et al. (2010). Low CD38 expression in lymphoblastoid cells and
haplotypes are both associated with autism in a family-based study. Autism Research:
Official Journal of the International Society for Autism Research,3(6), 293–302.
Lewontin, R.C. (1964). The interaction of selection and linkage. I. General considerations;
heterotic models. Genetics,49,49–67.
McCullough, M.E., Kilpatrick, S.D., Emmons, R.A., Larson, D.B. (2001). Is gratitude a
moral affect? Psychological Bulletin,127(2), 249–66.
McNulty, J.K., Fincham, F.D. (2012). Beyond positive psychology? Toward a contextual
view of psychological processes and well-being. The American Psychologist,67(2),
Munesue, T., Yokoyama, S., Nakamura, K., et al. (2010). Two genetic variants of CD38 in
subjects with autism spectrum disorder and controls. Neuroscience Research,67(2),
Polzonetti, V., Carpi, F.M., Micozzi, D., Pucciarelli, S., Vincenzetti, S., Napolioni, V.
(2012). Population variability in CD38 activity: correlation with age and significant
effect of TNF-alpha -308G>A and CD38 184C>G SNPs. Molecular Genetics and
Raudenbush, S.W., Bryk, A.S., Cheong, Y.F., Congdon, R.T. (1996). HLM 5. Hierarchical
Linear and Nonlinear Modeling. Chicago: Scientific Software International.
Reis, H., Clark, M.S., Holmes, J.G. (2004). Perceived partner responsiveness as an organiz-
ing construct in the study of intimacy and closeness. In: Mashek, D.J., Aron, A.P.,
editors. Handbook of Closeness and Intimacy. Mahwah, NJ: Lawrence Erlbaum
Associates Publishers, pp. 201–25.
Riebold, M., Mankuta, D., Lerer, E., et al. (2011). All-trans retinoic acid upregulates
reduced CD38 transcription in lymphoblastoid cell lines from autism spectrum disorder.
Molecular Medicine (Cambridge, Mass.),17(7–8), 799–806.
Saborit-Villarroya, I., Vaisitti, T., Rossi, D., et al. (2011). E2A is a transcriptional regulator
of CD38 expression in chronic lymphocytic leukemia. Leukemia : Official Journal of the
Leukemia Society of America, Leukemia Research Fund, U.K, 25(3), 479–88.
Sauer, C., Montag, C., Worner, C., Kirsch, P., Reuter, M. (2012). Effects of a common
variant in the CD38 gene on social processing in an oxytocin challenge study: possible
links to autism. Neuropsychopharmacology,37(6), 1474–82.
Schneiderman, I., Zagoory-Sharon, O., Leckman, J.F., Feldman, R. (2012). Oxytocin during
the initial stages of romantic attachment: relations to couples’ interactive reciprocity.
Shiota, M.N., Neufeld, S.L., Yeung, W.H., Moser, S.E., Perea, E.F. (2011). Feeling good:
autonomic nervous system responding in five positive emotions. Emotion,11(6),
Smith, T.W., Uchino, B.N., MacKenzie, J., et al. (2103). Effects of couple interactions and
relationship quality on plasma oxytocin and cardiovascular reactivity: empirical findings
and methodological considerations. International Journal of Psychophysiology,88,
Taylor, S.E., Saphire-Bernstein, S., Seeman, T.E. (2010). Are plasma oxytocin in women
and plasma vasopressin in men biomarkers of distressed pair-bond relationships?
Psychological Science,21(1), 3–7.
Trivers, R.L. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology,46,
Uvnas-Moberg, K. (1996). Neuroendocrinology of the mother-child interaction. Trends in
Endocrinology and Metabolism,7(4), 126–31.
Uvnas-Moberg, K. (1998). Oxytocin may mediate the benefits of positive social interaction
and emotions. Psychoneuroendocrinology,23(8), 819–35.
Wigginton, J.E., Cutler, D.J., Abecasis, G.R. (2005). A note on exact tests of Hardy-
Weinberg equilibrium. The American Journal of Human Genetics,76(5), 887–93.
Williams, J.R., Insel, T.R., Harbaugh, C.R., Carter, C.S. (1994). Oxytocin administered
centrally facilitates formation of a partner preference in female prairie voles (Microtus
ochrogaster). Journal of Neruoendocrinology,6, 247–50.
Expressed gratitude and CD38 SCAN (2014) 1861