Zahn 2016. HRV and Self Control

Biological Psychology 115 (2016) 926
Contents lists available at ScienceDirect
Biological Psychology
journal homepage: www.elsevier.com/locate/biopsycho
Review
Heart rate variability and self-controlA meta-analysis

Daniela Zahn a, , Johanna Adams a,b , Jeanette Krohn c , Mario Wenzel a , Caroline G. Mann a ,
Lara K. Gomille a , Vera Jacobi-Scherbening a , Thomas Kubiak a
a
Johannes Gutenberg-University Mainz, Health Psychology, Binger Str. 14-16, D-55122 Mainz, Germany
University Medicine, Johannes Gutenberg-University Mainz, Institute for Teachers Health, University Medical Center of the Johannes
Gutenberg-University Mainz, Kupferbergterrasse 17-19, D-55116 Mainz, Germany
c
Johannes Gutenberg-University Mainz, Personality Psychology and Psychological Assessment, Johannes Gutenberg-University Mainz, Binger Str. 14-16,
D-55122 Mainz, Germany
b
a r t i c l e
i n f o
Article history:
Received 4 January 2015
Received in revised form
11 December 2015
Accepted 21 December 2015
Available online 30 December 2015
Keywords:
Meta-analysis
Self-control
Heart rate variability
Executive functions
a b s t r a c t
Heart rate variability (HRV) has been suggested as a biological correlate of self-control. Whereas many
studies found a relationship between HRV at rest and self-control, effect sizes vary substantially across
studies in magnitude and direction. This meta-analysis evaluated the association between HRV at rest
and self-control in laboratory tasks, with a particular focus on the identication of moderating factors
(task characteristics, methodological aspects of HRV assessment, demographics). Overall, 24 articles with
26 studies and 132 effects (n = 2317, mean age = 22.44, range 18.457.8) were integrated (random effects
model with robust variance estimation). We found a positive average effect of r = 0.15, 95% CI [0.088;
0.221], p < 0.001 with a moderate heterogeneity (I2 = 56.10%), but observed evidence of publication bias.
Meta-regressions did not reveal signicant moderators. Due to the presence of potential publication bias,
our results have to be interpreted cautiously.
2015 Elsevier B.V. All rights reserved.
Contents
1.
2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.1.
HRV and self-control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.
Association between HRV and self-control performance in laboratory tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.
Moderators of the HRV-self-control connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.1.
Task characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.2.
Assessment of HRV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.3.
Study quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.4.
Age- and sex-differences in HRV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.
The present study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.
Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.
Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.
Coding of effect sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1.
2.3.2.
Methods of HRV assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.3.
Study quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.
Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4.1.
Computation of effect sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.1.
Average association between HRV and self-control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Corresponding author. Tel.: +49 6131 3939168; fax: +49 6131 3939186.
E-mail address: zahn@uni-mainz.de (D. Zahn).
http://dx.doi.org/10.1016/j.biopsycho.2015.12.007
0301-0511/ 2015 Elsevier B.V. All rights reserved.
10
D. Zahn et al. / Biological Psychology 115 (2016) 926
3.2.
4.
Moderation analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1.
3.2.2.
Methodological factors of HRV assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.3.
Age and sex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.4.
Study quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.
Publication bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Appendix A.
Coding scheme and classication for task characteristics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1. Introduction
Self-control refers to ones capacity to inhibit or modify dominant impulses related to thoughts, behaviors or emotions (de
Ridder, Lensvelt-Mulders, Finkenauer, Stok, & Baumeister, 2012)
and is considered a key factor for success in goal-oriented behaviors within a wide range of domains including exercising and
adopting an active lifestyle, healthy eating, job and school performance (de Ridder et al., 2012). Self-control is also associated
with psychological well-being and positive interpersonal relationships (Tangney, Baumeister, & Boone, 2004). The Strength Model of
Self-Control (Muraven & Baumeister, 2000) posits that self-control
is a limited resource. In turn, self-control research has focused
on identifying underlying correlates of self-control, which may
reect self-control capacity on a physiological level. In this line of
research, heart rate variability (HRV) has been suggested as a correlate of self-control (Baumeister, Vohs, & Tice, 2007). HRV refers
to the beat-to-beat variation in heart rate and reects the interplay
between sympathetic and parasympathetic inuences on heart rate
(Appelhans & Luecken, 2006). It is assumed that HRV is an indicator of the exibility of the autonomous nervous system which is
necessary to modulate cardiac activity according to changing situational demands arising from changes in physiological as well as
psychological states (Appelhans & Luecken, 2006). Based on this
notion, HRV has been discussed as an index of physiological and
psychological self-regulation (Thayer, Hansen, & Johnsen, 2010).
Self-regulation and self-control are closely linked: whereas some
researchers refer to the term self-regulation to describe automatic
regulatory processes and to self-control to describe the deliberate
regulation of impulses and behaviors, they are often used interchangeably (Baumeister et al., 2007).
Neurovisceral Integration Model (Thayer, Hansen, Saus-Rose, &
Johnsen, 2009; Thayer & Lane, 2000) as well as Porges Polyvagal Theory (Porges, 2001, 2007) provide theoretical frameworks
linking HRV to concomitants of self-control such as cognitive and
emotional regulation and social engagement. Neurovisceral Integration Model (Thayer et al., 2009; Thayer & Lane, 2000) posits
that trait (i.e., at rest) HRV is a proxy for the inhibitory capacity
of a central autonomic network (CAN) that regulates behavioral,
cognitive, and emotional responses (Thayer et al., 2010). The CAN
comprises brain regions related to inhibition and executive functions such as the prefrontal cortex and is reciprocally connected
with the heart as well as the periphery via parasympathetic and
sympathetic neural pathways. Through this neural network, the
prefrontal cortex can exert inhibitory control on subcortical structures, thereby enabling the individual to exibly and adaptively
respond to situational demands. Polyvagal Theory (Porges, 2001,
2007) argues that the vagal tone is part of a social engagement
system that had been originally developed to regulate ight/ght
reactions triggered by sympathetic nervous system. Due to functional connections between vagal outow and structures related
to emotion processing, attention, and communication, the social
engagement system enables behavioral responses such as social
communication, self-soothing, and the inhibition of physiological
arousal (Porges, 2007). Respiratory sinus arrhythmia (RSA), which

is closely related to trait HRV, serves an index of tonic vagal tone.
According to Porges, RSA indicates an individuals ability to maintain homeostasis and responsiveness to changing demands (Porges,
2001, 2007). In summary, both models assume that higher HRV
should reect a better ability to inhibit dominant impulses and
behaviors, which can be considered a key component of automatic
regulatory processes (i.e., self-regulation) as well as deliberate regulation (i.e., self-control). Hence, higher HRV may also indicate a
better capability for self-control. Whereas many studies hint at a
relationship between HRV and concomitants of self-control, the
magnitude of the relationship remains still unclear.
In the present study, we aimed at (1) quantifying the magnitude
of the association between trait HRV and self-control in laboratory
settings in a meta-analysis and (2) evaluating potential moderators
of the HRV-self-control connection. In the following sections, we
briey review empirical evidence on the association between selfcontrol and trait HRV and discuss moderators of this relationship.
1.1. HRV and self-control
In the last decades, a considerable amount of empirical evidence
has accumulated linking trait HRV and self-control: HRV is associated with a wide range of self-control processes and outcomes, such
as addictive behaviors (e.g., Ingjaldsson, Laberg, & Thayer, 2003),
binge eating (e.g., Friederich et al., 2006), depression (Rottenberg,
2007), emotion regulation strategies (see Appelhans & Luecken,
2006; for an overview), and cognitive performance (Thayer et al.,
2009). The indicators of self-control that have been used in this
research may be broadly categorized as outcome-related measures (e.g., addictive behavior, diet), self-report measures of trait
self-control, and cognitive and affective processes reecting selfcontrol. However, there is evidence that these measures show only
modest overlap: for example, one meta-analysis found only small
to moderate associations between self-reported trait self-control
and self-control outcomes such as diet or addictive behavior (de
Ridder et al., 2012). In addition, self-report measures are only
weakly correlated with performance in self-control tasks assessing
basic cognitive processes (Duckworth & Kern, 2011). Outcomes and
self-report measures reect rather distal indicators of self-control,
whereas cognitive processes can be considered proximal indicators
of self-control, which constitute the basic mechanisms underlying
self-control success in a wide range of behavioral domains. The current meta-analysis will focus on the relationship between HRV and
basic cognitive processes as assessed by self-control tasks in the
laboratory.
1.2. Association between HRV and self-control performance in
laboratory tasks
Numerous experimental studies found evidence for positive
associations between trait HRV and self-control for a range of cognitive processes including goal setting (Geisler & Kubiak, 2009),
working memory and executive function (Hansen, Johnsen, &
Thayer, 2003; Hansen, Johnsen, & Thayer, 2009), emotion regulation after experimental induction of affect (Pu, Schmeichel, &
Demaree, 2010; Volokhov & Demaree, 2010), information processing of emotional stimuli (Park, Van Bavel, Vasey, & Thayer, 2012),
response inhibition (Krypotos, Jahfari, van Ast, Kindt, & Forstmann,
2011; Park et al., 2012), or persistence (Reynard, Gevirtz, Berlow,
Brown, & Boutelle, 2011; Segerstrom & Solberg Nes, 2007). Other
studies, however, examining similar cognitive processes revealed
conicting ndings: For example, in the study of Frazier et al.,
trait HRV was not related to emotion regulation after experimental
affect induction (Frazier, Strauss, & Steinhauer, 2004). With respect
to executive functions and selective attention, two studies did
not nd any relationship between HRV and self-control (Duschek,
Muckenthaler, Werner, & del Paso, 2009; Ohira et al., 2013), and
another study revealed negative associations (Cellini, Covassin, de
Zambotti, Sarlo, & Stegagno, 2013). Other studies indicate that the
association between HRV and self-control differs depending on different performance measures of the same self-control task (e.g.,
Krypotos et al., 2011; Luft, Takas, & Darby, 2009). Taken together,
ndings on the link between HRV and self-control are mixed and
the magnitude of the effect is difcult to appreciate, because a
wide range of different cognitive and affective processes reecting higher as well as lower order self-control processes have been
examined. In consequence, we assume that task characteristics and
also several other variables may moderate the association between
HRV and self-control.
1.3. Moderators of the HRV-self-control connection
As pointed out above, task characteristics might moderate the
relationship between HRV and self-control. In addition, there are
several methods for heart rate recording and HRV computation
that could impact study results. As HRV varies with age and sex,
these characteristics may also moderate the HRV-self-control connection. In the following paragraphs, we outline the role of task
characteristics, methodological aspects in HRV assessment and
computation as well as age- and sex-differences and discuss the
potential inuence of these aspects on the HRV-self-control association.
1.3.1. Task characteristics
Given the variability of different tasks employed in self-control
research and the different basic processes, which are assessed by
these tasks, we assume that different sets of classications might
be useful to categorize self-control tasks: a classication relying on
content domains of self-control tasks and a classication pertaining
to underlying processes necessary to solve self-control tasks.
With respect to content domains of self-control tasks, a
recent meta-analysis on self-control failure in dual-task paradigms
(Hagger, Wood, Stiff, & Chatzisarantis, 2010) differentiated selfcontrol tasks by self-control domains (self-control spheres), as
suggested by Baumeister et al. (Baumeister et al., 2007). The categories comprised controlling attention, controlling emotions,
controlling thoughts, controlling impulses, choice and volition, cognitive processing, and social processing (Hagger et al.,
2010). Subgroup analyses showed that the magnitude of the effect
size varied slightly across the different domains (Hagger et al.,
2010), indicating that self-control domain may also constitute a
potential moderating variable for the HRV-self-control connection.
As for a classication based on underlying basic cognitive processes, Hofmann, Schmeichel, and Baddeley (2012) argued for
adopting the taxonomy of executive functions derived by Miyake
et al. (Miyake et al., 2000) for self-control tasks. Miyake and colleagues distinguish three sets of executive functions: shifting,
updating, and inhibiting. As inhibition of dominant responses or
behaviors constitutes a core element of most denitions of self-
11
control (de Ridder et al., 2012), we hypothesize that the relationship

between HRV and self-control might perhaps be stronger in tasks
requiring inhibition compared to updating or shifting.
To conclude, task characteristics such as self-control domains
or underlying sets of executive functions be considered potential
moderators of the HRV-self-control connection.
1.3.2. Assessment of HRV
1.3.2.1. Recording device. HRV computation requires the recording
of beat-to-beat information to determine RR intervals (Task Force of
the European Society of Cardiology and the North American Society
of Pacing and Electrophysiology, 1996). Although a full-featured
electrocardiography (ECG) is considered the gold standard (Task
Force of the European Society of Cardiology and the North American
Society of Pacing and Electrophysiology, 1996), other devices such
as portable easy-to-use heart rate monitors (HRM) or pulse wave
based devices (PWD) have been amply used as well (e.g., Schafer
& Vagedes, 2013; Weippert et al., 2010). Studies comparing HRM
and ECG showed equivalent performance of HRMs with regard to
RR interval detection (Nunan et al., 2008, 2009; Weippert et al.,
2010). However, there are also studies indicating that HRMs provide higher HRV values than ECG (Nunan et al., 2008; Wallen,
Hasson, Theorell, Canlon, & Osika, 2012). Although this overestimation may be due to the use of HRM software for (automated)
artifact control commonly supplied by the HRM manufacturers
rather than to a lack of accuracy in RR interval detection (Quintana,
Heathers, & Kemp, 2012), some authors still doubt the use of HRMs
for HRV assessment in general (Wallen et al., 2012). PWDs rely on
the detection of pulse cycles to monitor heart rate and can also be
used to determine HRV. Compared to ECGs, HRV values derived by
PWDs may also tend to overestimate HRV and are more susceptible to motion artifacts (Schafer & Vagedes, 2013). Taken together,
different recording devices may inuence HRV measures and in
consequence, may also affect the HRV-self-control connection.
1.3.2.2. Computational methods to derive HRV measures. Two computational approaches are common to determine HRV: time domain
methods use the duration (mostly measured in milliseconds) of
successive interbeat intervals to derive HRV estimates such as the
standard deviation of RR intervals (SDNN) or the root mean square
of successive differences of RR intervals (RMSSD). SDNN reects
overall HRV and RMSSD reects mostly parasympathetic inuence
(Task Force of the European Society of Cardiology and the North
American Society of Pacing and Electrophysiology, 1996; Thayer
et al., 2010). Frequency domain methods separate the variance of
interbeat intervals into a spectrum of given frequencies (Task Force
of the European Society of Cardiology and the North American
Society of Pacing and Electrophysiology, 1996). The high-frequency
component (HF; 0.150.40 Hz) refers to rapid changes in heart rate,
which are assumed to reect parasympathetic inuence, whereas
the low-frequency component (LF; 0.040.15 Hz) refers to slower
changes in HR. The meaning and origin of the LF component is controversially discussed (Goldstein, Bentho, Park, & Sharabi, 2011):
the LF component has been suggested to primarily indicate sympathetic outow to the heart, in particular if adjusted for HF power,
respiration, or total power. Recent evidence rather indicates that
LF power reects baroreex function (Goldstein et al., 2011).
For frequency domain versus time domain measures, a systematic review on HRV values in samples of healthy adults showed
a larger range in HRV for frequency domain measures compared
to time domain measures (Nunan, Sandercock, & Brodie, 2010).
The authors assume that this difference may be explained by the
higher susceptibility of frequency domain measures for technical
errors such as artifacts, poor preprocessing of RR data, and nonstationarity.
12
In addition to time and frequency domain measures of trait

HRV, research within Porges Polyvagal Theory has often used the
RSA as a measure of HRV. RSA is a naturally occurring rhythm in
the heart rate pattern at approximately the frequency of spontaneous breathing (Porges, 2007; p. 122). Different methods have
been used to estimate RSA (see Lewis, Furman, McCool, & Porges,
2012, for an overview): as the frequency of spontaneous breathing
is similar to the HF component of HRV, RSA has been estimated by
the HF component. But other methods such as the PorgesBohrermethod (Porges & Bohrer, 1990) or the Peak-to-trough algorithm
(e.g., Grossman, van Beek, & Wientjes, 1990; Katona & Jih, 1975)
have also been employed to quantify RSA. The PorgesBohrermethod rst removes variance associated with activity slower than
spontaneous breathing by polynomial lters. RSA is estimated by
averaging logarithmically transformed RSA estimates for several
short time segments (e.g., 30 s intervals) of the HR assessment.
This method is assumed to minimize inuences of non-stationarity
and ensures normal distribution of RSA estimates (Lewis et al.,
2012). The peak-to-trough-method averages the difference of the
longest heart rate period during expiration and the shortest heart
rate period in inspiration for each breath cycle. This RSA estimate reects the mean range (in milliseconds) of the heart rate
period oscillations associated with synchronous respiration (Lewis
et al., 2012). Although the different RSA metrics are highly correlated, there is evidence that they may differ in susceptibility for
non-stationarity, inuence of respiration and sensitivity to vagal
mechanisms (Lewis et al., 2012), indicating that the three metrics may not be equivalent. Lewis and colleagues conclude that
RSA determined by the PorgesBohrer-method is less inuenced by
respiration and non-stationarity than the HF component or Peakto-trough metric. To summarize, different HRV metrics have been
suggested, which may moderate the HRV-self-control connection.
Although the Task Force guidelines (1996) recommends only time
domain and frequency domain measures, Porges strongly argues
for the use of RSA estimated by the PorgesBohrer method. However, research beyond the Polyvagal Theory more often employs
HRV metrics recommended by the Task Force guidelines. Thus, we
limited the current meta-analysis on HRV metrics as recommended
by the Task Force.
1.3.2.3. Duration of HRV assessment. Short-term recordings for HRV
assessment at rest should last two minutes minimum with an optimal length of ve minutes (Task Force of the European Society
of Cardiology and the North American Society of Pacing and
Electrophysiology, 1996). Time domain measures, in particular
SDNN, are strongly inuenced by length of recording, resulting in
higher SDNN in longer recordings (Nunan et al., 2010). Thus, the
duration of HRV assessment at rest may also inuence the HRVself-control connection.
1.3.2.4. Control of respiration during HRV assessment. There is much
controversy over the inuence of respiration on HRV (Thayer et al.,
2010; Thayer, Loerbroks, & Sternberg, 2011). Whereas there is
agreement that respiration and cardiac change are related, opinions
differ on the direction of this relationship. Researchers assuming
that respiration causally inuences HRV posit that individual differences in respiration should be controlled. They either suggest
to monitor respiration and statistically adjust HRV or to employ
paced breathing during HRV assessment to minimize individual differences in respiration rate (Grossman, 1992). Others assume that
cardiac changes may causally inuence respiration as evidenced for
example by Tzeng, Larsen, and Galletly (2003), and argue that controlling or minimizing the inuence of respiration may remove an
important part of the variability in HRV (Thayer et al., 2010; Thayer
et al., 2011). Furthermore, HRV metrics may differ in susceptibility to respiratory changes. For example, the RMSSD (Penttila et al.,
2001) or the RSA estimated with the PorgesBohrer-method seem

to be less affected by changes in respiration than the HF component
(Lewis et al., 2012). To conclude, there is no consensus about dealing with respiration during HRV assessment. In consequence, it is
difcult to determine how control of respiration may exactly affect
the HRV-self-control connection. Assuming a causal inuence of
respiration on HRV, one would expect that controlling for respiration might increase the relationship between HRV and self-control.
Supposing that cardiac change causally inuences respiration, control for respiration rate might rather reduce the HRV-self-control
connection.
In summary, these ndings indicate that several aspects related
to HRV recording and computation such as recording device
(ECG versus HRM versus PWD), computational method (frequency
versus time domain method), length of recording, and control of
respiration rate (i.e., statistical adjustment versus paced breathing
versus no control for respiration) may inuence HRV.
1.3.3. Study quality
A recent meta-analysis on functional somatic disorders and HRV
(Tak et al., 2009) pointed out that the overall quality of studies
shows a great variation in several methodological aspects such as
appropriate selection of participants, quantication and reporting
of HRV, and appropriate adjustment of confounders. Some criteria
included in the quality index by Tak et al. (2009) may be rather
specic for studies comparing HRV in patients with somatic disorders versus healthy controls (e.g., assessment and stating disease
characteristics or HRV assessment blinded for disease status of participants). But many of the quality criteria suggested by Tak et al.,
particularly those referring to reporting and quantication of HRV,
may also play a role in studies on the relationship between HRV
and self-control. Furthermore, other meta-analyses provide evidence that study quality inuences the magnitude of the effect
size, indicating that poorer study quality is related to larger effect
sizes (e.g., Cuijpers, van Straten, Bohlmeijer, Hollon, & Andersson,
2010; Rosenbaum, Tiedemann, Sherrington, Curtis, & Ward, 2014).
Thus, we believe that study quality may also moderate the HRVself-control connection.
1.3.4. Age- and sex-differences in HRV
HRV reaches its maximum in young adulthood (1524 years
of age) and tends to decline with age until the 6th decade (see
Sosnowski, 2010; for an overview), which is caused mostly by a
specic age-related change in vagal modulation of the heart rate
(De Meersman & Stein, 2007; Fukusaki, Kawakubo, & Yamamoto,
2000). Results on sex differences in HRV are less clear: Some studies
found higher HRV in women compared to men over all age groups
(Antelmi et al., 2004; Britton et al., 2007), others conclude that
higher HRV in women is more likely observable in younger age
groups (up to 30 years of age) than in older age (De Meersman &
Stein, 2007; Stein, Kleiger, & Rottman, 1997). In contrast to these
results, a recent systematic review found lower HRV in women
compared to men, in particular for time domain measures (Nunan
et al., 2010). The consequences of these age- and sex differences on
the HRV-self-control connection are difcult to determine, because
moderating effects resulting in differential associations for younger
and older participants or women and men as well as mediating
inuences of age and sex are possible.
1.4. The present study
The current study aimed at the quantitative integration of
research results on trait HRV and self-control performance. We
limited ourselves to studies assessing self-control performance
in laboratory tasks, because we assume that cognitive processes
measured with these tasks reect the basic mechanisms underly-
ing self-control success. The focus of the quantitative integration

was twofold: (1) to compute the average effect size between
HRV and self-control by meta-analysis and (2) to evaluate and
identify potential variables, which may moderate the relationship
between HRV and self-control, by meta-regressions and subgroup
analyses. Possible moderating variables included characteristics of
self-control tasks (self-control domains, sets of executive functions), aspects of HRV assessment (recording device, time versus
frequency domain measures, length of heart rate assessment, control of respiration), study quality, age, and sex.
2. Method
2.1. Literature review
We conducted a comprehensive literature search in Pubmed,
Web of Science, and Medpilot (JA, DZ). The search was restricted
to research in humans and to articles in peer-reviewed scientic
English language journals published until December 2013. We combined search terms for heart rate variability with search terms for
self-control as well as with specic tests such as working memory tests (WMT) and continuous performance tests (CPT), because
Thayers Neurovisceral Model is also based on studies using these
specic tests. The full electronic search strategy in Pubmed syntax was as follows: (((heart rate variability) OR (respiratory sinus
arrhythmia) OR (heart period variability) OR (rr variability) OR
(cycle length variability)) AND ((self-control) OR (regulation) OR
(emotion) OR (CPT) OR (cognition) OR (attention) OR (WMT) OR
(self-regulation))) AND English [language] lters: humans; adult:
19+ years; middle aged: 4564 years. In addition, a manual search
was performed by scanning reference lists of reviewed articles. The
literature search yielded 5551 records total (see Fig. 1).
2.2. Selection
Eligibility criteria for the meta-analysis were:
(1) The study had to be conducted in a controlled laboratory setting.
(2) Self-control had to be conceptualized in line with notion of de
Ridder et al. (2012), i.e., self-control as the capacity to inhibit
or modify dominant impulses related to thoughts, behaviors,
and/or emotions. This is a very broad denition that applies to
a wide range of different laboratory tasks.
(3) At rest heart rate recording had to be of a two minutes minimum
according to the requirement laid down by the Task Force of the
European Society of Cardiology and the North American Society
of Pacing and Electrophysiology (1996).
(4) HRV had to be computed by frequency domain methods using
non-parametric or parametric power spectral density analysis
or statistical time domain methods according to the guidelines
of the Task Force of the European Society of Cardiology and the
North American Society of Pacing and Electrophysiology (1996).
(5) The studies had to report HRV measures, which reected
mostly parasympathetic inuence, i.e., the HF component or
RMSSD. Studies reporting SDNN were only included, if they also
reported HF or RMSSD.
(6) Studies conducted in clinical populations (somatic and/or mental disorders) were excluded.
(7) Studies examining children and adolescents (i.e., those under
age of 18) were excluded. To avoid a confoundation of HRV with
undetected cardiovascular diseases, we excluded studies with
participants over an age of 65 years, because incidence of heart
diseases is increased in those aged 65 years and above (Go et al.,
2014).
13
All records were screened by two independent raters (JA, DZ)

that led to the identication of a total of 24 publications that met
the eligibility (Fig. 1).
2.3. Coding of effect sizes
All effects pertaining to a relationship between HRV and selfcontrol were extracted. Extracted data for moderation analyses
included demographic data (mean age of study samples, percentage of male participants), task characteristics, and methodological
characteristics of HRV assessment and computation, and study
quality. The coding of study characteristics was performed by three
pairs of independent raters (JA/CM; DZ/MW; DZ/LG). Inter-rater
reliability was determined by intra class correlation coefcients
(ICC) for continuous variables and Cohens Kappa () for categorical variables. Possible discrepancies between raters were solved by
consensus. Initial inter-rater reliability for demographic data was
excellent (ICC = 1 for percentage of male participants, sample size,
and mean age of participants). Task characteristics and methodological characteristics of HRV assessment are described in more
detail below.
We employed two distinct sets of categories pertaining to task
characteristics to cover (1) the content domain of the task (selfcontrol domains) and (2) underlying cognitive processes (sets of
executive functions). Each effect was classied according to the two
sets of categories. Appendix A provides an overview about the
two sets of task characteristics and the respective classication
of task measures. Initial inter-rater reliability for the two sets of
task characteristics was excellent and ranged from = 0.89 (executive functions) to = 0.94 (self-control domains). The remaining
discrepancies between raters were resolved by consensus.
2.3.1.1. Self-control domains. For self-control domains, we used a
set of categories adapted from the self-control taxonomies by
Hagger et al. (2010) and Baumeister et al. (2007). These self-control
domains divide self-control tasks according to broad content features of the tasks. Because only one identied study (Laborde &
Raab, 2013) reported on tasks pertaining to choice and volition,
and none of the studies employed tasks pertaining to social processing, we limited the self-control domains to three categories (a)
emotion regulation (b) attention control (c) cognitive control. Hagger
and colleagues also distinguished between subdomains controlling thoughts, controlling impulses, and cognitive processing.
In the current study, we combined these three subdomains to the
category cognitive control, because the subdomains could not be
clearly separated in the tasks employed in the included studies.
Emotion regulation covered task measures that involved an
active manipulation of experienced emotions, i.e., the suppression of emotions or exaggeration of experienced emotions after
experimental mood induction or cognitive processing in the presence of emotional stimuli, i.e., changes in task performance after
experimental mood induction or information processing of emotional stimuli. Studies assessing pre-post differences in affect
after mood induction without specic instruction to manipulate
emotion experience or emotion expression were excluded. Tasks
involving discrimination of emotional expressions were excluded
for the same reason.
The category Attention control included tasks measuring attentional processes such as selective attention, attention control, and
concentration. To avoid a confoundation with the category emotion
regulation, only attention control tasks employing neutral stimuli
were included, task measures pertaining to emotional stimuli were
categorized as emotion regulation.
Included
Eligibility
Screening
Identification
14
Records identified through database searching

(n = 8531)
Duplicates removed
(n = 2985)
Records screened based on title

(n = 5551)
Records excluded
(n = 4167)
Records screened based on abstract

(n = 1384)
Records excluded
(n = 1057)
Full-text articles assessed for eligibility

(n = 327)
Full-text articles excluded (n = 303)

due to
Clinical population (n = 29)
Age (n = 17)
No experimental self-control
assessment (n = 146)
HRV-assessment (n = 102)
No effect size reported (n = 9)
Publications included in quantitative synthesis

(meta-analysis)
(n = 24 with 26 studies)
Fig. 1. Flow chart of study selection.
The category Cognitive control included task measures where

dominant impulses or cognitions had to be suppressed such
as resisting temptations, suppressing thoughts, overriding welllearned responses, or switching from a given mental set to a novel
mental set, as successful switching requires the suppression of the
previous mental set. Only tasks involving emotionally neutral stimuli were categorized as cognitive control. Task measures involving
cognitive control of emotional stimuli were classied as emotion
regulation.
2.3.1.2. Sets of executive functions. Executive functions were classied into three categories (a) Shifting, (b) Updating and (c) Inhibiting
according to the taxonomy by Miyake et al. (2000).
In the category Shifting, tasks were included which require to
switch forth and back between two different mental sets or from
one task to another task, thereby involving the ability to perform a
new operation in the face of proactive interference (Miyake et al.,
2000). Tasks assessing reappraisal after mood induction were also

classied as Shifting, as reappraisal reects a shift from the induced
negative mood towards a more positive way of thinking.
In the category Updating, we included task measures which
require the updating and monitoring of working memory representations. Updating includes monitoring and coding information
relevant for the current task as well as the active manipulation
of information currently held in working memory, for example
replacing older, no longer relevant information by new information
(Miyake et al., 2000). The ability to successfully exaggerate facial
expression after mood induction was also classied as Updating, as
the continuous display of an exaggerated facial expression requires
continuous monitoring of the facial expression.
The category Inhibiting referred to tasks which require the ability to deliberately suppress dominant or automated responses or
behaviors. We also classied tasks measuring performance after
induction of negative mood or in the presence of emotional stimuli
15
Table 1
Overview about study quality criteria and scoring system.
(1) Assessment of basic confounding variables (i.e., age, sex)
Descriptive statistics for age and sex provided or exclusion criteria based on age and sex (1)
Basic confounding variables not clearly stated (0)
(2) Assessment of further confounders
Descriptive statistics for further confounders (i.e., body mass index, psychiatric or somatic diseases, medication use, and smoking status) provided or any of
these variables mentioned as exclusion criteria (1)
Information of further confounding variables not clearly stated (0)
(3) Reporting of artifact correction
Information about identifying and handling artifacts clearly stated (1)
Information about handling artifacts not clearly stated (0)
(4) Pre-rest period before HRV measurement
Reporting of a pre-rest period before HRV measurement (1)
No pre-rest period before HRV measurement or not clearly stated (0)
(5) Posture during HRV measurement
Reporting information about posture during HRV measurement (1)
Posture during HRV measurement not clearly stated (0)
(6) Reporting of the frequency bands dening HF componenta
Information about the frequency bands dening the HF-component stated (1)
Information not clearly stated (0)
(7) Movement during HRV measurement
Measurement of movement during HRV measurement (1)
Measurement of movement during HRV measurement not clearly stated (0)
(8) Adequate reporting of HRV metrics.ab
HF component dened as the frequency range of 0.150.40 Hz and reported in adequate units (i.e., ms2 or normalized units) as recommended by Task Force
(1996) (1)
Units or frequency bands not clearly stated or use of other units/frequency bands (0)
(9) Assessment of HRV and self-control performance at the same session
HRV and self-control task assessed at the same session (1)
HRV and self-control task assessed at different time points (time lag of at least one day) or no information provided (0)
(10) Reporting of adjusted effectsc
Effects reported adjusted for confounders mentioned in criterion 1 or 2 (1)
Effects reported without adjustment for confounders (0)
(11) Self-control performance measure
Objective measure of self-control performance such as errors or reaction time (1)
Subjective measure of self-control performance such as self-report or observer-ratings (0)
(12) Measurement of HRV
Measurement by ECG device (1)
Measurement with other devices than ECG (0)
(13) Duration of HRV assessment
Duration of HRV assessment of at least two minutes (1)
Duration of HRV assessment less than two minutes or not clearly stated (0)
Note: a Criterion only applicable to effect sizes based on the HF component. b If log-transformed values based on ms2 or normalized units were reported, the criterion was
also considered as fullled. c Only applicable, if criteria 1 or 2 were fullled.
as inhibiting, because successful performance requires the suppression of ones affective state or the inhibition of emotional
information, respectively. Self-reported rumination after mood
induction was also categorized as inhibiting, as rumination indicates failure to inhibit thoughts and worries.
2.3.2. Methods of HRV assessment
We extracted information about (1) duration of heart rate assessment (in minutes), (2) HRV measures (RMSSD, SDNN, HF), (3)
recording devices, and (4) control of respiration during HRV assessment. Initial inter-rater reliability for methodological aspects of
HRV measurement ranged from = 0.51 (handling of respiration),
ICC = 0.63 (duration of HRV assessment), = 0.86 (recording device)
to = 0.98 (HRV measure). The remaining discrepancies between
raters were resolved by consensus. Discrepancies in duration of
HRV assessment were mostly caused by studies reporting different
durations for recording of HRV and computation of HRV measures.
Thus, we extracted only the duration of the sequence used for computation of HRV measures for all studies.
2.3.2.1 Recording devices. Recording devices were categorized into
(a) stationary ECG monitors (e.g., Biopac ), (b) ambulatory ECG monitoring systems (e.g., VU-AMS , Life shirt ), (c) HRMs (e.g., Polar
watch ), (d) pulse wave based devices (e.g., Finapres ).
2.3.2.2. Control of respiration. Control of respiration was rated as

follows: (a) no assessment of respiration, (b) paced breathing during HRV assessment, and (c) recording respiration rate and reporting
respiration-adjusted HRV measures. As reported above, inter-rater
reliability for handling of respiration was very low, which might
be due to the fact that the information about handling of respiration was difcult to extract because most studies did not explicitly
report on the handling of respiration. During discussion to reach
consensus about the coding, most studies were categorized as no
assessment of respiration. In consequence, out of the 132 effects,
125 effects were based on HRV measurement without any assessment or control of respiration. Due to the lack of variance and low
inter-rater reliability, we refrained from formal moderation testing
for respiration.

We adapted the quality criteria suggested by Tak et al. (2009)
and rated 13 criteria of study quality (see Table 1). For criteria 8,
12, and 13 we referred to the Task Force guidelines (1996), which
dene the frequency range for the HF component as 0.150.40 Hz
and advise to report the HF-component in ms2 or normalized
units. The Task Force guidelines also consider the ECG as the gold
standard for HRV assessment and recommend a minimum of two
minutes recording time for short-term assessment of HRV. Criterion 9 (Assessment of HRV and self-control performance at the same
16
session) was chosen as we assumed that a time lag between the

assessments of the two variables may lead to an underestimation of the HRV-self-control association. Criterion 11 (Self-control
performance measure) was added because we assumed that assessing self-control performance by measuring errors or reaction time
might yield a more objective performance measure than performance based on self-report or observer-ratings. Quality criteria
were coded separately for each extracted effect, with coding 1
for each fullled criterion and 0 in case the criterion was not
met or not explicitly stated. As the number of applicable criteria
varied for each effect depending on the HRV metrics (HF versus
RMSSD/SDNN), we computed the mean of the applicable criteria
and multiplied the resulting score with 100. The quality index could
vary between zero and 100 and reects the percentage of fullled
criteria, i.e., a score of 100 indicates that 100% of the quality criteria were met. Inter-rater reliability was excellent for the overall
score of study quality (ICC = 0.95). Cohens for single components
of study quality ranged from = 0.60 (assessment of HRV and selfcontrol performance at the same session) to = 1 (e.g., adjustment
for confounders, artifact correction).
2.4. Data analysis
2.4.1. Computation of effect sizes
We used Pearsons correlation coefcient r as the common effect
size (ES) measure. When studies reported results from t-tests, analyses of variance, or regression analyses, r was computed from
reported data using recommended transformations (Borenstein,
Hedges, Higgins, & Rothstein, 2009; Lipsey & Wilson, 2001). Positive correlation coefcients indicate a positive HRV-self-control
association, that is higher HRV is associated with better self-control
performance.
In case the information provided in the publications was not sufcient to compute effect sizes, authors were contacted and asked
to provide the (raw) data needed. When studies reported results as
nonsignicant and information to compute exact estimates of effect
sizes was not available or not provided by the authors upon request,
these effects (n = 21) were included with a conservative effect size
of r = 0 (Rosenthal, 1995). The response rate of authors was 46%. All
effect sizes were transformed in Fishers z-correlations for aggregation and were back-transformed in r afterwards (Pigott, 2012).
All effects pertaining to a relationship between HRV and selfcontrol were extracted, resulting in multiple effects per study.
Most studies reported multiple outcomes based on the same sample, which results in statistically dependent effect size estimates.
As integrating dependent effect size estimates yields biased estimates of effect sizes and standard errors (Hedges, Tipton, & Johnson,
2010), we employed robust variance estimation (RVE) to account
for dependency in effect sizes as suggested by Hedges et al. (2010)
and Tipton (2014). As simulations show, the RVE methods yields
valid estimates of standard errors, point estimates, condence
intervals, and signicance tests in meta-analysis and metaregression of dependent effects (Hedges et al., 2010). The advantage
of RVE methods compared to other methods to deal with statistically dependent effects such as multivariate models is that knowledge about the exact covariance structure between dependent
effect size is not necessary (Hedges et al., 2010). Instead, the RVE
method choses a certain value to reect the covariance of dependent effects (). To ensure that the chosen value is an adequate estimate of the covariance, sensitivity analyses are performed repeating the estimation of effect size, standard errors and 2 (i.e., the
estimate of the between-study covariance structure) with different
values of (Hedges et al., 2010; Tanner-Smith & Tipton, 2013).
To compute the average effect size we used the RVE method
in combination with the random effects model to account for
the heterogeneity in effect sizes among studies (Borenstein et al.,
2009; Pigott, 2012). As multiple effects of a study often resulted

from different outcome measures of the one task, e.g., reaction
time and percentage of errors, we assumed a high correlation
between the effect sizes and chose a = 0.80. Sensitivity analyses
with ranging from = 0.00 to = 1 did not reveal changes in any
estimates. We also applied small sample correction to all analyses,
which corrects the degrees of freedoms and covariance structure,
as recommended for meta-analysis with fewer than 40 studies
(Tipton, 2014). Heterogeneity was assessed using the I2 statistic,
which measures the percentage of total variation in effect sizes
across studies due to heterogeneity.
We evaluated the inuence of moderators using metaregressions with RVE and small sample corrections in combination
with the random effects model. Categorical covariates were
weighted effects-coded in case of variables with more than two
levels (recording device, HRV measures, self-control domains, sets
of executive functions) to compare the levels against the weighted
grand mean. In addition, we computed average effect sizes for each
category of moderating variables employing the RVE method with
small sample correction in combination with the random effects
models.
To detect outliers and to judge publication bias, a second data
set consisting of one weighted mean effect size per study was
created, because outlier diagnostics or methods for publication
bias are not available for RVE models. Outlier diagnostics followed the recommendation of Viechtbauer and Cheung (2010) and
included inspection of externally standardized residuals, DFFITS
value, Cooks distance, covariance ratio, the leave-one-out amount
of (residual) heterogeneity, and the leave-one-out test statistic for
the test of (residual) heterogeneity.
We judged potential publication bias by visual inspection of the
funnel plot, by Eggerss regression test (Sterne & Egger, 2005), and
the trim-and-ll method (Duval & Tweedie, 2000).
Meta-analysis was performed using R software version 3.1.1
(R Core Team, 2014) packages robumeta (Fisher & Tipton, 2014),
metafor (Viechtbauer, 2010), and rmeta (Lumley, 2012). All aggregated effect sizes are presented as correlation coefcients r;
coefcients and standard errors from meta-regressions are presented in the Fishers z metric.
3. Results
Out of 327 screened full text articles, 24 published articles with
26 studies met all criteria for the meta-analysis with an overall
n = 2317 participants and k = 132 effect sizes. Effect sizes per study
ranged from one to 16 effects with a mean number of 5.1 4.34
effects per study. Table 2 provides an overview of the basic characteristics of the included studies.
3.1. Average association between HRV and self-control
The pooled effect size of the HRV-self-control association
amounted to r = 0.15, 95% CI [0.088; 0.221], p < 0.001 with moderate
heterogeneity (I2 = 56.10%, Q = 57.26, df = 25.13, p < 0.001). The forest plot is displayed in Fig. 2. Repeating the analysis without the 21
effect sizes included as zero effects slightly increased the pooled
effect size (r = 0.17, 95% CI [0.10; 0.240], p < 0.001) but slightly
increased heterogeneity to I2 = 60.31%.
3.2. Moderation analyses
3.2.1.1. Self-control domains. Of the 132 extracted effects, 59 effects
were classied as pertaining to emotion regulation, 68 effects as
attention control, and ve effects as cognitive control. Weighted
effect coding was used to compare differences in effect sizes by
Table 2
Characteristics of the included primary studies.
N
Mean
age
(years)
PercentageSelf-control task
of men
HRV measure
Duration of HRV
recording (minutes)
Recording device
Handling of respiration
Number of effects
Burg, Wolf, and

Michalak (2012)
Cellini et al. (2013)
Cocia, Uscatescu, and
Rusu (2012)
Demaree, Robinson,
Everhart, and
Schmeichel (2004)
Duschek et al. (2009)
Gaebler, Daniels,
Lamke, Fydrich, and
Walter (2013)*
Geisler and Kubiak
(2009)
Gillie, Vasey, and
Thayer (2013)
Hansen et al. (2009)
23
23.8
13
Mindful breathing exercise
RMSSD, SDNN
Ambulatory ECG
No control of respiration
14
74
22
19.0
0
15
D2 task
Emotional spatial cueing task
HF
HF
3
5
Ambulatory ECG
ECG
4
2
103
18.7
50
Exaggeration of facial expression

during mood induction
HF
ECG
60
21
24.5
29.1
47
24
D2 task
Trail-making test
HF
HF
5
5
ECG
PWD
3
2
53
21.6
HF
HRM
85
18.4
36
HF
ECG
65
23.1
100
HF
Ambulatory ECG
12
49
23
100
RMSSD
Ambulatory ECG
16
Kaufmann, Vgele,
Stterlin, Lukito, and
Kbler (2012)
Kimhy et al. (2013)
34
26.3
47
Rumination after negative

feedback
Memory inhibition (think/no-think
task)
Working memory test
California Computerized
Assessment Package (CALCAP)
Working memory test
California Computerized
Assessment Package (CALCAP)
P300 visual spelling Matrix
RMSSD
HF
ECG
817
57.1
44.2
HF
ECG
Respiration adjusted effects
Krypotos et al. (2011)

Laborde and Raab
(2013)
Laborde and Raab
(2013)
54
90
21.4
25.5
19
100
RMSSD
HF RMSSD SDNN
10
3
ECG
Ambulatory ECG
6
6
30
22.4
100
Backward counting
Backward digit span
Stop and go switch task
Emotional stop signal Task
Decision making after mood
induction
Decision Making after Mood
Induction
Ambulatory ECG
Luft et al. (2009)

Meule, Vgele, and
Kbler (2012)
Ohira et al. (2013)
Park et al. (2012)
Park, Van Bavel, Vasey,
and Thayer (2013)a
30
47
18.5
23.7
77
0
Cogstate Computerized System

Flanker Task
2
5
Ambulatory ECG
HRM
2
4
20
45
29
32.6
20
20
100
29
0
Stochastic reversal learning task

HF
RMSSD
SDNN
HF
HF
RMSSD
HF
HF
HF
RMSSD
5
5
5
PWD
ECG
ECG
1
10
8
Park, Van Bavel et al.

(2013)b
25
20
HF
RMSSD
ECG
Park, Vasey, Van Bavel,

& Thayer (2013)
Pu et al. (2010)
73
20
42
Letter detection task
HF
ECG
No control of
respiration/respiration adjusted
effects c
No control of
respiration/respiration adjusted
effects c
136
18.9
49
HF
ECG
16
Reynard et al. (2011)

Segerstrom and
Solberg Nes (2007)
Volokhov and Demaree
(2010)
59
168
27.2
19.1
27
47
Working memory test after mood

induction
Persistence on Anagram task
Persistence on Anagram task
RMSSD
RMSSD
5
n.a.
Ambulatory ECG
ECG
2
2
113
19.2
44
HF
Ambulatory ECG
respiration adjusted effects
Emotion regulation after mood

induction
Authors
17
Note: ECG = electrocardiography. HF = high frequency component. HRM = heart rate monitor. PWD = pulse wave based device. RMSSD = root mean square of successive differences of RR intervals. SDNN = standard deviation of NN
intervals. n.a. = not assessed. * = only healthy subsample used in this meta-analysis. a, b = Experiment 1 and Experiment 2 in one study. c = some effects were reported adjusted for respiration.
18
self-control domains against the grand mean of all three categories.

Meta-regression did not reveal a signicant relationship between
self-control domains and effect size magnitude (see Table 3).
Averaged effect sizes for each category are displayed in Table 3.
Signicant effect sizes were found for emotion regulation and for
attention control.
Study
Kaufmann et al. (2012)
Kaufmann et al. (2012)
Ohira et al. (2013)
Park et al. (2013) Exp 1
Park, Vasey et al. (2013)
Cocia et al. (2012)
Cocia et al. (2012)
Meule et al. (2012)
Meule et al. (2012)
Meule et al. (2012)
Meule et al. (2012)
Gaebler et al. (2013)
Gaebler et al. (2013)
Tasks
P300 Visual Spelling Matrix
P300 Visual Spelling Matrix
Stochastic Reversal Learning Task
Emotional Spatial Cueing Task
Letter Detection Task
Flanker Task
Flanker Task
Flanker Task
Flanker Task
TrailMaking Test
TrailMaking Test
3.2.1.2. Executive functions. Of the 132 extracted effects, four

effects were classied as shifting, 48 effects as updating and
78 effects as inhibiting. The remaining two effects could not be
classied into one of the three categories and were excluded
from meta-regression. Again, weighted effect coding was used.
Meta-regression did not reveal a signicant relationship between
Effect size Lower CI Higher CI Weight

0.24
0.42
0.00
0.42
0.07
0.17
0.14
0.24
0.07
0.41
0.38
0.41
0.00
0.13
0.40
0.17
0.09
0.01
0.18
0.46
0.10
0.27
0.38
1.05
0.12
0.05
0.28
0.34
0.14
0.04
0.10
0.05
0.58
0.56
0.11
0.07
0.48
0.04
0.46
0.55
0.24
0.14
0.31
0.00
0.01
0.01
0.42
0.55
0.02
0.59
0.51
0.42
0.24
0.02
0.13
0.77
0.10
0.48
0.60
0.19
0.04
0.11
0.16
0.34
0.19
0.35
0.10
0.09
0.59
0.78
0.48
0.81
0.31
0.22
0.53
0.63
0.45
0.82
0.76
0.83
0.42
0.29
0.82
0.25
0.32
0.41
0.60
0.90
0.33
0.24
0.85
1.61
0.35
0.28
0.51
0.58
0.44
0.25
0.40
0.24
1.05
1.04
1.74
1.74
2.35
0.39
0.39
0.39
0.39
0.39
0.39
0.39
0.39
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.44
0.44
0.44
0.44
0.44
0.44
2.58
2.58
1.05
1.05
1.05
1.05
1.17
1.17
1.5
0.5
0.5
1.5
Effect size (Fisher's z metric)
Study
Kimhy et al. (2013)
Kimhy et al. (2013)
Kimhy et al. (2013)
Laborde & Raab (2013) Exp 1
Burg et al. (2012)
Burg et al. (2012)
Demaree et al. (2004)
Luft et al. (2009)
Luft et al. (2009)
Geisler & Kubiak (2009)
Tasks
Backward Counting
Digits Backward Span
Stop and Go Switch Task
D 2 Task
D 2 Task
D 2 Task
D 2 Task
Decision Making after Mood Induction
Mindful Breathing Exercise
Mindful Breathing Exercise
Exaggerate Expression during Mood Induction
D 2 Task
D 2 Task
D 2 Task
Rumination after Negative Feedback

0.02
0.04
0.01
0.78
0.73
0.73
0.03
0.17
0.22
0.01
0.26
0.13
0.11
0.29
0.27
0.18
0.38
0.41
0.12
0.38
0.38
0.10
0.08
0.45
0.45
0.00
0.04
0.04
0.04
0.52
0.09
0.03
0.08
1.37
1.32
1.32
0.62
0.21
0.15
0.37
0.12
0.25
0.27
0.09
0.11
0.20
0.00
0.03
0.26
0.06
0.06
0.39
0.33
0.02
0.02
0.38
0.30
0.30
0.30
0.11
0.05
0.10
0.06
0.18
0.13
0.13
0.56
0.55
0.60
0.39
0.63
0.51
0.49
0.66
0.64
0.56
0.75
0.79
0.50
0.82
0.82
0.19
0.49
0.88
0.88
0.38
0.22
0.22
0.22
0.94
2.63
2.63
2.63
0.42
0.42
0.42
0.42
0.53
0.53
0.53
0.53
0.53
0.53
0.53
0.53
0.53
0.53
0.53
0.53
1.31
1.31
1.06
1.06
1.06
1.47
1.47
1.58
1.58
1.58
2.80
1.5
0.5
0.5
1.5
Fig. 2. Forest plot displaying effect sizes (in Fishers z metric), condence intervals, weights for each effect size and the average effect size based on the random effects
model with RVE and small sample corrections. The edges of average effect size polygon represent 95% condence intervals of the average effect size. CALCAP = California
Computerized Assessment Package. CI = condence interval.

Study
19
Tasks
0.32
0.26
0.26
0.21
0.23
0.22
0.22
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.24
0.23
0.00
0.00
0.00
0.00
0.31
0.22
0.24
0.27
0.28
0.21
0.00
0.00
0.02
0.07
0.04
0.34
0.20
0.29
Working Memory Test

CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
Working Memory Test
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
Working Memory Test
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
CALCAP
Working Memory Test
Working Memory Test
CALCAP
CALCAP
CALCAP
Emotional Stop Signal Task
0.03
0.03
0.03
0.08
0.05
0.07
0.51
0.29
0.29
0.29
0.29
0.29
0.29
0.29
0.04
0.05
0.25
0.25
0.25
0.25
0.06
0.03
0.01
0.52
0.03
0.46
0.25
0.25
0.26
0.21
0.23
0.07
0.07
0.02
0.61
0.54
0.54
0.50
0.52
0.51
0.07
0.29
0.29
0.29
0.29
0.29
0.29
0.29
0.53
0.52
0.25
0.25
0.25
0.25
0.56
0.46
0.49
0.02
0.53
0.04
0.25
0.25
0.29
0.34
0.32
0.62
0.48
0.56
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.75
0.75
0.75
0.75
0.75
0.75
1.5
0.5
0.5
1.5
Study
Tasks
Pu et al. (2010)
Working Memory Test after Mood Induction
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Pu et al. (2010)
Persistence on Anagram Task
Segerstrom & Solberg Nes (2007)
Segerstrom & Solberg Nes (2007)
Emotion Regulation after Mood Induction
Volokhov & Demaree (2010)
Volokhov & Demaree (2010)
Emotion Regulation after Mood Induction
Gillie, Vasey & Thayer (2013)
Memory Inhibition
Gillie, Vasey & Thayer (2013)
Memory Inhibition
Summary
Average effect size
0.30
0.17
0.01
0.12
0.05
0.09
0.13
0.02
0.08
0.00
0.00
0.00
0.00
0.12
0.00
0.00
0.13
0.31
0.22
0.22
0.36
0.30
0.41
0.23
0.54
0.41
0.23
0.12
0.19
0.15
0.37
0.22
0.32
0.24
0.24
0.24
0.24
0.36
0.24
0.24
0.39
0.05
0.00
0.00
0.03
0.03
0.18
0.00
0.06
0.07
0.25
0.36
0.29
0.33
0.11
0.26
0.16
0.24
0.24
0.24
0.24
0.12
0.24
0.24
0.13
0.57
0.44
0.44
0.69
0.63
0.64
0.47
0.16
0.09
0.23
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
2.34
2.34
2.68
2.68
1.86
1.86
2.60
2.60
1.5
0.5
0.5
1.5
Fig. 2. (Continued).
executive functions and effect size magnitude (see Table 3). Table 3
shows estimates of the effect sizes within the three categories. A
signicant pooled effect size was only observed for inhibiting.
Taken together, these results provide little evidence that task
characteristics may moderate the HRV-self-control connection.
3.2.2. Methodological factors of HRV assessment
3.2.2.1. Recording device. Of the 132 effects, 74 were categorized as
recorded with stationary ECGs, 50 with ambulatory ECG systems,
ve with HRMs and three with pulse wave based measures. Metaregression using the weighted-effects coded category system for
recording devices did not show a signicant relationship between

kind of device and magnitude of effect size (see Table 4). Average
effects for the four recording devices are displayed in Table 4.
3.2.2.2. HRV measures. Of the 132 effects, 78 were based on the
HF component, 49 were based on RMSSD and ve effects were
based on SDNN. As displayed in Table 4, meta-regression with the
weighted-effects coded variable HRV measure did not yield a signicant relationship between type of HRV measure and magnitude
of effect size. Averaging effect sizes within the three HRV measures
(see Table 4) yielded signicant effect sizes for HF and RMSSD, with
20
Table 3
Meta-regression results and average effect sizes for self-control task characteristics.
Meta-regression
B
Self-control domains
0.039
Emotion regulation
0.023
Attention control
Cognitive control
0.145
Sets of executive functions
Shifting
0.067
0.039
Updating
0.021
Inhibiting
Average effect sizes
SE
df
95% CI
df
I2 (%)
k1 (effects)
k2 (studies)
0.033
0.029
0.035
0.248
0.426
0.122
1.19
0.81
4.21
19.0
20.5
1.15
0.171
0.136
0.006
0.064; 0.274
0.010; 0.231
0.130; 0.119
0.005
0.035
0.667
3.54
3.00
0.58
10.3
13.2
1
38.66
62.63
0
59
68
5
12
16
2
0.141
0.035
0.023
0.654
0.293
0.379
0.49
1.11
0.91
3.53
10.4
11.5
0.189
0.082
0.158
0.271; 0.578
0.033; 0.195
0.072; 0.241
0.270
0.122
0.001
1.38
1.89
3.90
2.68
4.64
15.9
69.13
47.63
52.35
4
48
78
4
7
18
Note: meta-regression for self-control domains and sets of executive functions based on weighted effects coding. df = degrees of freedom. SE = Standard error. k1 = number of
effects. k2 = number of studies.
Table 4
Average effect sizes for HRV assessment (based on subgroup analyses).
Meta-regression
Recording device
Stationary ECG
Ambulatory ECG
HRM
PWD
HRV measures
HF
RMSSD
SDNN
Average effect sizes
SE
df
95% CI
0.004
0.012
0.084
0.135
0.034
0.045
0.232
0.281
0.901
0.815
0.771
0.709
0.13
0.24
0.24
0.48
14.9
11.6
1.17
1.11
0.143
0.138
0.252
0.278
0.056; 0.230
0.064; 0.277
0.992; 0.997
0.997; 0.999
0.005
0.002
0.052
0.022
0.031
0.100
0.833
0.944
0.653
0.21
0.07
0.10
16.2
14.0
2.12
0.171
0.166
0.177
0.08; 0.258
0.108; 0.230
0.188; 0.499
df
I2 (%)
k1 (effects)
k2 (studies)
0.004
0.058
0.480
0.500
3.55
2.31
1.07
1.00
11.8
6.34
1.0
1.0
60.32
48.82
70.98
0
74
50
5
3
114
8
2
2
<0.001
<0.001
0.169
3.98
5.85
2.13
18.1
8.52
1.96
60.67
8.83
0
78
49
5
21
11
3
Note: meta-regression for recording device and HRV measures based on weighted effects coding. ECG = electrocardiography. HRM = heart rate monitor. PWM = pulse wave
based measure. HF = high frequency component. RMSSD = root mean square of successive differences of RR intervals. SDNN = standard deviation of NN intervals. df = degrees
of freedom. SE = standard error. k1 = number of effects. k2 = number of studies.
large heterogeneity in HF (I2 = 60.67%) and a homogeneous effect

size for RMSSD (I2 = 8.83%).
3.2.2.3. Duration of HRV assessment. Duration of assessment of
HRV at rest varied from two minutes to ten minutes, with an
overall mean of 4.46 1.72 min. One study (Segerstrom & Solberg
Nes, 2007) did not provide information about duration of HRV
assessment and was excluded from the following analysis. Metaregression did not reveal a signicant relationship between length
of measurement and magnitude of effect size (b = 0.012, SE = 0.018,
t = 0.68, df = 3.25, p = 0.543, 95% CI [0.042; 0.065], I2 = 58.23%).
To conclude, moderation analyses did not reveal a moderation of
the HRV-self-control connection by aspects related to HRV assessment.
3.2.3. Age and sex
Mean age of the study samples varied between 18.5 and 57.1
years, with a mean age of 21.4 5.9 years over all studies. Metaregression revealed a trend for a negative relationship between
mean age and magnitude of effect size (b = 0.005, SE = 0.001,
t = 3.97, df = 1.43, p = 0.097, 95% CI [0.013; 0.003], I2 = 45.71). Percentage of male participants varied between 0% and 100% with an
overall mean of 50.22% 37.8 of males. Meta-regression did not
reveal a signicant relationship between percentage of males and
magnitude of effect size (b = 0.061, SE = 0.093, t = 0.65, df = 10.0,
p = 0.539, 95% CI [0.268; 0.147]).
Taken together, we did not nd signicant moderation of the
HRV-self-control connection by age or sex.
Study quality ranged from 36% to 85% with a mean of 63 0.12%,
indicating that some studies only met 36% of the 13 quality criteria.
Only 15 out of 26 studies provided information or stated exclusion
criteria referring to confounders such as medication, BMI, psychiatric or somatic diseases, or smoking status. In 16 studies, posture
during HRV assessment was stated, artifact correction was stated in

19 studies, nine reported a pre-rest period before HRV assessment,
and 18 studies dened frequency bands for the HF component. Only
one study reported the association between HRV and self-control
adjusted for confounders.
Overall, there was a negative, but nonsignicant relationship between study quality and the magnitude of the effect
size, (b = 0.288, SE = 0.228, p = 0.248, t = 1.26, df = 6.85, 95% CI
[0.831; 0.284], I2 = 54.8). Meta-regressions analyzing the single
components of the study quality score yielded signicant moderating effects only for adjustment of confounders (b = 0.168,
SE = 0.033, p < 0.001, t = 5.09, df = 22.1, 95% CI [0.236; 0.099],
I2 = 46.39). However, only one study with three effects reported
effects adjusted for confounders. Objective measures of self-control
performance were also signicantly associated with smaller effect
sizes (b = 0.178, SE = 0.072, p = 0.046, t = 2.48, df = 6.2, 95% CI
[0.351; 0.004], I2 = 52.23). Only six studies (14 effects) used selfreport measures.
3.3. Publication bias
The visual inspection of the funnel plot based on the second data
set consisting of one weighted mean effect size per study showed
asymmetry indicative of publication bias. Eggers test revealed evidence for publication bias (t = 3.003, df = 24, p = 0.006). Applying the
trim-and-ll method resulted in the estimation of 10 studies missing on the left side of the funnel plot (Fig. 3). Adjusting the average
effect size for missing studies revealed a nonsignicant average
effect of r = 0.04, 95% CI [0.038; 0.110], p = 0.334; compared to
an average effect size of r = 0.14, 95% CI [0.074; 0.208], p < 0.001
without adjustment for missing studies.
Outlier diagnostics identied one study (Kimhy et al., 2013)
as an outlier and inuential case. The study had the largest sample size (n = 817) and the highest mean age (57.1 11.2 years)
of study participants. In addition, HRV assessment and experi-
0.151
0.302
Standard Error
0.000
0.60
0.20
0.20
0.40
0.60
0.80

Fig. 3. Funnel plot displaying effect sizes by standard errors. White circles represent studies missing from the funnel plot as estimated by the trim-and-ll method.
Funnel plot based on averaged effect sizes per study.
mental self-control tasks were conducted during two separate

sessions with a time lag between sessions of one to 61 months
(average time 24.2 14.1 months). Repeating meta-analysis and
meta-regression after exclusion of this study only had an almost
negligible effect on the pooled effect size (average effect size
r = 0.16, 95% CI [0.098; 0.223], p < 0.001). However, exclusion of this
study led to a reduction of the observed publication biasin consequence neither Eggers test (t = 0.227, df = 23, p = 0.823) nor the
trim-and-ll-method indicated signicant evidence of publication
bias.
4. Discussion
The current meta-analysis aimed at evaluating the association
between HRV at rest and self-control performance in laboratory
settings. We also investigated potential moderators such as sex
and age, methodological aspects of HRV assessment, task characteristics, and study quality. Our meta-analysis showed a signicant
small association between HRV measured at rest and self-control
performance with moderate heterogeneity in the true effect size
between studies. This result indicates that higher trait-HRV is
indeed related to better self-control performance, but the effect
size is rather small. However, adjusting for the observed publication bias drastically reduced the relationship between HRV and
self-control, indicating no signicant association between HRV and
self-control.
Heterogeneity was moderate suggesting moderation by third
variables. Based on the current literature, possible moderators were
included in the analyses. Formal moderation testing with metaregressions did not yield any signicant results. But one has to take
into account that our analyses were based on a medium-size sample
of studies (k = 26) and that many covariates were unbalanced, possibly leading to power issues for the meta regression with robust
variance estimation (Tanner-Smith & Tipton, 2013; Tipton, 2014).
Given the observed variability in average effect sizes within the
categories of sets of executive functions, self-control domains, and
recording devices, it seems too early to conclude that these aspects
do not moderate the HRV-self-control connection.
In addition to the reduced power for moderator analyses, several other reasons may explain the lack of moderating effects: As
for sets of executive functions, most complex self-control tasks may
require more than one executive component (Duckworth & Kern,
2011; Hofmann et al., 2012; Miyake et al., 2000). This notion is also
supported by the results of Miyakes latent class analysis which
showed that the three executive functions are separable constructs,
but are also moderately correlated. However, to categorize the
21
tasks we had to determine the most prominent executive function

of each task. This procedure may not adequately cover the complex interplay of different executive functions, which are necessary
to enable successful self-control. Finally, denitions and classications of executive functions vary widely among researchers
(Suchy, 2009) and Miyakes taxonomy of executive functions constitutes one way to classify executive functions, which may not be
exhaustive (Miyake et al., 2000). Whereas further research is necessary to derive consensus about the denition and classication
of executive functions, it has yet to be determined which executive functions play a role for successful performance in self-control
tasks. In terms of self-control domains, the association between
HRV and self-control was quite similar for emotion regulation and
attention control. Smaller effects were observed for the cognitive
control category, but the number of effects pertaining to cognitive control was very small. In sum, we assume that the lack of a
moderating effect of self-control domains and executive function
may indicate that the inhibitory capacity reected by trait HRV is
of equal importance for aspects of cognitive as well as of emotional
regulation, as hypothesized by the Neurovisceral Model (Thayer &
Lane, 2000) and Polyvagal Theory (Porges, 2007).
With respect to methodological factors related to HRV assessment and computation, we limited the meta-analysis to studies
reporting HRV metrics meeting the guidelines of the Task Force of
the European Society of Cardiology and the North American Society
of Pacing and Electrophysiology (1996), i.e., RMSSD, SDNN, and HF
component. This led to the exclusion of studies that used RSA estimated with the PorgesBohrer-method or Peak-to-trough-metric.
In consequence, we may have already excluded some sources of
variance in HRV metrics, and cannot rule out that studies conducted
in the framework of Polyvagal Theory may have been underrepresented in our meta-analysis. However, only one study out of 327
screened full text articles was excluded due to the use of these RSA
metrics. Thus, we assume that the exclusion of these RSA metrics
might not heavily impact the generalization of our results.
Reduced variance may also explain the lack of a moderating
effect of age: the effect of aging on HRV is more pronounced in
the age groups older than 30 years of age (De Meersman & Stein,
2007). Except one study (Kimhy et al., 2013), all studies included
in the current meta-analysis examined participants aged between
18 and 30 years. This limits the generalization of our results to
this age group. The study of Kimhy and colleagues indicated that
the effect sizes for participants with a mean age of 57 years were
almost zero, but this might also be caused by the time lag (one to
61 months) between assessment of self-control performance and
HRV measurement. Thus, we cannot determine the role of age on
the HRV-self-control connection. Further research on the association between HRV and self-control in older participants is necessary
to explore, if an age-specic decline in HRV also affects self-control
performance. We also limited our analysis to adult populations and
excluded studies on infants, children, and adolescents. However,
much research conducted in the framework of Porges Polyvagal
Theory focused on developmental aspects such as RSA and adaptive functioning in infants, children, and adolescents (Beauchaine,
2001). Many of these studies indicate differential relationships
between RSA and psychological adaptation to changing demands
in infancy and childhood (Beauchaine, 2001). Thus, the exclusion of
studies on these age groups may also explain the lack of a moderating effect of age. Our analysis did not reveal a moderating effect
of gender on the HRV-self-control connection. However, it can be
assumed that higher HRV in women does not automatically lead
to better self-control performance, but may also depend on task
characteristics and the use of self-control strategies.
The current meta-analysis supports the assumptions of the
Polyvagal Theory of Porges (Porges, 2001, 2007) and Thayers Neurovisceral Model (Thayer et al., 2009; Thayer & Lane, 2000), which
22
postulate that higher HRV is related to concomitants of better selfcontrol. However, the magnitude of the effect size was actually
lower than expected. Several reasons might have contributed to
the small effect size: rst, we limited the meta-analysis to studies on trait HRV. Whereas the Neurovisceral Integration Model
focusses on the role of trait HRV, Polyvagal Theory emphasizes
both trait measures of HRV and cardiac reactivity as reected by
state or phasic HRV (Beauchaine, 2001; Porges, 2001, 2007). A
large body of evidence supports the importance of cardiac reactivity for psychological functioning (see Graziano & Derenko,
2013; for a meta-analysis). Integrating studies which examine
cardiac reactivity during self-control tasks might perhaps have
yielded a larger effect size. Second, we included only studies which
assessed self-control performance in a laboratory setting. Laboratory self-control tasks may have little in common with self-control
performance in daily life. Furthermore, performance in these tasks
might be of little personal relevance for participants. This may lead
to a reduced motivation to give their best in the tasks. In turn,
the relationship between HRV and self-control performance may
be underestimated. Research within the Strength model of SelfControl supports the notion that motivation affects self-control
performance in the laboratory: Participants receiving incentives
or information about the relevance of the task at hand performed
better at self-control tasks (e.g., Muraven & Slessareva, 2003).
Taken together, these reasons might have contributed to the small
overall effect size. However, previous meta-analyses on HRV and
different self-control outcomes also suggested small to moderate
association with depression (Rottenberg, 2007), alcohol dependence (Quintana, McGregor, Guastella, Malhi, & Kemp, 2013) or
externalizing and internalizing problems in children (Graziano &
Derenko, 2013). Thus, our results support the notion that HRV
shows small to moderate associations with different aspects of
self-control.
The results of the current meta-analysis have to be interpreted
with great caution, as we observed a considerable publication bias.
Adjustment for missing studies indicated that the average effect
size may be reduced drastically, resulting in a nonsignicant effect
size close to zero. Publication bias was reduced after exclusion of
one study identied as an outlier and inuential case. Whereas
this study (Kimhy et al., 2013) differed from other included studies in mean age and the time lag between self-control and HRV
assessment, the study also had the largest sample size (n = 817) and
reported effects after adjustment for medical conditions, demographics, health behaviors and respiratory rate. The impact of this
study on the results of the meta-analysis is difcult to determine:
on the one hand, the study might constitute an outlier due to age
differences or the time lag between HRV and self-control assessment. On the other hand, based on the large sample size and the
careful control for confounding variables, the study might provide a more valid estimate of the HRV-self-control connection.
To conclude, we believe that this study cannot fully explain the
observed publication bias. In consequence, we assume that the
association between HRV and self-control could actually be lower
than the observed effect of r = 0.15. This notion is also supported
by the results of the moderator-analysis of study quality: although
nonsignicant, higher study quality was related to smaller effect
sizes. This observation is also known from meta-analysis in other
elds of research (Cuijpers et al., 2010; Rosenbaum et al., 2014).
As our study quality score was based on information provided in
the manuscripts, this quality scores mostly reect the quality of
reporting and only in part the quality of the study conduction. In
accordance with an earlier meta-analysis on HRV in functional disorders (Tak et al., 2009), our study highlights the need of a thorough
reporting of methodological aspects covering detailed information about the setting of HRV assessment (posture, pre-rest-period,
device, length of monitoring), the detection and handling of artifacts, and clear information about HRV metrics (frequency bands,
units). Researchers should also consider reporting clear eligibility criteria and adequate control of confounding variables such as
medical conditions or medication that might affect HRV.
In addition to publication bias, we also observed evidence for
selective reporting of results, with many studies reporting exact
statistics only for signicant results. In consequence, of the 132
extracted effect sizes 21 effects had to be included as zero effects
due to a lack of further information to compute exact effect sizes. As
a meta-analysis can only provide a reliable and valid estimate of an
overall effect size, if based on a representative sample of research,
our results highlight the need to further encourage authors as well
as journal editors to report and publish nonsignicant and counterintuitive results.
Several limitations of the current meta-analysis have to be mentioned: rst, the number of studies that met eligibility criteria was
moderate, which limits the power of the analyses, in particular for
moderator analyses. Second, when studies reported results as nonsignicant and information to compute exact estimates of effect
sizes was not available or not provided by the authors upon request,
these effects (n = 21) were included with a conservative effect size
of r = 0 (Rosenthal, 1995). Although repeating the analysis without
these effects did not affect the overall estimate of the effect size,
model-based methods to impute missing effect sizes such as multiple imputation might have been a more adequate approach to
deal with missing effect sizes (Pigott, 2012). However, the robust
variance estimation method for dependent effects is not available
for multiple imputed meta-analytic data sets. Third, we had to
refrain from formal moderation testing for respiration, as information about the control of respiration during HRV assessment was
difcult to extract, because most studies did not explicitly report
on the control of respiration. Fourth, although we aimed at reducing the potential inuence of cardiovascular disease by limiting the
study to participants aged 65 years or younger, we cannot fully
rule out the presence of undetected cardiovascular disease in the
included studies. Finally, due to the exclusion of studies on children and studies reporting RSA estimated with the PorgesBohrer
method, our results may only partially be generalizable to Polyvagal
Theory.
To conclude, to our knowledge, the current meta-analysis is
the rst to integrate studies examining the relationship between
HRV and self-control in laboratory tasks. The available evidence is
not sufcient to conclusively determine the association between
HRV and performance in self-control tasks: Our meta-analysis
supported the notion that higher HRV may be related to better
self-control, but the effect size was small. One study identied
as an outlier further complicated the interpretation of the association between HRV and self-control, because the study had the
largest sample size, but differed in several aspects from the other
included studies. Finally, we observed signicant publication bias
that suggests that there might not be an association between HRV
and self-control performance.
Appendix A. Coding scheme and classication for task
characteristics.
23
Self-control domains
Emotion regulation
Attention control
Cognitive control
Emotion regulation after mood induction (reappraisal,

negative lm)
Exaggeration of emotion expression during mood
induction (overall facial valence, positive and negative
lms)
induction (facial valence, positive lms)
induction (facial valence, negative lms)
WMT after mood induction (reaction time, post-lm,
verbal, negative lm)
WMT after mood induction (error rate, post-lm, spatial,
negative lm)
spatial, negative lm)
Emotional spatial cueing task (engagement, LSF fearful
faces)
Emotional spatial cueing task (engagement, HFS fearful
faces)
Emotional spatial cueing task (engagement, BSF fearful
faces)
Emotional spatial cueing task (disengagement, HSF
fearful faces)
Emotional spatial cueing task (disengagement, BSF
fearful faces)
Emotional spatial cueing task (disengagement, LSF
fearful faces)
Letter detection task (overall reaction time)
Letter detection task (reaction time, fearful faces, low
load)
Letter detection task (reaction time, fearful faces, high
load)
Emotional spatial cueing task (reaction time,
engagement to threat)
Emotional spatial cueing task (reaction time,
disengagement to threat)
Decision making after mood induction (mean quality of
options, negative condition)
Decision making after mood induction (decision time,
negative condition)
Rumination after negative feedback
Emotional stop signal task (reaction time, initiate
reaction overall)
reaction, negative trails)
reaction, neutral trials)
Emotional stop signal task (reaction time, inhibit
reaction, negative trials)
reaction, neutral trails)
reaction overall)
Emotional spatial cueing task (reaction time, cue
validity, BSF fearful faces)
validity, LSF fearful faces)
WMT after mood induction (reaction time differences,
spatial task, negative lm)
verbal task, negative lm)
WMT after mood induction (accuracy, spatial task,
negative lm) WMT after mood induction (accuracy,
verbal task, negative lm)
WMT after mood induction (error rate, post-lm, verbal,
negative lm)
Emotion regulation after mood induction (suppression,
negative lm)
validity, fearful faces, valid cues)
validity, fearful faces, invalid cues)
Flanker task (overall errors)

Flanker task (reaction time
difference incongruentcongruent
trials)
Stop and go switch task (reaction
time)
Stochastic reversal learning task (response bias)

Trail-making test (part B)
P300 visual spelling matrix (accuracy)
Backward counting
Digit backwards
Cogstate computerized system (reaction time)
Cogstate computerized system (errors)
WMT (correct answers)
CPT (total reaction time)
CPT (correct answers)
CPT (errors)
SPM 1 (reaction time)
SPM 1 (correct answers)
SPM 1 (errors)
SPM 2 (errors)
WMT (errors)
SRT (reaction time)
SRT (mean reaction time)
CRT (true positive)
CRT (reaction time)
CRT (errors)
WMT (nonresponses)
CRT (overall correct)
CRT (overall errors)
CPT (mean reaction time)
spatial task, neutral lm)
verbal task, neutral lm)
WMT after mood induction (accuracy, spatial task,
neutral lm)
WMT after mood induction (accuracy, verbal task,
neutral lm)
neutral lm)
verbal, neutral lm)
WMT after mood induction (error rate, post-lm, spatial,
neutral lm)
spatial, neutral lm)
Letter detection task (reaction time, neutral faces, high
load)
Letter detection task (overall errors)
D2 task (number of Items)
D2 task (accuracy)
D2 task (concentration performance)
D2 task (percentage of errors)
D2 task (total number of stimuli)
D2 task (attentional capacity)
validity, HSF neutral faces)
validity, BSF neutral faces)
validity, LSF neutral faces)
Persistence on anagrams (time to complete solvable
anagrams)
Persistence on unsolvable anagrams (persistence time,
rst anagram)
Memory inhibition (same probe test, independent probe
test)
Letter detection task (reaction time, neutral faces, low
load)
Trail-making test (part A)
validity, neutral faces, valid cues)
validity, neutral faces, invalid cues)
24
Sets of executive functions

Shifting
Updating
Inhibition
Emotion regulation after mood induction

(reappraisal, negative lm)
Stochastic reversal learning task
(response bias)
Trail-making test (part B)
P300-visual spelling matrix (accuracy)

Backward counting
Digit backwards
Cogstate computerized system (reaction
time)
Cogstate computerized system (errors)
WMT (correct answers)
CPT (total reaction time)
CPT (correct answers)
CPT (errors)
SPM 1 (errors)
SPM 2 (errors)
WMT (errors)
SRT (reaction time)
SRT (mean reaction time)
CRT (true positive)
CRT (reaction time)
CRT (errors)
WMT (nonresponses)
CRT (overall correct)
CRT (overall errors)
CPT (mean reaction time)
WMT after mood induction (reaction time
differences, spatial task, neutral lm)
WMT after mood induction (reaction time
differences, verbal task, neutral lm)
WMT after mood induction (accuracy,
spatial Task, neutral lm)
WMT after mood induction (accuracy, verbal
task, neutral lm)
WMT after mood induction (error rate,
post-lm, verbal, neutral lm)
WMT after mood induction (reaction time,
post-lm, verbal, neutral lm)
WMT after mood induction (error rate,
post-lm, spatial, neutral lm)
WMT after mood induction (reaction time,
post-lm, spatial, neutral lm)
Emotional spatial cueing task (engagement, LSF fearful faces)

Emotional spatial cueing task (engagement, HFS fearful faces)
Emotional spatial cueing task (engagement, BSF fearful faces)
Emotional spatial cueing task (disengagement, HSF fearful faces)
Emotional spatial cueing task (disengagement, BSF fearful faces)
Memory inhibition (same probe test, independent probe test)
Emotional spatial cueing task (disengagement, LSF fearful faces)
Letter detection task (overall reaction time)
Letter detection task(reaction time, fearful faces, low load)
Letter detection task (reaction time, fearful faces, high load)
Emotional spatial cueing task (reaction time, engagement to
threat)
Emotional spatial cueing task (reaction time, disengagement to
threat)
Decision making after mood induction (mean quality of options,
negative condition)
Decision making after mood induction (decision time, negative
condition)
Rumination after negative feedback
Emotional stop signal task (reaction time, initiate reaction
overall)
Emotional stop signal task (reaction time, initiate reaction,
negative trails)
Emotional stop signal task (reaction time, initiate reaction,
neutral trials)
Emotional stop signal task (reaction time, inhibit reaction,
negative trials)
Emotional Stop signal task (reaction time, inhibit reaction,
neutral trails)
Emotional stop signal task (reaction time, inhibit reaction
overall)
Emotional spatial cueing task (reaction time, cue validity, BSF
fearful faces)
Emotional spatial cueing task (reaction time, cue validity, LSF
fearful faces)
WMT after mood induction (reaction time differences, spatial
Task, negative lm)
WMT after Mood Induction (reaction time differences, verbal
task, negative lm)
WMT after mood induction (accuracy, spatial task, negative lm)
WMT after mood induction (accuracy, verbal Task, negative lm)
negative lm)
Emotion regulation after mood induction (suppression, negative
lm)
Letter detection task (reaction time, neutral faces, high load)
Letter detection task (overall errors)
D2 task (number of Items)
D2 task (accuracy)
D2 task (concentration performance)
D2 task (total number of stimuli)
D2 task (attentional capacity)
Emotional spatial cueing task (reaction time, cue validity, HSF
neutral faces)
Flanker task (overall errors)
Flanker task (reaction time difference incongruentcongruent
trials)
Stop and go switch task (reaction time)
Emotional spatial cueing task (reaction time, cue validity, fearful
faces, valid cues)
Emotional spatial cueing task (reaction time, cue validity, fearful
faces, invalid cues)
Emotional spatial cueing task (reaction time, cue validity,
neutral faces, valid cues)
Emotional spatial cueing task (reaction time, cue validity,
neutral faces, invalid cues)
Note: CPT = Continuous performance test (California Computerized Assessment Package Abbreviated Version (CALCAP)), CRT = CRT (CALCAP). SPM = SPM (CALCAP). SRT = simple reaction time (CALCAP). BSF = broad spatial frequency. HSF = high spatial frequency, LSF = low
spatial frequency. WMT = working memory test.
References
Antelmi, I., de Paula, R. S., Shinzato, A. R., Peres, C. A., Mansur, A. J., & Grupi, C. J.
(2004). Inuence of age, gender, body mass index, and functional capacity on
heart rate variability in a cohort of subjects without heart disease. American
Journal of Cardiology, 93(3), 381385. http://dx.doi.org/10.1016/j.amjcard.
2003.09.065
Appelhans, B. M., & Luecken, L. J. (2006). Heart rate variability as an index of
regulated emotional responding. Review of General Psychology, 10(3), 229240.
http://dx.doi.org/10.1037/1089-2680.10.3.229
Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of
self-control. Current Directions in Psychological Science, 16(6), 351355. http://
dx.doi.org/10.1111/j.1467-8721.2007.00534.x
Beauchaine, T. (2001). Vagal tone, development, and Grays motivational theory:
toward an integrated model of autonomic nervous system functioning in
psychopathology. Developmental Psychopathology, 13(2), 183214. http://dx.
doi.org/10.1017/s0954579401002012
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction
to meta-analysis. Chichester, U.K: John Wiley & Sons.
Britton, A., Shipley, M., Malik, M., Hnatkova, K., Hemingway, H., & Marmot, M.
(2007). Changes in heart rate and heart rate variability over time in
middle-aged men and women in the general population (from the Whitehall II
Cohort Study). American Journal of Cardiology, 100(3), 524527. http://dx.doi.
org/10.1016/j.amjcard.2007.03.056
*Burg, J. M., Wolf, O. T., & Michalak, J. (2012). Mindfulness as self-regulated
attention associations with heart rate variability. Swiss Journal of Psychology,
71(3), 135139. http://dx.doi.org/10.1024/1421-0185/a000080
*Cellini, N., Covassin, N., de Zambotti, M., Sarlo, M., & Stegagno, L. (2013).
Relationship between cardiovascular resting state and visual attention. Clinical
Autonomic Research, 23(3), 157161. http://dx.doi.org/10.1007/s10286-0130194-x
*Cocia, I. R., Uscatescu, L. C., & Rusu, A. S. (2012). Attention bias to threat in
anxiety-prone individuals evidence from disengagement, but not engagement
bias using cardiac vagal tone. Journal of Psychophysiology, 26(2), 7482. http://
dx.doi.org/10.1027/0269-8803/a000069
Cuijpers, P., van Straten, A., Bohlmeijer, E., Hollon, S. D., & Andersson, G. (2010).
The effects of psychotherapy for adult depression are overestimated: a
meta-analysis of study quality and effect size. Psychological Medicine, 40(2),
211223. http://dx.doi.org/10.1017/S0033291709006114
De Meersman, R. E., & Stein, P. K. (2007). Vagal modulation and aging. Biological
Psychology, 74(2), 165173. http://dx.doi.org/10.1016/j.biopsycho.2006.04.008
de Ridder, D. T., Lensvelt-Mulders, G., Finkenauer, C., Stok, F. M., & Baumeister, R. F.
(2012). Taking stock of self-control: a meta-analysis of how trait self-control
relates to a wide range of behaviors. Personality and Social Psychology Review,
16(1), 7699. http://dx.doi.org/10.1177/1088868311418749
*Demaree, H. A., Robinson, J. L., Everhart, D. E., & Schmeichel, B. J. (2004). Resting
RSA is associated with natural and self-regulated responses to negative
emotional stimuli. Brain and Cognition, 56(1), 1423. http://dx.doi.org/10.1016/
j.bandc.2004.05.001
Duckworth, A. L., & Kern, M. L. (2011). A meta-analysis of the convergent validity of
self-control measures. Journal of Research in Personality, 45(3), 259268. http://
dx.doi.org/10.1016/j.jrp.2011.02.004
*Duschek, S., Muckenthaler, M., Werner, N., & del Paso, G. A. R. (2009).
Relationships between features of autonomic cardiovascular control and
cognitive performance. Biological Psychology, 81(2), 110117. http://dx.doi.org/
10.1016/j.biopsycho.2009.03.003
Duval, S., & Tweedie, R. (2000). Trim and ll: a simple funnel-plot-based method of
testing and adjusting for publication bias in meta-analysis. Biometrics, 56(2),
455463. http://dx.doi.org/10.1111/j.0006-341X.2000.00455.x
Fisher, Z., Tipton, E., (2014). Robumeta: robust variance meta-regression (Version R
package version 1.3.), retrieved from http://CRAN.R-project.org/
package=robumeta.
Frazier, T. W., Strauss, M. E., & Steinhauer, S. R. (2004). Respiratory sinus
arrhythmia as an index of emotional response in young adults.
Psychophysiology, 41(1), 7583. http://dx.doi.org/10.1046/j.1469-8986.2003.
00131.x
Friederich, H. C., Schild, S., Schellberg, D., Quenter, A., Bode, C., Herzog, W., & Zipfel,
S. (2006). Cardiac parasympathetic regulation in obese women with binge
eating disorder. International Journal of Obesity, 30(3), 534542. http://dx.doi.
org/10.1038/sj.ijo.0803181
Fukusaki, C., Kawakubo, K., & Yamamoto, Y. (2000). Assessment of the primary
effect of aging on heart rate variability in humans. Clinical Autonomic Research,
10(3), 123130. http://dx.doi.org/10.1007/bf02278016
*Gaebler, M., Daniels, J. K., Lamke, J. P., Fydrich, T., & Walter, H. (2013). Heart rate
variability and its neural correlates during emotional face processing in social
anxiety disorder. Biological Psychology, 94(2), 319330. http://dx.doi.org/10.
1016/j.biopsycho.2013.06.009
*Geisler, F. C. M., & Kubiak, T. (2009). Heart rate variability predicts self-control in
goal pursuit. European Journal of Personality, 23(8), 623633. http://dx.doi.org/
10.1002/per.727
*Gillie, B. L., Vasey, M. W., & Thayer, J. F. (2013). Heart rate variability predicts
control over memory retrieval. Psychological Science, http://dx.doi.org/10.1177/
0956797613508789
Go, A. S., Mozaffarian, D., Roger, V. L., Benjamin, E. J., Berry, J. D., Blaha, M. J., . . .
Turner, M. B., (2014). Heart disease and stroke statistics2014 update: a report
25
from the American Heart Association. Circulation, 129(3), e28e292.

doi:10.1161/01cir.0000441139.02102.80.
Goldstein, D. S., Bentho, O., Park, M. Y., & Sharabi, Y. (2011). Low-frequency power
of heart rate variability is not a measure of cardiac sympathetic tone but may
be a measure of modulation of cardiac autonomic outows by baroreexes.
Experimental Physiology, 96(12), 12551261. http://dx.doi.org/10.1113/
expphysiol.2010.056259
Graziano, P., & Derenko, K. (2013). Cardiac vagal control and childrens adaptive
functioning: a meta-analysis. Biological Psychology, 94(1), 2237. http://dx.doi.
org/10.1016/j.biopsycho.2013.04.011
Grossman, P. (1992). Respiratory and cardiac rhythms as windows to central and
autonomic biobehavioral regulation: selection of window frames, keeping the
panes clean and viewing the neural topography. Biological Psychology, 34(23),
131161. http://dx.doi.org/10.1016/0301-0511(92)90013-k
Grossman, P., van Beek, J., & Wientjes, C. (1990). A comparison of three
quantication methods for estimation of respiratory sinus arrhythmia.
Psychophysiology, 27(6), 702714. http://dx.doi.org/10.1111/j.1469-8986.1990.
tb03198.x
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. (2010). Ego depletion and
the strength model of self-control: a meta-analysis. Psychological Bulletin,
136(4), 495525. http://dx.doi.org/10.1037/a0019486
*Hansen, A. L., Johnsen, B. H., & Thayer, J. E. (2003). Vagal inuence on working
memory and attention. International Journal of Psychophysiology, 48(3),
263274. http://dx.doi.org/10.1016/s0167-8760(03)00073-4
*Hansen, A. L., Johnsen, B. H., & Thayer, J. F. (2009). Relationship between heart rate
variability and cognitive function during threat of shock. Anxiety, Stress, and
Coping, 22(1), 7789. http://dx.doi.org/10.1080/10615800802272251
Hedges, A. J., Tipton, E., & Johnson, M. C. (2010). Robust variance estimation in
meta-regression with dependent effect size estimates. Research Synthesis
Methods, 1, 3965. http://dx.doi.org/10.1002/jrsm.5
Hofmann, W., Schmeichel, B. J., & Baddeley, A. D. (2012). Executive functions and
self-regulation. Trends in Cognitive Science, 16(3), 174180. http://dx.doi.org/
10.1016/j.tics.2012.01.006
Ingjaldsson, J. T., Laberg, J. C., & Thayer, J. F. (2003). Reduced heart rate variability in
chronic alcohol abuse: relationship with negative mood, chronic thought
suppression, and compulsive drinking. Biological Psychiatry, 54(12),
14271436. http://dx.doi.org/10.1016/s0006-3223(02)01926-1
Katona, P. G., & Jih, F. (1975). Respiratory sinus arrhythmia: noninvasive measure of
parasympathetic cardiac control. Journal of Applied Physiology, 39(5), 801805.
*Kaufmann, T., Vgele, C., Stterlin, S., Lukito, S., & Kbler, A. (2012). Effects of
resting heart rate variability on performance in the P300 brain-computer
interface. International Journal of Psychophysiology, 83(3), 336341. http://dx.
doi.org/10.1016/j.ijpsycho.2011.11.018
*Kimhy, D., Crowley, O. V., McKinley, P. S., Burg, M. M., Lachman, M. E., Tun, P. A.,
. . . Sloan, R. P., (2013). The association of cardiac vagal control and executive
functioningndings from the MIDUS study. Journal of Psychiatric Research,
47(5), 628635. doi:10.1016/j.jpsychires.2013.01.018.
*Krypotos, A. M., Jahfari, S., van Ast, V. A., Kindt, M., & Forstmann, B. U. (2011).
Individual differences in heart rate variability predict the degree of slowing during
response inhibition and initiation in the presence of emotional stimuli, 2(278)
http://dx.doi.org/10.3389/fpsyg.2011.00278
*Laborde, S., & Raab, M. (2013). The tale of hearts and reason: the inuence of mood
on decision making. Journal of Sport & Exercise Psychology, 35(4), 339357.
Lewis, G. F., Furman, S. A., McCool, M. F., & Porges, S. W. (2012). Statistical
strategies to quantify respiratory sinus arrhythmia: are commonly used
metrics equivalent? Biological Psychology, 89(2), 349364. http://dx.doi.org/10.
1016/j.biopsycho.2011.11.009
Lipsey, M. W., & Wilson, D. B. (2001). The way in which intervention studies have
personality and why it is important to meta-analysis. Evaluation and the Health
Professions, 24(3), 236254. http://dx.doi.org/10.1177/016327870102400302
*Luft, C. D., Takas, E., & Darby, D. (2009). Heart rate variability and cognitive
function: effects of physical effort. Biological Psychology, 82(2), 196201. http://
dx.doi.org/10.1016/j.biopsycho.2009.07.007
Lumley, T., (2012). rmeta: Meta-analysis, retrieved from http://CRAN.R-project.
org/package=rmeta.
*Meule, A., Vgele, C., & Kbler, A. (2012). Restrained eating is related to
accelerated reaction to high caloric foods and cardiac autonomic dysregulation.
Appetite, 58(2), 638644. http://dx.doi.org/10.1016/j.appet.2011.11.023
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T.
D. (2000). The unity and diversity of executive functions and their
contributions to complex Frontal Lobe tasks: a latent variable analysis.
Cognitive Psychology, 41(1), 49100. http://dx.doi.org/10.1006/cogp.1999.0734
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of limited
resources: does self-control resemble a muscle? Psychological Bulletins, 126(2),
247259. http://dx.doi.org/10.1037/0033-2909.126.2.247
Muraven, M., & Slessareva, E. (2003). Mechanisms of self-control failure:
motivation and limited resources. Personality & social psychology bulletin,
29(7), 894906. http://dx.doi.org/10.1177/0146167203029007008
Nunan, D., Donovan, G., Jakovljevic, D. G., Hodges, L. D., Sandercock, G. R., & Brodie,
D. A. (2009). Validity and reliability of short-term heart-rate variability from
the Polar S810. Medicine and Science in Sports and Exercise, 41(1), 243250.
http://dx.doi.org/10.1249/mss.0b013e318184a4b1
Nunan, D., Jakovljevic, D. G., Donovan, G., Hodges, L. D., Sandercock, G. R., & Brodie,
D. A. (2008). Levels of agreement for RR intervals and short-term heart rate
26
variability obtained from the Polar S810 and an alternative system. European
Journal of Applied Physiology, 103(5), 529537. http://dx.doi.org/10.1007/
s00421-008-0742-6
Nunan, D., Sandercock, G. R., & Brodie, D. A. (2010). A quantitative systematic
review of normal values for short-term heart rate variability in healthy adults.
Pacing and Clinical Electrophysiology, 33(11), 14071417. http://dx.doi.org/10.
1111/j.1540-8159.2010.02841.x
*Ohira, H., Matsunaga, M., Osumi, T., Fukuyama, S., Shinoda, J., Yamada, J., & Gidron,
Y. (2013). Vagal nerve activity as a moderator of brain-immune relationships.
Journal of Neuroimmunology, 260(12), 2836. http://dx.doi.org/10.1016/j.
jneuroim.2013.04.011
*Park, G., Van Bavel, J. J., Vasey, M. W., & Thayer, J. F. (2012). Cardiac vagal tone
predicts inhibited attention to fearful faces. Emotion, 12(6), 12921302. http://
dx.doi.org/10.1037/a0028528
*Park, G., Van Bavel, J. J., Vasey, M. W., & Thayer, J. F. (2013). Cardiac vagal tone
predicts attentional engagement to and disengagement from fearful faces.
Emotion, 13(4), 645656. http://dx.doi.org/10.1037/a0032971
*Park, G., Vasey, M. W., Van Bavel, J. J., & Thayer, J. F. (2013). Cardiac vagal tone is
correlated with selective attention to neutral distractors under load.
Psychophysiology, 50(4), 398406. http://dx.doi.org/10.1111/psyp.12029
Penttila, J., Helminen, A., Jartti, T., Kuusela, T., Huikuri, H. V., Tulppo, M. P., . . .
Scheinin, H., (2001). Time domain, geometrical and frequency domain analysis
of cardiac vagal outow: effects of various respiratory patterns. Clinical
Physiology, 21(3), 365376. doi:10.1046/j1365-2281.2001.00337.x.
Pigott, T. D. (2012). Statistics for social and behavioral sciences: advances in
meta-analysis. Boston, MA, USA: Springer.
Porges, S. W. (2001). The polyvagal theory: phylogenetic substrates of a social
nervous system. International Journal of Psychophysiology, 42(2), 123146.
http://dx.doi.org/10.1016/s0167-8760(01)00162-3
Porges, S. W. (2007). The polyvagal perspective. Biological Psychology, 74(2),
116143. http://dx.doi.org/10.1016/j.biopsycho.2006.06.009
Porges, S. W., & Bohrer, R. E. (1990). Analyses of periodic processes in
psychophysiological research. In J. T. Cacioppo, & L. G. Tassinary (Eds.),
Principles of psychophysiology: physical, social, and inferential elements (pp.
708753). New York: Cambridge University Press.
*Pu, J., Schmeichel, B. J., & Demaree, H. A. (2010). Cardiac vagal control predicts
spontaneous regulation of negative emotional expression and subsequent
cognitive performance. Biological Psychology, 84(3), 531540. http://dx.doi.org/
10.1016/j.biopsycho.2009.07.006
Quintana, D. S., Heathers, J. A., & Kemp, A. H. (2012). On the validity of using the
Polar RS800 heart rate monitor for heart rate variability research. European
s00421-012-2453-2
Quintana, D. S., McGregor, I. S., Guastella, A. J., Malhi, G. S., & Kemp, A. H. (2013). A
meta-analysis on the impact of alcohol dependence on short-term
resting-state heart rate variability: implications for cardiovascular risk.
Alcoholism: Clinical and Experimental Research, 37(Suppl. 1), E23E29. http://dx.
doi.org/10.1111/j.1530-0277.2012.01913.x
R Core Team. (2014). R: a language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing. http://www.R-project.
org/
*Reynard, A., Gevirtz, R., Berlow, R., Brown, M., & Boutelle, K. (2011). Heart rate
variability as a marker of self-regulation. Applied Psychophysiology and
Biofeedback, 36(3), 209215. http://dx.doi.org/10.1007/s10484-011-9162-1
Rosenbaum, S., Tiedemann, A., Sherrington, C., Curtis, J., & Ward, P. B. (2014).
Physical activity interventions for people with mental illness: a systematic
review and meta-analysis. Journal of Clinical Psychiatry, 75(9), 964974. http://
dx.doi.org/10.4088/jcp.13r08765
Rosenthal, R. (1995). Writing metaanalytic reviews. Psychological Bulletin, 118(2),
183192. http://dx.doi.org/10.1037//0033-2909.118.2.183
Rottenberg, J. (2007). Cardiac vagal control in depression: a critical analysis.
Biological Psychology, 74(2), 200211. http://dx.doi.org/10.1016/j.biopsycho.
2005.08.010
Schafer, A., & Vagedes, J. (2013). How accurate is pulse rate variability as an
estimate of heart rate variability? A review on studies comparing
photoplethysmographic technology with an electrocardiogram. International
Journal of Cardiology, 166(1), 1529. http://dx.doi.org/10.1016/j.ijcard.2012.03.
119
*Segerstrom, S. C., & Solberg Nes, L. (2007). Heart rate variability reects
self-regulatory strength, effort, and fatigue. Psychological Science, 18(3),
275281. http://dx.doi.org/10.1111/j.1467-9280.2007.01888.x
Sosnowski, M. (2010). Heart rate variability. In P. W. Macfarlane, A. van Oosterom,

O. Pahlm, P. Kligeld, M. Janse, & J. Camm (Eds.), Comprehensive
electrocardiology (pp. 15131675). Springer.
Stein, P. K., Kleiger, R. E., & Rottman, J. N. (1997). Differing effects of age on heart
rate variability in men and women. American Journal of Cardiology, 80(3),
302305. http://dx.doi.org/10.1016/s0002-9149(97)00350-0
Sterne, J. A. C., & Egger, M. (2005). Regression methods to detect publication and
other bias in meta-analysis. In H. R. Rothstein, A. J. Sutton, & M. Borenstein
(Eds.), Publication bias in meta-analysis: prevention, assessment, and adjustments
(pp. 99110). Chichester, England: Wiley.
Suchy, Y. (2009). Executive functioning: overview, assessment, and research issues
for non-neuropsychologists. Annals of Behavioral Medicine, 37(2), 106116.
http://dx.doi.org/10.1007/s12160-009-9097-4
Tak, L. M., Riese, H., de Bock, G. H., Manoharan, A., Kok, I. C., & Rosmalen, J. G.
(2009). As good as it gets? A meta-analysis and systematic review of
methodological quality of heart rate variability studies in functional somatic
disorders. Biological Psychology, 82(2), 101110. http://dx.doi.org/10.1016/j.
biopsycho.2009.05.002
Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts
good adjustment, less pathology, better grades, and interpersonal success.
Journal of Personality, 72(2), 271324. http://dx.doi.org/10.1111/j.0022-3506.
2004.00263.x
Tanner-Smith, E. E., & Tipton, E. (2013). Robust variance estimation with
dependent effect sizes: practical considerations including a software tutorial in
Stata and SPSS. Research Synthesis Methods, 5, 1330. http://dx.doi.org/10.
1002/jrsm.1091
Task Force of the European Society of Cardiology and the North American Society
of Pacing and Electrophysiology, (1996). Heart rate variability: standards of
measurement, physiological interpretation and clinical use. Task Force of the
European Society of Cardiology and the North American Society of Pacing and
Electrophysiology. Circulation, 93(5), 1043-1065. doi:10.1161/01
CIR.93.5.1043.
Thayer, J. F., Hansen, A. L., & Johnsen, B. H. (2010). The non-invasive assessment of
autonomic inuences on the heart using impedance cardiography and heart
rate variability. In A. Steptoe (Ed.), Handbook of behavioral medicine (pp.
723740). Springer.
Thayer, J. F., Hansen, A. L., Saus-Rose, E., & Johnsen, B. H. (2009). Heart rate
variability, prefrontal neural function, and cognitive performance: the
neurovisceral integration perspective on self-regulation, adaptation, and
health. Annals of Behavioral Medicine, 37(2), 141153. http://dx.doi.org/10.
1007/s12160-009-9101-z
Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion
regulation and dysregulation. Journal of Affective Disorders, 61(3), 201216.
http://dx.doi.org/10.1016/s0165-0327(00)00338-4
Thayer, J. F., Loerbroks, A., & Sternberg, E. M. (2011). Inammation and
cardiorespiratory control: the role of the vagus nerve. Respiratory Physiology
and Neurobiology, 178(3), 387394. http://dx.doi.org/10.1016/j.resp.2011.05.
016
Tipton, E. (2014). Small sample adjustments for robust variance estimation with
meta-regression. Psychological Methods, http://dx.doi.org/10.1037/
met0000011
Tzeng, Y. C., Larsen, P. D., & Galletly, D. C. (2003). Cardioventilatory coupling in
resting human subjects. Experimental Physiology, 88(6), 775782. http://dx.doi.
org/10.1113/eph8802606
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package.
Journal of Statistical Software, 36, 148. http://www.jstatsoft.org/v36/i03/
Viechtbauer, W., & Cheung, M. W.-L. (2010). Outlier and inuence diagnostics for
meta-analysis. Research Synthesis Methods, 1, 112125. http://dx.doi.org/10.
1002/jrsm.11
*Volokhov, R. N., & Demaree, H. A. (2010). Spontaneous emotion regulation to
positive and negative stimuli. Brain and Cognition, 73(1), 16. http://dx.doi.org/
10.1016/j.bandc.2009.10.015
Wallen, M. B., Hasson, D., Theorell, T., Canlon, B., & Osika, W. (2012). Possibilities
and limitations of the Polar RS800 in measuring heart rate variability at rest.
European Journal of Applied Physiology, 112(3), 11531165. http://dx.doi.org/10.
1007/s00421-011-2079-9
Weippert, M., Kumar, M., Kreuzfeld, S., Arndt, D., Rieger, A., & Stoll, R. (2010).
Comparison of three mobile devices for measuring RR intervals and heart rate
variability: Polar S810i, Suunto t6 and an ambulatory ECG system. European
s00421-010-1415-9

Zahn 2016. HRV and Self Control

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Zahn 2016. HRV and Self Control

Transféré par

Droits d'auteur :

Formats disponibles

Biological Psychology 115 (2016) 926

Contents lists available at ScienceDirect

Heart rate variability and self-controlA meta-analysis

D. Zahn et al. / Biological Psychology 115 (2016) 926

arousal (Porges, 2007). Respiratory sinus arrhythmia (RSA), which

D. Zahn et al. / Biological Psychology 115 (2016) 926

control (de Ridder et al., 2012), we hypothesize that the relationship

D. Zahn et al. / Biological Psychology 115 (2016) 926

In addition to time and frequency domain measures of trait

2001) or the RSA estimated with the PorgesBohrer-method seem

D. Zahn et al. / Biological Psychology 115 (2016) 926

ing self-control success. The focus of the quantitative integration

All records were screened by two independent raters (JA, DZ)

D. Zahn et al. / Biological Psychology 115 (2016) 926

Records identified through database searching

Records screened based on title

Records screened based on abstract

Full-text articles assessed for eligibility

Full-text articles excluded (n = 303)

Publications included in quantitative synthesis

Fig. 1. Flow chart of study selection.

The category Cognitive control included task measures where

2000). Tasks assessing reappraisal after mood induction were also

D. Zahn et al. / Biological Psychology 115 (2016) 926

2.3.2.2. Control of respiration. Control of respiration was rated as

2.3.3. Study quality

D. Zahn et al. / Biological Psychology 115 (2016) 926

session) was chosen as we assumed that a time lag between the

2009; Pigott, 2012). As multiple effects of a study often resulted

Burg, Wolf, and

Mindful breathing exercise

Exaggeration of facial expression

Hansen et al. (2003)

Rumination after negative

Respiration adjusted effects

Krypotos et al. (2011)

Luft et al. (2009)

Cogstate Computerized System

Stochastic reversal learning task

Park, Van Bavel et al.

Emotional spatial cueing task

Park, Vasey, Van Bavel,

Letter detection task

Reynard et al. (2011)

Working memory test after mood

respiration adjusted effects

Emotion regulation after mood

D. Zahn et al. / Biological Psychology 115 (2016) 926

D. Zahn et al. / Biological Psychology 115 (2016) 926

self-control domains against the grand mean of all three categories.

3.2.1.2. Executive functions. Of the 132 extracted effects, four

Effect size Lower CI Higher CI Weight

Effect size (Fisher's z metric)

Effect size Lower CI Higher CI Weight

Effect size (Fisher's z metric)

D. Zahn et al. / Biological Psychology 115 (2016) 926

Effect size Lower CI Higher CI Weight

Working Memory Test

Effect size (Fisher's z metric)

Effect size Lower CI Higher CI Weight

Average effect size

Effect size (Fisher's z metric)

recording devices did not show a signicant relationship between

D. Zahn et al. / Biological Psychology 115 (2016) 926