Effects of Circadian Cortisol on the Development of a Health Habit

Abstract

Objective:
Given the impact of individuals' habits on health, it is important to study how behaviors can become habitual. Cortisol has been well documented to have a role in habit formation. This study aimed to elucidate the influence of the circadian rhythm of cortisol on habit formation in a real-life setting.

Method:
Forty-eight students were followed for 90 days during which they attempted to adopt a health behavior (psoas iliac stretch). They were randomly assigned to perform the stretch either upon waking in the morning, when cortisol concentrations are high, or before evening bedtime, when cortisol levels approach the nadir. A smartphone application was used to assess the Self-Report Behavioural Automaticity Index every day and to provide reminders for salivary measurements every 30 days. The speed of the health habit formation process was calculated by modeling the learning curves.

Results:
Extrapolation of the curves indicated that the morning group achieved automaticity at an earlier time point (105.95 days) than did the evening group (154.01 days). In addition, the cortisol level during the performance of the health behavior was identified as a significant mediator of the time point when the health behavior became habitual.

Conclusion:
The present findings suggest that the time course of the development of healthy habits depends on the time of the day and that the effect is mediated through diurnal variation in cortisol levels. Future studies are now needed to determine to what extent cortisol rhythmicity can help individuals to adopt new health behaviors. (PsycINFO Database Record

Full text

Health Psychology
Effects of Circadian Cortisol on the Development of a
Health Habit
Marion Fournier, Fabienne d’Arripe-Longueville, Carole Rovere, Christopher S. Easthope, Lars
Schwabe, Jonathan El Methni, and Rémi Radel
Online First Publication, June 26, 2017. http://dx.doi.org/10.1037/hea0000510
CITATION
Fournier, M., d’Arripe-Longueville, F., Rovere, C., Easthope, C. S., Schwabe, L., El Methni, J., &
Radel, R. (2017, June 26). Effects of Circadian Cortisol on the Development of a Health Habit.
Health Psychology. Advance online publication. http://dx.doi.org/10.1037/hea0000510

BRIEF REPORT
Effects of Circadian Cortisol on the Development of a Health Habit
Marion Fournier and Fabienne d’Arripe-Longueville
Université Côte d’Azur
Carole Rovere
Université de Nice Sophia Antipolis
Christopher S. Easthope
Balgrist University Hospital
Lars Schwabe
University of Hamburg
Jonathan El Methni
Université Paris Descartes
Rémi Radel
Université Côte d’Azur
Objective: Given the impact of individuals’ habits on health, it is important to study how behaviors can
become habitual. Cortisol has been well documented to have a role in habit formation. This study aimed
to elucidate the influence of the circadian rhythm of cortisol on habit formation in a real-life setting.
Method: Forty-eight students were followed for 90 days during which they attempted to adopt a health
behavior (psoas iliac stretch). They were randomly assigned to perform the stretch either upon waking
in the morning, when cortisol concentrations are high, or before evening bedtime, when cortisol levels
approach the nadir. A smartphone application was used to assess the Self-Report Behavioural Automa-
ticity Index every day and to provide reminders for salivary measurements every 30 days. The speed of
the health habit formation process was calculated by modeling the learning curves. Results: Extrapolation
of the curves indicated that the morning group achieved automaticity at an earlier time point (105.95
days) than did the evening group (154.01 days). In addition, the cortisol level during the performance of
the health behavior was identified as a significant mediator of the time point when the health behavior
became habitual. Conclusion: The present findings suggest that the time course of the development of
healthy habits depends on the time of the day and that the effect is mediated through diurnal variation
in cortisol levels. Future studies are now needed to determine to what extent cortisol rhythmicity can help
individuals to adopt new health behaviors.
Keywords: health, habits, learning, stress, cortisol
Supplemental materials: http://dx.doi.org/10.1037/hea0000510.supp
Attention should be paid to behavioral habits, given their important
role in health (Wood & Neal, 2016). A behavioral habit is a process
by which a stimulus automatically generates an impulse toward ac-
tion, based on learned stimulus–response associations (Gardner,
2015). Recent studies have pointed to the critical role of stress and the
stress hormones (mainly cortisol) in the development of habit behav-
ior. For instance, it has been shown that stress induces a shift in the
control of instrumental behavior from goal-directed toward habitual
responding (e.g., Schwabe, Schächinger, de Kloet, & Oitzl, 2010).
The endogenous level of cortisol varies according to a circadian
rhythm. In humans, cortisol levels are low at midnight and increase
overnight to a peak in the morning. Following this morning peak,
cortisol levels slowly decline throughout the day (Weitzman et al.,
1971). Although the impact of cortisol on habit behavior has been
well established through stress manipulation or cortisol injection (e.g.,
Fournier, d’Arripe-Longueville, & Radel, 2017; Quirarte et al., 2009),
the influence of cortisol’s circadian rhythm on habit formation re-
mains to be elucidated.
Many studies have demonstrated the predictive capacity of
habits on health behavior (e.g., de Bruijn & Rhodes, 2011), but
less is known about the process of habit formation in real-life
settings. A previous study followed participants over 84 days
while they adopted a new, daily health-promoting behavior such
as eating fruit or exercising (Lally, van Jaarsveld, Potts, &
Marion Fournier and Fabienne d’Arripe-Longueville, Faculty of Sport
Science, Université Côte d’Azur; Carole Rovere, Institut de Pharmacologie
Moléculaire et Cellulaire, Université de Nice Sophia Antipolis; Christopher
S. Easthope, Spinal Cord Injury Center, Balgrist University Hospital; Lars
Schwabe, Department of Cognitive Psychology, Institute for Psychology,
University of Hamburg; Jonathan El Methni, Department of Statistics,
Université Paris Descartes; Rémi Radel, Faculty of Sport Science, Univer-
sité Côte d’Azur.
Correspondence concerning this article should be addressed to Marion
Fournier, Faculty of Sport Science, University Côte d’Azur, LAMHESS,
Nice, France. E-mail: fou2marion@gmail.com
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Health Psychology © 2017 American Psychological Association
2017, Vol. 999, No. 999, 000 0278-6133/17/$12.00 http://dx.doi.org/10.1037/hea0000510
1

Wardle, 2010). Although this protocol was innovative in track-
ing the formation of habits by collecting daily reports of expe-
rienced automaticity, the collected data proved challenging to
model. In the present study, a similar protocol was used to
investigate the influence of diurnal variations in cortisol con-
centrations on the dynamics of habit formation using a real-
world application. To better fit the evolution of experienced
automaticity in time, we used logistic curves (Murre, 2014).
The level of circadian cortisol was manipulated by having two
groups of participants perform a new health behavior at differ-
ent times of the day. Given that cortisol levels are higher in the
morning than in the evening, we hypothesized that the new
behavior would become habitual more quickly if executed in
the morning than in the evening. The cortisol level was ex-
pected to mediate this effect.
Method
Participants
On average, diurnal variations of cortisol can lead to a 20%
difference in learning and memory (May, Hasher, & Stoltzfus,
1993; Petros, Beckwith, & Anderson, 1990; Wyatt, Ritz-De
Cecco, Czeisler, & Dijk, 1999), which suggests that at least 46
participants were needed to observe such an effect with a power
of .80. Forty-eight French students (28 male; age ⫽21.7 ⫾1.78
years, range ⫽20 –25) participated in exchange for course
credit. Participation was limited to healthy nonsmokers with
normal body mass index (22.4 ⫾2.17 kg/m
2
, range ⫽18.2–
27.7 kg/m
2
) not under contraceptive medication according to
the recommendations for salivary cortisol measurements
(Kirschbaum, Kudielka, Gaab, Schommer, & Hellhammer,
1999). The study was approved by the local ethics committee
for the protection of individuals (Université de Nice Sophia-
Antipolis) and conducted in accordance with the Declaration of
Helsinki (1964) ethical guidelines.
Procedure
Participants were invited to a first meeting with the experi-
menter. After providing written informed consent, they were
randomly assigned to a morning or evening group according to
a minimization algorithm, balancing for gender. Here, a chro-
notype questionnaire and a measure of intention to adopt the
behavior was completed. The intervention commenced within a
week of the first meeting. It consisted of performing a new
behavior once daily for 90 days. The behavior was a stretching
exercise that is highly recommended to maintain flexibility and
prevent low back pain (i.e., psoas iliac stretch; see the online
supplemental materials). Depending on group allocation, the
stretch was completed in bed upon waking or before sleeping.
Adherence and a visual analog scale version of the Self-Report
Behavioural Automaticity Index (SRBAI; Gardner, Abraham,
Lally, & de Bruijn, 2012) were recorded via a smartphone
reminder application on a daily basis. A salivary sample was
collected every 30 days. Details on the measures are provided in
the online supplemental materials.
Data Analysis
To evaluate the evolution of automaticity, we fitted a four-
parameter logistic function to participants’ daily responses to the
SRBAI (see the online supplemental materials). Although habit
formation has previously been modeled using a power function
(Lally et al., 2010), this method led to mixed results, providing a
moderate fit (R
2
⬎.70) for only 48% of the participants. Because
a logistic function can outperform a power function in modeling
learning curves (Murre, 2014), this approach was adopted. We
considered that, in line with Lally et al. (2010), automaticity was
achieved at the time point at which 95% of the asymptote was
reached (x
.95
). This value served as the dependent variable.
To determine the group effect and the potential mediation effect
of cortisol on x
.95
, we employed a mediation model using the
PROCESS toolbox (Hayes, 2012). Group (morning vs. evening)
was applied as the independent variable, cortisol concentration as
a mediator, and sex and intention as covariates for the prediction
of the dependent variable. A bias-corrected bootstrapping method
with 2,000 samples was used to evaluate the effects. The total
effect of the independent variable and its indirect effect through
the mediator are presented. To control for the chronotype, we used
a moderation analysis to examine how the effect of group on x
.95
was influenced by the chronotype score. The mediation model was
also used in a further configuration with the independent variable
replaced by a Group ⫻Chronotype interaction term. The interac-
tion term adequately represented the hypothesis of a match be-
tween the moment of execution and the participants’ chronotype
score (Hasher, Goldstein, & May, 2005).
Results
Of the 48 recruited participants, 42 completed the experiment:
19 participants in the morning group and 23 in the evening group.
The intention to adopt the behavior was high (4.74 ⫾0.39), with
no difference between groups (p⫽.24). The response rate to the
daily questionnaire was high (89.2%), with no difference between
groups (p⬎.25). Adherence was high (94.9%), with no difference
between groups (p⬎.25). The curve-fitting process was highly
successful. Fitting converged for each participant using the logistic
function (by comparison, only 31 participants could be fitted using
a power curve). A high adjustment quality index was obtained
(R
2
⫽.945 ⫾.053). Figure 1 presents the mean traces of the
morning and evening groups obtained by averaging each of the
four function parameters for each condition.
Automaticity (x
.95
) was slightly positively skewed but did not
significantly deviate from normality (p⫽.10). The results of the
first mediation model populated with group as the independent
variable are presented in Table 1. After controlling for covariates
(sex and intention), we observed a significant total effect of time
of day until habit development (effect ⫽⫺26.74, SE ⫽10.25,
95% confidence interval [CI: ⫺47.52, ⫺5,96]), indicating that the
behavior became habitual more quickly when it was performed in
the morning (M⫽105.95 ⫾46.72 days) as opposed to the evening
(M⫽154.01 ⫾71.05 days). Group (morning vs. evening) had a
significant effect on the cortisol level, indicating that cortisol
levels differed between groups (morning: 2.16 ⫾.87 ng/ml; eve-
ning: 1.10 ⫾.86 ng/ml). When included in the model predicting
the time to form a habit (x
.95
), cortisol was a significant predictor
and the group effect was no longer significant, suggesting a me-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2FOURNIER ET AL.

diating role. Accordingly, the group effect had no significant direct
effect (effect ⫽⫺14.21, SE ⫽11.21, 95% CI [⫺36.93, 8.50]), but
a significant indirect effect on the dependent variable through the
effect of the cortisol (effect ⫽⫺11.89, SE ⫽6.25, 95% CI
[⫺30.39, ⫺2.98]). Figure 2 illustrates the nature of the relation
between cortisol levels and x
.95
.
The second model, in which the independent variable group was
replaced by the Group ⫻Chronotype interaction, led to a similar
pattern (see Table 2). The Group ⫻Chronotype interaction had a
significant total effect on x
.95
(effect ⫽⫺.57, SE ⫽.22, 95% CI
[⫺1.02, ⫺.12]). The total effect was decomposed into a nonsig-
nificant direct effect (effect ⫽⫺.29, SE ⫽.25, 95% CI [⫺.79,
2.13]) and a significant indirect effect through cortisol (ef-
fect ⫽⫺.27, SE ⫽.15, 95% CI [⫺.65, ⫺.04]).
Discussion
Because it has been shown that cortisol level has a strong
connection to habit memory formation and that cortisol concen-
trations vary markedly over the day, this study aimed to test the
influence of the circadian rhythm of cortisol on the development of
a healthy habit. The presented findings indicate that the stretching
behavior that participants intended to adopt became habitual more
quickly when it was performed in the morning than in the evening.
In addition, it seems that cortisol played a mediating role, because
there was a significant indirect effect of the time of day on
automatization speed through the level of cortisol displayed at the
moment of behavior execution. Cortisol concentrations in the
morning were higher than in the evening samples, and the cortisol
level was negatively associated with the time taken to make the
behavior habitual (see Figure 2). To our knowledge, this is the first
time that habit formation has been studied through the prism of
chronobiology. Our findings are consistent with those in other
studies that have manipulated glucocorticoides GC levels with
either pharmacological injection or stress induction, because they
have all indicated that cortisol level has an impact on habit for-
mation (e.g., Quirarte et al., 2009; Schwabe & Wolf, 2009).
Our results show that the indirect effect of time of day on habit
formation through the circulating cortisol level persisted, and was
even slightly stronger, when we took into account the match
between chronotype and the time of behavior execution. This
match effect contributes to the large body of literature on the
effects of chronobiology on behavior (e.g., Hasher et al., 2005)
showing that synchrony between individual preferences and the
time of the day has an important effect on performance and
particularly on tasks relying on attentional demands such as learn-
ing. It is thus not surprising that this individual trait further refines
Figure 1. Fitted logistic function representating the evolution of be-
havioral automaticity (Self-Reported Behavioural Automaticity Index,
SRBAI) in time in the morning and evening conditions. Shaded areas
represent standard error of the mean. Data from after the measurement
period (90 days) is extrapolated and indicated using light gray. The
moment when participants reach 95% of the maximal asymptote (dotted
line, x
.95
) represents the time taken to form the habit. Traces are
normalized for representation purposes.
Table 1
Regression Models Used to Determine the Mediating Role of Cortisol in the Effect of the
Condition on the Time Taken to Form a Behavioral Habit
Model and variable Coefficient SE t p
95% CI
LL UL
Model predicting cortisol: R
2
⫽.28, F(1, 40) ⫽14.89, p⫽.0004
Constant 1.627 .138 11.831 .000 1.349 1.905
Condition .531 .138 3.859 .001 .252 .809
Model predicting x
.95
without inclusion of the mediator: R
2
⫽.20, F(3, 38) ⫽2.34, p⫽.09
Constant 288.495 172.121 1.671 .102 ⫺59.950 636.941
Condition ⫺26.601 10.169 ⫺2.616 .013 ⫺47.187 ⫺6.015
Intention ⫺30.528 35.453 ⫺.861 .395 ⫺102.299 41.244
Sex ⫺25.043 20.269 ⫺1.236 .224 ⫺66.075 15.990
Model predicting x
.95
with inclusion of the mediator: R
2
⫽.29, F(4, 37) ⫽2.58, p⫽.053
Constant 290.459 150.281 1.933 .061 ⫺14.044 594.962
Cortisol ⫺22.413 10.779 ⫺2.079 .045 ⫺44.254 ⫺.572
Condition ⫺14.218 11.213 ⫺1.268 .213 ⫺36.937 8.502
Intention ⫺23.578 31.366 ⫺.752 .457 ⫺87.133 39.976
Sex ⫺22.316 19.782 ⫺1.128 .267 ⫺62.398 17.767
Note.CI⫽confidence interval; LL ⫽lower limit; ULCI ⫽upper limit.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
CORTISOL AND HEALTHY HABIT DEVELOPMENT

the effect of the moment of execution on the acquisition of behav-
ioral habits.
Our study was designed to test habit formation in real life and
therefore makes an original contribution to the literature by having
greater ecologic validity than do the laboratory tasks (i.e., instru-
mental tasks) that are generally used to study the habit formation
process. The short SRBAI and smartphone reminder application
made it possible for us to track the participants for 3 months.
However, self-reported habit measures have been criticized (Hag-
ger, Rebar, Mullan, Lipp, & Chatzisarantis, 2015) because people
may have limited conscious knowledge of automatic behaviors.
Nevertheless, some researchers have still contested the belief that
people can provide valid information on their habits using such
self-reports (Orbell & Verplanken, 2015). In addition, there are no
current alternatives for measuring behavioral habits and automa-
ticity, particularly for frequent repeated measurements. To suc-
cessfully analyze the evolution of automaticity we used a new type
of learning curve that seems to better represent the habit-formation
process. Unlike power learning curves, which show a substantial
gain at the beginning, the logistic function suggests that some
Figure 2. Relation between salivary cortisol values and the time necessary for behavior automatization (x
.95
)
for the morning and evening groups.
Table 2
Regression Models Used to Determine the Mediating Role of Cortisol on the Effect of the
Product Representing the Condition ⫻Chronotype Interaction on the Time Taken to Form a
Behavioral Habit
Model and
variable Coefficient SE t p
95% CI
LL UL
Model predicting Cortisol: R
2
⫽.33, F(1, 40) ⫽20.699, p⫽.0001
Constant 1.634 .132 12.366 .000 1.367 1.901
Condition ⫻MEQ .012 .003 4.544 .000 .007 .018
Model predicting x
.95
without inclusion of the mediator: R
2
⫽.20, F(3, 38) ⫽2.18, p⫽.10
Constant 281.140 171.910 1.635 .110 ⫺66.879 629.159
Condition ⫻MEQ ⫺.566 .219 ⫺2.591 .013 ⫺1.009 ⫺.124
Intention ⫺28.987 35.409 ⫺.819 .418 ⫺100.671 42.696
Sex ⫺25.203 20.360 ⫺1.238 .223 ⫺66.419 16.014
Model predicting x
.95
with inclusion of the mediator: R
2
⫽.28, F(4, 37) ⫽2.51, p⫽.058
Constant 284.753 148.792 1.914 .063 ⫺16.733 586.239
Cortisol ⫺22.404 11.129 ⫺2.013 .051 ⫺44.955 .146
Condition ⫻MEQ ⫺.283 .248 ⫺1.141 .261 ⫺.787 .220
Intention ⫺22.369 31.232 ⫺.716 .478 ⫺85.652 40.914
Sex ⫺22.369 19.934 ⫺1.122 .269 ⫺62.760 18.023
Note.CI⫽confidence interval; LL ⫽lower limit; ULCI ⫽upper limit; MEQ ⫽Morningness-Eveningness
Questionnaire.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4FOURNIER ET AL.

participants may actually start more slowly (see Murre, 2014),
which is often the case for complex learning such as the adoption
of a new healthy behavior that requires plenty of conscious effort
at the beginning to accommodate the behavior in the living context
and evaluate its costs and benefits (Wood & Neal, 2016). Only
when these obstacles have been overcome may the automatization
process begin.
A limitation of our work concerns the relatively small sample
size of this study due to a limited recruitment capacity and attrition
throughout the duration of the intervention. Although the signifi-
cant results indicate that sufficient evidence was obtained to reject
the null hypothesis despite the limited power, this study should be
replicated to ensure reproducibility. In future studies, other medi-
ators should also be tracked. We could not demonstrate a full
mediation effect, and it therefore seems likely that other factors
may come into play. As Schwabe, Tegenthoff, Höffken, and Wolf
(2012) indicated, cortisol alone does not affect habitual behavior,
and concurrent glucocorticoid and noradrenergic activity may have
a role since it is known that this concurrent activity is needed to
shift learning from goal-directed to habitual control. Some psy-
chological factors might also explain the effect. For example, it is
possible that the behavior was perceived as less difficult, more
satisfying, or more easily cued in the morning than in the evening.
Moreover, people assigned to the morning group were given a
prior cue (i.e., do the stretch after waking up), but those in the
evening group were not (i.e., do the stretch before going to bed).
Because cueing is a critical aspect of habit formation (Gardner,
2015), it could have led to reinforcement of automatization in the
morning condition.
Conclusion
In this study, we demonstrated that a newly adopted stretching
behavior became habitual more quickly when it was performed in
the morning as opposed to the evening. This effect was mediated
by cortisol levels— higher cortisol levels in the morning resulted in
accelerated automatization. By matching the intervention time
with individuals’ chronobiology, faster automatization and there-
with higher success rates are likely. The extent to which these
results can be translated to applied settings and more complex
behaviors should be further investigated.
References
Courneya, K. S. (1994). Predicting repeated behavior from intention: The
issue of scale correspondence. Journal of Applied Social Psychology, 24,
580 –594. http://dx.doi.org/10.1111/j.1559-1816.1994.tb00601.x
de Bruijn, G. J., & Rhodes, R. E. (2011). Exploring exercise behavior,
intention and habit strength relationships. Scandinavian Journal of Med-
icine & Science in Sports, 21, 482– 491. http://dx.doi.org/10.1111/j
.1600-0838.2009.01064.x
Declaration of Helsinki. (1964, June.). Recommendation physicians in
biomedical research involving human subjects. 18th World Medical
Assembly, Helsinki, Finland.
Fournier, M., d’Arripe-Longueville, F., & Radel, R. (2017). Effects of
psychosocial stress on the goal-directed and habit memory systems
during learning and later execution. Psychoneuroendocrinology, 17,
275–283. http://dx.doi.org/10.1016/j.psyneuen.2016.12.008
Gardner, B. (2015). A review and analysis of the use of “habit” in
understanding, predicting and influencing health-related behaviour.
Health Psychology Review, 9, 277–295. http://dx.doi.org/10.1080/
17437199.2013.876238
Gardner, B., Abraham, C., Lally, P., & de Bruijn, G.-J. (2012). Towards
parsimony in habit measurement: Testing the convergent and predictive
validity of an automaticity subscale of the Self-Report Habit Index.
International Journal of Behavioral Nutrition and Physical Activity, 9,
102. http://dx.doi.org/10.1186/1479-5868-9-102
Hagger, M. S., Rebar, A. L., Mullan, B., Lipp, O. V., & Chatzisarantis,
N. L. D. (2015). The subjective experience of habit captured by self-
report indexes may lead to inaccuracies in the measurement of habitual
action. Health Psychology Review, 9, 296 –302. http://dx.doi.org/10
.1080/17437199.2014.959728
Hasher, L., Goldstein, D., & May, C. P. (2005). It’s about time: Circadian
rhythms, memory, and aging. Mahwah, NJ: Erlbaum.
Hayes, A. F. (2012). PROCESS: A versatile computational tool for ob-
served variable mediation, moderation, and conditional process model-
ing. Retrieved from http://www.afhayes.com/public/process2012.pdf
Horne, J. A., & Ostberg, O. (1976). A self-assessment questionnaire to
determine morningness-eveningness in human circadian rhythms. Inter-
national Journal of Chronobiology, 4, 97–110.
Kirschbaum, C., Kudielka, B. M., Gaab, J., Schommer, N. C., & Hellham-
mer, D. H. (1999). Impact of gender, menstrual cycle phase, and oral
contraceptives on the activity of the hypothalamus-pituitary-adrenal
axis. Psychosomatic Medicine, 61, 154 –162. http://dx.doi.org/10.1097/
00006842-199903000-00006
Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010).
How are habits formed: Modelling habit formation in the real world.
European Journal of Social Psychology, 40, 998 –1009. http://dx.doi
.org/10.1002/ejsp.674
MATLAB and Statistics Toolbox Release. (2012). The MathWorks, Inc.
Natick, MA.
May, C. P., Hasher, L., & Stoltzfus, E. R. (1993). Optimal time of day and
the magnitude of age differences in memory. Psychological Science, 4,
326 –330. http://dx.doi.org/10.1111/j.1467-9280.1993.tb00573.x
Murre, J. M. J. (2014). Learning and categorization in modular neural
networks. New York, NY: Psychology Press.
Orbell, S., & Verplanken, B. (2015). The strength of habit. Health Psy-
chology Review, 9, 311–317. http://dx.doi.org/10.1080/17437199.2014
.992031
Petros, T. V., Beckwith, B. E., & Anderson, M. (1990). Individual differ-
ences in the effects of time of day and passage difficulty on prose
memory in adults. British Journal of Psychology, 81, 63–72. http://dx
.doi.org/10.1111/j.2044-8295.1990.tb02346.x
Quirarte, G. L., Ledesma de la Teja, I. S., Casillas, M., Serafín, N.,
Prado-Alcalá, R. A., & Roozendaal, B. (2009). Corticosterone infused
into the dorsal striatum selectively enhances memory consolidation of
cued water-maze training. Learning & Memory, 16, 586 –589. http://dx
.doi.org/10.1101/lm.1493609
Schwabe, L., Schächinger, H., de Kloet, E. R., & Oitzl, M. S. (2010).
Corticosteroids operate as a switch between memory systems. Journal of
Cognitive Neuroscience, 22, 1362–1372. http://dx.doi.org/10.1162/jocn
.2009.21278
Schwabe, L., Tegenthoff, M., Höffken, O., & Wolf, O. T. (2012). Simul-
taneous glucocorticoid and noradrenergic activity disrupts the neural
basis of goal-directed action in the human brain. Journal of Neurosci-
ence, 32, 10146 –10155. http://dx.doi.org/10.1523/JNEUROSCI.1304-
12.2012
Schwabe, L., & Wolf, O. T. (2009). Stress prompts habit behavior in
humans. Journal of Neuroscience, 29, 7191–7198. http://dx.doi.org/10
.1523/JNEUROSCI.0979-09.2009
Taillard, J., Philip, P., Chastang, J.-F., & Bioulac, B. (2004). Validation of
Horne and Ostberg Morningness-Eveningness Questionnaire in a
middle-aged population of French workers. Journal of Biological
Rhythms, 19, 76 – 86. http://dx.doi.org/10.1177/0748730403259849
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
CORTISOL AND HEALTHY HABIT DEVELOPMENT

Weitzman, E. D., Fukushima, D., Nogeire, C., Roffwarg, H., Gallagher,
T. F., & Hellman, L. (1971). Twenty-four hour pattern of the episodic
secretion of cortisol in normal subjects. Journal of Clinical Endocri-
nology and Metabolism, 33, 14 –22. http://dx.doi.org/10.1210/jcem-
33-1-14
Wood, W., & Neal, D. T. (2016). Healthy through habit: Interventions for
initiating & maintaining health behavior change. Behavioral Science &
Policy, 2, 71– 83. http://dx.doi.org/10.1353/bsp.2016.0008
Wyatt, J. K., Ritz-De Cecco, A., Czeisler, C. A., & Dijk, D. J. (1999).
Circadian temperature and melatonin rhythms, sleep, and neurobehav-
ioral function in humans living on a 20-h day. American Journal of
Physiology: Regulatory, Integrative and Comparative Physiology, 277,
R1152–R1163.
Received September 7, 2016
Revision received February 24, 2017
Accepted March 3, 2017 䡲
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6FOURNIER ET AL.