Psychoneuroendocrinology 86 (2017) 144–151

Contents lists available at ScienceDirect

Psychoneuroendocrinology
journal homepage: www.elsevier.com/locate/psyneuen

Acute physical exercise in humans enhances reconsolidation of emotional MARK

memories
⁎
Dharani Keyan, Richard A. Bryant
School of Psychology, University of New South Wales, NSW, 2052, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: Increasing evidence suggests that when a memory is reactivated through retrieval, it becomes temporarily
Exercise vulnerable to environmental or pharmacological manipulation, which can consequently update or strengthen the
Glucocorticoid memory. Physical exercise has been shown to modulate the maintenance of fear memories in animals following
Memory reconsolidation memory reactivation. This study investigated the effect of intense exercise in modulating the reconsolidation of
Trauma memories
trauma memories. Fifty-four undergraduate students watched a trauma film depicting the aftermath of a
highway car crash. Two days later, participants engaged in either (a) 20–25 min of incremental cycling following
a memory reactivation induction (Reactivation/Exercise), (b) 20–25 min of mild cycling (Reactivation/No
Exercise) following memory reactivation, or (c) 20–25 min of incremental cycling but no memory reactivation
(No Reactivation/Exercise). Saliva samples were collected to index salivary amylase and cortisol at baseline and
post activity. Participants completed memory questionnaires relating to declarative and intrusive memory recall
two days after memory reactivation. Reactivation/Exercise participants recalled more central details of the
trauma film relative to other participants. Increased cortisol predicted better total memory recall in the
Reactivation/Exercise, but not in the other conditions. These findings suggest that intense exercise during the
period of memory reactivation enhances subsequent trauma memory, and provides human evidence consistent
with recent findings of exercise-induced fear reconsolidation in animals.

1. Introduction
Memory for emotionally arousing events are typically stronger than
non-emotional events, with particularly greater recall of central details
of these memories (Adolphs et al., 2005; Shields et al., 2017). This
memory bias for emotionally significant information is driven by the
release of noradrenaline and cortisol at the time of an emotionally
arousing event (Cahill et al., 2003; Roozendaal et al., 2004,2006),
where an interaction between these catecholamines is modulated
within the basolateral amygdala (McGaugh, 2004). Whilst recall of
central information is arguably adaptive in navigating subsequent
threatening experiences, overconsolidated memories for emotionally
arousing details with accompanying high levels of distress can lead to
debilitating psychopathological states, such as posttraumatic stress
disorder (PTSD; Pitman, 1989).
Increasing evidence suggests that when a memory is reactivated it
returns to a labile, protein-synthesis dependant state (Nader et al.,
2000), making it susceptible to modifications in the form of environ-
mental (Schiller et al., 2010) or pharmacological (Kindt et al., 2009)
manipulations. Research suggests that a mere reminder is not sufficient
in inducing this labile state (Forcato et al., 2009), instead a prediction
error involving a mismatch between expected and actual experience
must be accompanied in order for the original memory trace to be
destabilised (Sevenster et al., 2014). Empirical evidence suggests that a
destabilised memory trace can be updated by incorporating new in-
formation (Forcato et al., 2007; Hupbach et al., 2007), or modulating its
strength (Forcato et al., 2014). Glucocorticoid administration following
memory reactivation impairs recall of a contextual fear memory in
animals (Cai et al., 2006), with some evidence suggesting that this ef-
fect may be selective to strong emotional memories (Abrari et al.,
2008). In humans, glucocorticoid administration following memory
reactivation enhances conditioned fear (Drexler et al., 2015), and de-
clarative emotional memory (Marin et al., 2010). In contrast, beha-
vioral stress following memory reactivation resulted in a memory im-
pairing effect (Drexler and Wolf, 2017), which suggests that cortisol
administration and behavioral stress have differential effects on
memory reconsolidation. However, there is evidence that experiencing
stress during memory reactivation may not influence declarative
memory but does increase intrusive memories (Cheung et al., 2015).
Further, stress following memory reactivation can impair memory for

⁎
Corresponding author.
E-mail address: r.bryant@unsw.edu.au (R.A. Bryant).

http://dx.doi.org/10.1016/j.psyneuen.2017.09.019
Received 23 May 2017; Received in revised form 26 August 2017; Accepted 21 September 2017
0306-4530/ © 2017 Published by Elsevier Ltd.
D. Keyan, R.A. Bryant
negative stimuli encoded five weeks prior (Tollenaar et al., 2008). The
discrepancies in terms of declarative memory recall for emotional sti-
muli may be attributable to the strength of memory, nature of stimuli,
and differential modulation of the hypothesized reconsolidation
window and related timing of the stressor (i.e., cortisol exposure during
vs after memory reactivation). Additionally, the timing of glucocorti-
coid administration in relation to new learning appears critical in fa-
cilitating subsequent memory (de Quervain et al., 2017), with evidence
to suggest that delayed glucocorticoid exposure can impair retrieval of
memory (Joels et al., 2006). Further, noradrenergic activity appears to
play a role in the persistence of emotional memories, with evidence that
blocking the noradrenergic system (with propranolol) prior to re-
activation eliminates conditioned fear (Kindt et al., 2009), and selec-
tively impairs negative episodic memories (Schwabe et al., 2012).
Given the encoding (van Stegeren et al., 2007), and storage of emo-
tionally arousing experiences (van Stegeren, 2008) depend upon the
interactive relationship between glucocorticoids & noradrenaline, it is
conceivable that a similar relationship may at least in part underlie the
reconsolidation of emotional memories as well.
A role for physical exercise in modulating emotional memory re-
consolidation arises from evidence that exercise-induced mechanisms
underpin stabilisation of memories. In particular, memory stabilisation
appears to be time-dependant with the specific activation of molecular
mechanisms responsible for learning and memory (Hötting and Röder,
2013; Roig et al., 2013,2016). Additionally, acute aerobic exercise in-
duces significant increases in glucocorticoids (Hötting et al., 2016;
Wahl et al., 2010), and blockade of exercise-induced glucocorticoid
production attenuates memory retention in animals (Hajisoltani et al.,
2011). Similarly, acute exercise activates noradrenaline (Segal et al.,
2012; Winter et al., 2007) and plays a crucial role in exercise-induced
memory consolidation (Segal et al., 2012). There is increasing evidence
in animals (Siette et al., 2014), and humans (Keyan and Bryant,
2017a,b) that acute exercise following encoding of fear memories in-
creases the strength of the memory trace. A recent animal study de-
monstrated a role for exercise in the reconsolidation of fear memories
by demonstrating that wheel running immediately after brief exposure
to a previously fear conditioned context subsequently displayed more
fear (as indexed by freezing behavior) on a delayed retention test re-
lative to sedentary controls (Siette et al., 2014). Moreover, the distance
covered in the wheel running following brief exposure was positively
correlated with levels of freezing (Siette et al., 2014), suggesting that
more intense exercise following memory reactivation corresponded
with increased emotional memory strength. In support of this notion,
the dose response of exercise on memory enhancement appears to be
critically mediated by brain derived neurotrophic factor (BDNF; Winter
et al., 2007), a growth factor implicated in the consolidation (Schulz-
Klaus et al., 2013), and recently the reconsolidation of contextual fear
memories in animals (Giachero et al., 2013). Given the available evi-
dence, it is possible that exercise plays a role in the reconsolidation of
distressing emotional memories, and this may be dependent on the
intensity of the exercise and subsequent activation of critical biological
mechanisms in implicated in memory reconsolidation.
The current study aimed to extend previous animal research (Siette
et al., 2014) by investigating the role of acute exercise in modulating
the reconsolidation of emotional memories. An aerobic cycling task was
chosen as this mode and level of intensity has been shown to induce
substantial learning and memory benefits (Roig et al., 2013), elevate
noradrenergic levels (Segal et al., 2012), and hypothalamus-pituitary-
adrenal (HPA) responsivity (Brownlee et al., 2005). An analogue
trauma (film) paradigm was utilized as it elicits significant stress re-
sponses (Cheung et al., 2015), and distressing emotional memories in
non-clinical populations (Holmes and Bourne, 2008). On Day 1 of the
experiment, participants viewed a trauma film. On Day 3, participants
were randomised to either, (a) Reactivation/Exercise (acute exercise
immediately following memory reactivation), (b) Reactivation
(memory reactivation alone), or (c) Exercise (exercise alone).On Day5
145
Psychoneuroendocrinology 86 (2017) 144–151
participants completed a surprise cued recall and measure of intrusive
memories. Based upon the available evidence that acute stress enhances
reactivated emotional memories (Bos et al., 2014; Cheung et al., 2015),
and exercise functions to enhance memory, we hypothesised that par-
ticipants who received both memory reactivation and exercise would
display better cued recall and more intrusions on a delayed memory
test. We also hypothesized that levels of cortisol and salivary amylase
would be positively associated with reconsolidated emotional memories
of the trauma film.
2. Method
2.1. Participants
The sample included 54 healthy University of New South Wales
undergraduate students (26 female, 28 male) aged between 17 and 36
years (M = 19.48, SD = 3.03) who participated in exchange for course
credit. Participants were randomly allocated to a Reactivation/Exercise
(n = 18; 11 male), Reactivation/No exercise (n = 18; 9 male), or No
Reactivation/Exercise condition (n = 18; 8 male).
2.2. Measures
2.2.1. Depression anxiety and stress scales-21 item version
Baseline depression was assessed using the Depression Anxiety and
Stress Scales (DASS; Lovibond and Lovibond, 1995) 21-item version,
which indexes depression, anxiety and stress. This measure possesses
strong reliability and internal consistency (Lovibond and Lovibond,
1995).
2.2.2. Intrusions questionnaire
Selected items from the intrusion subscale of the Impact of Event
Scale (IES; Horowitz et al., 1979) was used to measure intrusive
memories of the trauma film. Participants were instructed to rate the
degree to which they identified with each item on a five-point Likert
scale (0 = Not at all, 4 = Extremely). The items were Pictures about it
popped into mind, Other things kept making me think about it, and I thought
about it when I didn’t mean to. Items were framed to capture memory
occurrences of the film presented to participants initially, and in the
intervening two days since the experiment.
2.2.3. Cued recall test
A 25-item questionnaire that has been previously adapted to the car
accident film was used to index intentional recall in the current study
(Devilly et al., 2007). This questionnaire included 12 items relating
directly to the central features (e.g., How many of the injured victims had
dark skin and how many had light skin?), and 13 items relating to per-
ipheral/surroundings (e.g. How many police vehicles surrounded in the
scene) of the car accident. Central features of the film were defined as
elements directly relating to the victims, whereas peripheral features
were defined as elements not directly relating the victims, and therefore
elements that captured background details (Devilly et al., 2007).
2.2.4. Current distress
Participants’ distress associated with the trauma film was assessed
by rating an 8-point Likert scale (0 = Not at all, 7 = Extremely).
2.2.5. Pre-exercise questionnaire
Participants were asked to indicate if they were currently suffering
from any health and/or physical conditions that would prevent them
from engaging in strenuous physical activity (1 participant was ex-
cluded for this reason). Examples included high blood pressure, asthma,
chronic pain, and a current heart condition; endorsement of any con-
dition resulted in exclusion from the study. In addition, participants did
not report taking any hormonal contraceptives on this questionnaire.
D. Keyan, R.A. Bryant
2.2.6. Godin-Shephard leisure-time exercise questionnaire (LTEQ)
The Leisure-Time Exercise Questionnaire (LTEQ; Godin and
Shephard, 1997) was used to index the frequency of light-intensity,
moderate-intensity, and vigorous-intensity leisure-time physical ac-
tivity undertaken on a weekly basis. This questionnaire has good con-
struct validity and strong test re-test reliability (Godin and Shephard,
1997). The LTEQ was modified by asking participants to indicate how
much time (in min) is spent on strenuous, moderate and light activity
for more than 15 min in a given 7-day period, such that the duration of
engagement in each type of activity could be assessed, and factored into
when estimating weekly metabolic equivalent (METS); this was done
using the formula [(total METS = minutes of strenuous exercise/
15) × 9] + [(total minutes of moderate exercise/15) × 5] + [(total
minutes of light exercise/15) × 3] (Andrykowski et al., 2007).
2.2.7. Heart rate measurement
A Garmin FR70 heart rate watch and chest belt was used to record
participants’ heart rate. This monitoring belt consists of two smart
fabric sensors to acquire cardiac activity. HR data was wirelessly
transmitted to the Garmin Connect Training centre and was stored as
average beats per minute over the 10 min period. Heart rate data was
sampled at a rate of 2.4 GHz.
2.2.8. Perceived exertion
To determine how effortfully participants were working on the as-
signed exercise tasks, subjective ratings of exercise intensity were
measured using the Borg Rating of Perceived Exertion scale (RPE; Borg,
1982). This measure takes into account individual fitness level by re-
quiring participants to match how hard they feel they are working on a
15-point scale (6 = No exertion at all, 20 = Maximal exertion). Mod-
erate activity is classified as 11–14, and vigorous activity as 15 or
above. This measure provides an approximation of heart rate by mul-
tiplying a given rating by 10, for example, if an individual’s rating of
perceived exertion (RPE) is 15, then 15 × 10 = 150, suggesting that
their heart rate should approximate 150 beats per minute (Borg, 1982;
Johnson and Prins, 1991).
2.2.9. Salivary hormone collection
Circulating cortisol and alpha amylase (sAA) were indexed via
saliva sampling before and after the exercise interventions. Saliva
samples were acquired using the absorbent devise technique, which
required participants to place a cotton swab in their mouth for 1–2 min
until saturation after which the swab was placed back into the storage
tube. Collected saliva samples were immediately frozen at −20 °C until
assay. To reduce non-specific variations in stress hormone levels, par-
ticipants were requested to refrain from taking alcohol, caffeine or
smoking 3 h, eating 1 h, and exercising 24 h prior to both experimental
sessions. No participants were taking medications interacting with the
noradrenergic response (e.g., beta blockers). Peak levels of endogenous
hormones were assessed immediately, and 20–25 min post the exercise
intervention for sAA and cortisol, respectively.
2.3. Materials
2.3.1. Trauma film
A film depicting the actual aftermath of a highway car crash was
employed as it has been validated as an effective emotional memory
stimulus (Small et al., 2011). This film involves 10 min of live footage
depicting emergency workers attending the scene of a motor vehicle
road accident. The video focuses on assistance given to surviving vic-
tims, and a fatality resulting from the car crash. Additionally, sounds of
distressed victims and sirens, and words spoken by emergency workers
are presented.
2.3.2. Memory reactivation
A combination of a reminder cue and retrieval was utilised to
146
Psychoneuroendocrinology 86 (2017) 144–151
reactivate participants’ memory of the trauma film. An audio clip
(2 min duration), including ambulance sirens and the voice of dis-
tressed victims, was extracted from the original trauma film and served
as a reminder cue, which was presented on a 12 in. laptop with head-
phones. Following this audio clip, participants were required to bring to
mind the car accident film and imagine it in as much detail possible for
a further three minutes. Participants were informed that they would be
asked some questions about the video at the end of the imaginal period.
The questions at the end included What percentage of the time did you
spend thinking about the video?, and How vivid is the memory of the video,
on a scale of 1–10? which served as manipulation checks for the memory
reactivation.
2.3.3. Exercise task
An adaption of the maximal cycle protocol test of aerobic fitness
validated by Storer et al., 1990 was employed as the exercise inter-
vention. This task required participants to engage in 20–25 min of in-
cremental cycling at a cadence between 60 and 70 revolutions per
minute (RPM) on a manual cycle ergometer, (Monark 828E, Sweden).
Participants began with a 3-min warm up period, after which the
workload was increased by 1/4 kg (i.e., kilopond) per minute until
participants reached their limit of tolerance, which was defined as
ability to no longer maintain a cadence of at least 60 RPM (i.e., max-
imal resistance; Storer et al., 1990). On average, it was observed that
participants took 10 min to reach their maximal resistance level, and
thereafter engaged at their submaximal resistance level for at least
12 min of the exercise intervention. In order to obtain a genuine max-
imal resistance, participants’ exertion levels were monitored using in-
termittent RPE ratings, and were also verbally encouraged by the ex-
perimenter to provide a true maximal effort. After this, frictional
resistance on the bike was reduced just enough to meet the minimum
cadence (i.e., 60 RPM; submaximal resistance), which was thereafter
maintained until the end of the task. The experimenter closely mon-
itored participants’ exercise performance such that when heart rate
dropped below a minimum standard (arbitrary rate of 160 bpm), par-
ticipants were notified and encouraged to increase the intensity ac-
cordingly. Participants’ maximal and submaximal workload (or watt)
on the cycle ergometer was determined using the following formula,
Work (kgm) = force (kg) × distance (meters/min); where force relates
to the frictional resistance applied on the bike as measured in Kg, and
distance refers the cadence (rev/min) multiplied by the a constant of 6
(i.e., whereby 6 m per revolution of the flywheel applies to the Monark
ergometer). For example, if an individual is working at a frictional re-
sistance of 3 kg at 65 RPM, then the work level (kgm) engaged would be
defined as 3 kg × 65 RPM × 6m/rev = 1170. Work level is then ex-
pressed in watts approximately equivalent to 6kgm/min, thus the in-
dividual in the above example would be exercising at a 195 W
(1170 kpm/min divided by 6). Accordingly, maximal watt was calcu-
lated using the highest level of resistance endured by participants at a
minimum cadence of 65 rpm. Submaximal watt was calculated using
the submaximal resistance maintained at a minimum of speed of
65 rpm. The reactivation alone condition engaged in 20–25 min of easy
cycling at a 0-resistance level, and cadence between 60 and 70 RPM,
and this was designed to match the exercise conditions in terms of
cognitive interference (i.e., control for differences in distraction be-
tween conditions). Given research to suggest that fitness levels mod-
erate the extent to which exercise impacts learning and memory (Roig
et al., 2013), and related cortisol production (Hill et al., 2008), a va-
lidated maximal cycle protocol (Storer et al., 1990) was chosen. Whilst
participants’ true fitness level was not determined via prior exercise
testing, the protocol was modified to still incorporate individualised
workload and related thresholds during the exercise interventions.
2.4. Procedure
To reduce circadian variation in cortisol, all experimentation was
D. Keyan, R.A. Bryant
conducted between 12.00 and 1800 (Cheung et al., 2015). On Day 1
informed consent was obtained, questionnaires completed, and those
who responded with a “moderate” or greater level of depression, stress
or anxiety on the DASS-21 were excluded from partaking in the study so
as to not further aggravate any depressive and/or anxious symptoms by
exposure to the trauma film (nil were excluded for this reason). Re-
maining participants were randomised to Reactivation/Exercise, Re-
activation/No Exercise, or No Reactivation/Exercise. Following this,
participants watched the trauma film and were instructed to maintain
attention to the film content, which was monitored by the experimenter
who remained out of sight. Post viewing, participants were instructed to
close their eyes for 7 min (post encoding silence), after which the dis-
tress rating (Time 1) associated with the film was assessed. Finally, the
session was terminated and participants were reminded to attend ses-
sion 2 two days later. On Day 3, all participants returned to a different
room, and this change in context was implemented to introduce a novel
element prior to reactivation in the reactivation conditions, which is
recommended for facilitating updating during memory reactivation
(Bonin and De Koninck, 2015), and to also limit the trauma film
memory from being reactivated in No Reactivation/Exercise condition.
Participants habituated to this new environment for 20 min in order
collect saliva samples to index baseline alpha amylase and cortisol.
Following this, the Reactivation conditions received the 5 min re-
activation paradigm, whilst the Exercise alone condition completed a
neutral word search task for the same amount of time. This was fol-
lowed by either the exercise or no exercise interventions. A third saliva
sample was collected immediately after these tasks (to sample sAA),
and participants were then required to close their eyes for 5 min. Par-
ticipants read magazines of an emotionally neutral nature for a further
15 min until the final saliva sample was collected (to sample cortisol)
followed by a second distress rating. On Day 5, the online survey was
distributed, which included a cued recall test, intrusions questionnaire,
and another distress rating (see Fig. 1).
147
Psychoneuroendocrinology 86 (2017) 144–151
2.5. Saliva analyses
Saliva analyses were performed at the University of New South
Wales (Sydney, Australia) using commercially available ELISA kits
(Cayman Chemical, USA; Salimetrics, USA). All saliva samples were
stored frozen at −20 °C until assay, and underwent one freeze/thaw
cycle. Thawed samples were centrifuged at 1500 × g for 15 min to
collect clear saliva, and brought to room temperature prior to assay
performance. The Cayman Chemical assay for cortisol utilises a com-
petitive immunoassay whereby the concentration of cortisol for each
sample is determined by interpolation from a standard curve generated
by plotting known cortisol concentrations (6.6–4000 pg/mL) against
their average absorbance divided by the background absorbance (%B/
B0). The Salimetrics assay uses the principle that sAA catalyses the
substrate maltotriose linked with a chromagenic substrate, 2-chloro-p-
nitrophenol. The amount of sAA activity present in the sample is di-
rectly proportional to the increase in absorbance, and this enzymatic
action can be spectrophotometrically measured at 405 nm. The inter-
and intra-assay variability, respectively was 8.8% and 6.5% for cortisol,
and 6.8% and 4.3% for sAA. Raw concertation scores for sAA and CORT
were square-root transformed to increase normality of distributions
prior to stress hormone analyses.
2.6. Data analyses
SPSS statistical package (v.23) was used to conduct data analyses.
Separate 3 (Experimental Condition) × 2 (Time) mixed-model ANOVAs
indexed changes in sAA and cortisol levels pre and post the exercise
interventions. Separate one-way ANOVAs across groups were con-
ducted to analyse cued recall for central and peripheral details, and
intrusive memory scores. To determine the relative roles of stress hor-
mones in influencing the reconsolidation or stabilisation of emotional
memories, increase in sAA and cortisol (i.e., change from baseline to
post exercise/no exercise intervention) were entered into a linear re-
gression model to predict total cued recall and intrusions within each
Fig. 1. Timeline for experimental sequence.
D. Keyan, R.A. Bryant Psychoneuroendocrinology 86 (2017) 144–151

Table 1
Participant characteristics.

Reactivation and Exercise (n = 18) Reactivation alone (n = 18) Exercise alone (n = 18)

Age 18.78 (1.06) 19.56 (2.28) 20.11 (4.61)

DASS-Depression 3.89 (3.46) 3.72 (2.56) 4.00 (4.64)
DASS-Anxiety 3.28 (3.04) 3.83 (3.03) 3.72 (2.85)
DASS-Stress 5.67 (3.99) 5.33 (3.18) 6.33 (4.60)
Baseline sAA (U/mL) 63.79 (41.53) 51.62 (37.79) 78.53 (45.82)
Post sAA (U/mL) 144.04 (77.04) 79.38 (69.79) 173.88 (105.25)
Baseline CORT (μg/dL) 0.25 (0.29) 0.16 (0.10) 0.15 (0.09)
Post CORT (μg/dL) 0.55* (0.48) 0.09 (0.06) 0.30 (0.15)
LTEQ-strenuous minutes 80.00 (630.00) 52.50 (300.00) 105.00 (540.00)
LTEQ- moderate minutes 30.00 (240.00) 35.00 (240.00) 50.00 (300.00)
LTEQ- mild minutes 60.00 (150.000) 60.00 (1080.00) 10.00 (420.00)
Total METs 107.00 (378.00) 77.00 (261.00) 91.50 (377.33)
Baseline HR 72.12* (9.21) 68.82 (10.60) 63.28 (10.13)
Maximum HR 188.06 (7.38) 89.50 (6.98) 179.39 (10.22)
Maximal watt 162.50 (32.00) 0.00 168.82 (42.86)
Submaximal watt 125.49 (27.34) 0.00 127.29 (43.26)
%Thinking about trauma film 65.28 (19.53) 71.36 (18.35) N/A
Vividness of trauma film 6.47 (1.40) 7.08 (1.75) N/A
Central detail recall 6.89* (1.45) 5.50 (2.07) 5.33 (1.81)
Peripheral detail recall 2.83 (0.99) 2.83 (1.04) 2.78 (0.88)
Intrusive memory scores 5.72 (2.74) 6.56 (3.43) 6.33 (3.18)

Note. Means and standard deviations in parentheses have been reported. Raw concentration data for CORT and sAA have been provided. However square-root transformed values were
used in the analyses. Median and range (in parentheses) for LTEQ subscales and total METs are presented.
* p < 0.05.

group. To account for individual differences in daily levels of exercise, conditions relative to the Reactivation alone condition [t(2.48)
weekly energy expenditure (i.e., total METs) was added as a covariate = 37.22, p < 0.001]. These findings suggest that the exercise inter-
in these regression analyses. vention was effective in elevating participants’ heart rate.

3. Results 3.2.2. Physiological response

Fig. 2 presents mean concentrations of sAA and cortisol pre- and
3.1. Participant characteristics post-exercise interventions. A 3 (Experimental Condition) × 2 (Time,

Table 1 presents participant characteristics for each experimental

condition. One-way ANOVAs revealed no pre-existing differences be-
tween groups in age and DASS-21 scores. Additionally, separate one-
way ANOVAs revealed no pre-existing differences between groups in
and baseline sAA (F(2, 51) = 1.90, p = 0.161) and cortisol (F(2, 50)
= 1.71, p = 0.191) concentrations. Further, gender did not sig-
nificantly differ across experimental conditions (χ2 = 0.1.04,
p = 0.595). A Kruskal-Wallis H test displayed no pre-existing differ-
ences in engagement of regular physical activity (i.e., total METs)
across conditions,
χ2 = 0.80, p = 0.669. A chi-square test revealed no significant
differences between experimental groups in time of day in which ex-
ercise testing was conducted [χ2 = 9.45, p = 0.490].

3.2. Manipulation checks

3.2.1. Heart rate

Participants’ average resting and maximum heart rate data are
presented in Table 1. A one-way ANOVA revealed that resting [F(2.49)
= −3.51, p = 0.038] and maximum [F(2.48) = 695.62, p < 0.001]
heart rates differed across conditions. Follow-up comparisons revealed
that baseline HR was significantly higher in the Reactivation/Exercise
relative to the Exercise alone condition [t(2.49) = 2.61, p = 0.012].
Follow-up correlations revealed that baseline heart rate was sig-
nificantly correlated with baseline sAA (r = −0.484, p = 0.042), post
sAA (r = −0.493, p = 0.038), and total METS (r = −0.546,
p = 0.019) in the Exercise alone condition. Similar significant corre-
lations were not observed in Reactivation/Exercise condition
(ps > 0.05). Due to the potential, that baseline heart rate could
moderate differences in amylase response and cued recall; this is added Fig. 2. Salivary alpha amylase (sAA) and cortisol levels prior to-, and following the ex-
as a covariate in later ANOVA and regression analyses. Maximum HR ercise interventions (Day 3).
Note. Error bars reflect standard error of the means.
was significantly higher in the Reactivation/Exercise and Exercise alone

148
D. Keyan, R.A. Bryant Psychoneuroendocrinology 86 (2017) 144–151

baseline, post activity) mixed-model ANCOVA of sAA concentration Table 2

controlling for baseline heart rate revealed a non-significant main effect Summary of hierarchical regression models for cued recall.
of Condition [F(2.48) = 2.81, p = 0.070], and non-significant interac-
B SEB β p
tion effect [F(2.48) = 2.09, p = 0.135]. A 3 (Experimental Condi-
tion) × 2 (Time) mixed-model ANOVA for cortisol concentration re- Reactivation and exercise
vealed a main effect of Condition [F(2, 50) = 7.12, p = 0.002], and a Step 1 Total METs 0.00 0.004 −0.01 0.96
Step 2 sAA increase −1.134 0.76 −0.34 0.16
significant interaction effect [F(2, 50) = 3.73, p = 0.031]. Follow-up
CORT increase 2.79 1.04 0.63 0.02*
comparisons revealed that the Reactivation/Exercise relative to the
Reactivation alone condition experienced an increased cortisol response Reactivation alone
Step 1 Total METs 0.01 0.01 0.20 0.54
[F(1.50) = 6.86, p = 0.012]. Additionally, both exercise conditions Step 2 sAA increase 0.57 1.28 0.12 0.67
experienced an increased cortisol response relative to Reactivation CORT increase 2.77 10.32 0.09 0.79
alone condition [F(1.50) = 7.05, p = 0.011], suggesting that the ex-
Exercise alone
ercise intervention was effective in augmenting the glucocorticoid re- Step 1 Total METs −0.002 0.01 −0.09 0.76
sponse. A secondary analysis was conducted that accounted for poten- Step 2 sAA increase 0.34 0.60 0.17 0.54
tial gender effects in light of previous research suggesting a sex CORT increase 0.78 3.54 0.06 0.83
moderation in cortisol response (Seeman et al., 2001). This 2 (Experi-
Note. SEB = standard error of B. Reactivation and exercise, Step 1, R2 = 0.04; Step 2,
mental Condition) × (Gender) ANOVA did not reveal any significant
ΔR2 = 0.39; Reactivation alone, Step 1, R2 = 0.03; Step 2, ΔR2 = 0.02; Exercise alone,
sex moderation of cortisol response [non-significant con- Step 1, R2 = 0.01; Step 2, ΔR2 = 0.03.
dition × gender interaction, F(2.47) = 0.641, p = 0.531].
3.2.7. Intrusive memories
3.2.3. Exercise intensity Table 1 presents participants’ total intrusions scores. A one-way
Participants’ maximal and submaximal workload data are presented ANOVA of intrusions scores indicated no significant differences across
in Table 1. Planned comparisons revealed that maximal and sub- groups [F(1.51) = 0.34; p = 0.712].
maximal workload (watts) endured on the cycle ergometer did not
significantly differ between the Reactivation/Exercise and Exercise 3.3. Role of stress hormones
alone conditions (ps > 0.05), suggesting that exertion levels did not
differ between these conditions. Across the Reactivation/Exercise and Controlling for average weekly energy expenditure (i.e., Total
Exercise alone conditions, participants were working on average at 76% METs), cortisol increase following acute exercise predicted cued recall
of their maximal work load (as defined by submax W/max W × 100) of the trauma film in only the Reactivation/Exercise condition (see
for at least 12 min of the exercise intervention (see Table 1). Table 2). sAA was not a significant predictor. A similar finding was not
observed in relation to intrusions (p > 0.05). The same pattern of re-
3.2.4. Reactivation induction solts was observed when weekly energy expenditure was not included
Planned comparisons revealed that the Reactivation/Exercise and as a covariate.
Reactivation alone conditions did not significantly differ in the reported
percentage time spent thinking about the film [t(1.34) = −0.96, 4. Discussion
p = 0.342], or vividness rating [t(1.34) = −1.16, p = 0.255] sug-
gesting that these conditions similarly responded to the memory re- Acute exercise following memory reactivation improved subsequent
activation paradigm. declarative memory, and is the first evidence in humans to implicate
exercise in modulating the reconsolidation of emotional memories. To
this end, the current study extends recent animal work of exercise-in-
3.2.5. Distress
duced fear memory reconsolidation (Siette et al., 2014). Importantly,
A 3 (Experimental Condition) × 3 (Time) mixed-model ANOVA of
non-specific effects of exercise were not attributable to memory en-
distress ratings associated with the film revealed a significant main
hancement given that the exercise without memory reactivation was
effect of Time [F(2.49) = 64.56; p < 0.001]. Specifically, participants’
not beneficial to recall performance. These findings need to be under-
distress was highest immediately after viewing the film, and then di-
stood within the context of evidence that biological mechanisms that
minished over time. There was no significant difference between con-
can be induced by a bout of acute exercise have a modulatory role in
ditions over time.
the reconsolidation of emotionally arousing memories (Cai et al., 2006;
Marin et al., 2010; Bos et al., 2014).
3.2.6. Cued memory recal The specificity of exercise-induced memory enhancement for central
Mean cued recall scores are presented in Table 1. The mean recall of but not background/peripheral details of the trauma film requires ex-
central and peripheral memories are consistent with levels reported in planation. We speculate that this pattern of findings may be attributed
previous studies (Keyan and Bryant, 2017a,b). A one-way ANOVA of to the nature of the memory reactivation paradigm and subsequent
cued recall for central and peripheral details revealed a significant impact upon reconsolidation of memories. The use of a deliberate re-
between-group difference [F(2.51) = 4.08, p = 0.023]. Planned con- trieval strategy by asking participants to recall the film in ‘as much
trasts found that participants in the Reactivation/Exercise condition detail possible’, may have activated a covert rehearsal process whereby
recalled more central details of the trauma film relative to the Re- focal details of the trauma film were preferentially reactivated at the
activation alone [t(1.51) = 2.32; p = 0.024], and Exercise alone [t expense of contextual information. This proposition is in line with re-
(1.51) = 2.60; p = 0.012] conditions. Similar differences were not search suggesting a narrowing of attention towards arousing informa-
observed for peripheral details of the trauma film (ps > 0.05). A sec- tion during retrieval of traumatic memories (Christianson, 1992).
ondary analysis was conducted that accounted for potential gender Consequent acute exercise may have served to consolidate the storage
effects in memory recall, given previous research suggesting a sex of this emotionally significant information. In this instance, it is pos-
moderation of emotional memory modulation (Bryant et al., 2013). sible that intense exercise is acting as a strengthening mechanism in
This 2 (Experimental Condition) × (Gender) ANOVA of cued recall did updating the reconsolidation process given that the Reactivation alone
not reveal any significant sex moderation of memory recall [non-sig- condition did not yield similar memory benefits. Furthermore, the
nificant condition x gender interaction, F(2.48) = 1.75, p = 0.185]. heightened glucocorticoid response following acute exercise appears to

149
D. Keyan, R.A. Bryant
be playing an important role in the persistence of reactivated emotional
memories, and this finding is consistent with previous research evi-
dencing a modulatory role of stress in enhancing the reconsolidation of
neutral (Coccoz et al., 2011) and emotional (Marin et al., 2010; Bos
et al., 2014) material.
In this light, it is surprising that intense exercise did not enhance the
reconsolidation of intrusive memories. This is because activation of the
original memory trace in association with exercise-induced stress was
expected to enhance the reconsolidation of distressing intrusions. This
proposition is in line with previous research evidencing stress during
memory reactivation to enhance intrusive memories (Cheung et al.,
2015). A possible explanation for the apparent discrepancy in findings
may be attributed to differences in reactivation paradigms and sub-
sequent activation of different memory representations. Whilst an ex-
plicit recall instruction during memory reactivation has been evidenced
to modulate declarative memory (Marin et al., 2010), less deliberate
retrieval strategies encompassing brief reminder cues (Cheung et al.,
2015; James et al., 2015) have been found to modulate involuntary
intrusions. Thus, the current study’s more deliberate recall instruction
could have favored the activation of an explicit memory representation,
which in turn potentially favored the modulation of intentional re-
trieval. This raises the possibility of boundary conditions under which
explicit and implicit memory traces can be reconsolidated. Indeed,
theoretical models stipulate a dissociation between explicit and implicit
memory representations following encoding of an emotional event
(Brewin et al., 1996), whereby factors such as attentional capacity and
rehearsal are suspected to modulate the strength of the explicit memory
trace. Having stated this, it is worth noting here that the dissociation
between intrusive and deliberate memory performance is not an un-
common finding (Brewin et al., 1996). Alternatively, it is quite possible
that the lack of a difference in intrusive memory performance stems
from a methodological limitation of the current study, whereby a diary
method would have been more sensitive in tracking intrusions fol-
lowing memory reactivation on Day 3 (James et al., 2015). This tracked
method was not utilised due to the possibility that declarative memory
at test would be contaminated following the implementation of daily
memory records.
It is worth juxtaposing the current study’s findings within past re-
search that have evidenced stress following memory reactivation to
impair, rather than enhance, memory performance (Cai et al., 2006;
Abrari et al., 2008). Such discrepant findings may be explained by the
possibility that the impact of stress on emotional memory re-
consolidation follows a dose trajectory similar to consolidation whereby
extreme levels of glucocorticoids can impair memory (Elzinga et al.,
2005), whereas low doses exert a beneficial effect (Roozendaal, 2000).
It is also possible that the impact of glucocorticoids on memory recall
may be attributed to the influence of glucocorticoids on consolidation
of fear extinction rather than on memory reconsolidation (see Cai et al.,
2006).
It is important to consider that biological mechanisms underlying
acute exercise-induced memory benefits may extend beyond a heigh-
tened glucocorticoid response. The intensity of exercise achieved in the
current study is comparable to previous research in terms of consequent
impact upon BDNF production (Huang et al., 2014), and given the
crucial role of BDNF in modulating learning and memory processes
(Cunha et al., 2010), it is possible that this is another factor modulating
the impact of exercise on emotional memory reconsolidation. This is
noteworthy within the context of a recent study evidencing BDNF en-
hancement following retrieval in stressed animals exclusively (Giachero
et al., 2013). It is possible that exercise-induced BDNF has a strength-
ening influence in memory reconsolidation within the context of a
heightened glucocorticoid response. Evidence suggests that BDNF in-
teracts with glucocorticoid response in modulating the impact of ex-
ercise on emotional memory consolidation (Keyan and Bryant,
2017a,b). As exercise-induced BDNF production was not indexed in the
current study, its role in modulating the reconsolidation process must
150
Psychoneuroendocrinology 86 (2017) 144–151
be explored with further research.
In terms of limitations to the current study, the sample size was
limited in being able to explore the role of gender or the interaction
between stress hormones in modulating the reconsolidation of emo-
tional memories. Future replications should be adequately powered to
include gender as an additional factor because of the evidence that
gender can impact on cortisol response (Seeman et al., 2001) and
emotional memory modulation (Bryant et al., 2013). We also note that
timing of experimental sessions varied It is important to note that in-
dexation of sAA response within the context of cortisol response is
difficult to investigate given the opposing circadian patterns of each
stress hormone (Nater et al., 2007). All testing was conducted between
1200 and 1900 to reduce circadian variation of cortisol, however this
meant that natural sAA variation was largely increased given this time
window, which in turn could have potentially induced a ceiling effect of
sAA following exercise. This proposition may explain the lack of a re-
liable sAA response post activity following reactivation and exercise
relative to the reactivation alone condition. This is important to con-
sider in light of studies that have evidenced exercise induced nora-
drenergic activity in isolation of glucocorticoid response assessment
(Segal et al., 2012). Additionally, we note that the degree of cortisol
induced by physical exercise could be influenced by individuals’ car-
diovascular fitness and related workload capacity, and as such the
current study could be improved by matching participants based on a
prior exercise fitness test (e.g., Hill et al., 2008). Whilst the current
study encouraged participants to exercise to an individualized
threshold, cardiorespiratory fitness could be better accounted for via
fitness testing. Whilst an online survey were utilised to minimise at-
trition rates at follow-up and has been used in past research (e.g.,
Cheung et al., 2015), we recognize that this method does carry lim-
itations including variability in conditions under which questionnaires
are completed (e.g., time of day, distraction, and attention), and ex-
perimenter presence. Future replications of the study must standardise
testing times and contexts, and ideally bring participants back to the
laboratory. To avoid intentional learning of the stimuli, we did not
administer a baseline memory test of the trauma film following its
presentation. One confound of this is that we cannot preclude the
possibility that participants in the Reactivation/Exercise condition did
not consolidate their memories more strongly prior to the experimental
manipulation; this possibility is limited by the randomization of parti-
cipants to each condition. Finally, we note that we did not assess prior
trauma history; one study has shown that recency of a traumatic ex-
perience impacts on baseline cortisol levels in experimental studies
(Chou et al., 2014).
In summary, this study provides initial evidence that acute exercise
following activation of an emotional memory trace can subsequently
enhance intentional retrieval. These findings extend recent animal work
relating to exercise-induced fear memory reconsolidation, and point to
possible mechanisms for the maintenance of reactivated emotional
memories. Specifically, heightened levels of stress following reactiva-
tion of trauma memories coupled with extreme exercise-induced
arousal may contribute to the persistence of traumatic memories.
Future study of this issue may lead to possible management of clinical
disorders characterized by intrusive memories. For example, trials
could test whether limiting exercise after patients with PTSD reactivate
their trauma memories reduces memories of the trauma. Development
of our understanding of how brief exercise reconsolidates traumatic
memories potentially offers cheap and accessible means to modify the
intensity of this core feature of a range of clinical disorders.
Conflict of interests
No authors have any actual or potential conflicts of interest related
to this study.
D. Keyan, R.A. Bryant
Role of funding source
This work was supported by an NHMRC program grant (1073041).
References
Abrari, K., Rashidy-Pour, A., Semnanian, S., Fathollahi, Y., 2008. Administration of
corticosterone after memory reactivation disrupts subsequent retrieval of a con-
textual conditioned fear memory, dependence upon training intensity. Neurobiol.
Learn. Mem. 89 (2), 178–184.
Adolphs, R., Tranel, D., Buchanan, T.W., 2005. Amygdala damage impairs emotional
memory for gist but not details of complex stimuli. Nat. Neurosci. 8, 512–518.
Andrykowski, M.A., Beacham, A.O., Jacobsen, P.B., 2007. Prospective, longitudinal study
of leisure-time exercise in women with early-stage breast cancer. Cancer Epidemiol.
Biomark. Prev. 16 (3), 430–438.
Bonin, R.P., De Koninck, Y., 2015. Reconsolidation and the regulation of plasticity,
moving beyond memory. Trends Neurosci. 38 (6), 336–344.
Borg, G.A., 1982. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc. 14
(5), 377–381.
Bos, M.G., Schuijer, J., Lodestijn, F., Beckers, T., Kindt, M., 2014. Stress enhances re-
consolidation of declarative memory. Psychoneuroendocrinology 46, 102–113.
Brewin, C.R., Dalgleish, T., Joseph, S., 1996. A dual representation theory of posttrau-
matic stress disorder. Psychol. Rev. 103 (4), 670–686.
Brownlee, K.K., Moore, A.W., Hackney, A.C., 2005. Relationship between circulating
cortisol and testosterone, influence of physical exercise. J. Sports Sci. Med. 4 (1),
76–83.
Bryant, R.A., McGrath, C., Felmingham, K.L., 2013. The role of noradrenergic and glu-
cocorticoid interaction in the development of intrusive memories. PLoS One 8,
e62675.
Cahill, L., Gorski, L., Le, K., 2003. Enhanced human memory consolidation with post
learning stress, interaction with the degree of arousal at encoding. Learn. Mem. 10
(4), 270–274.
Cai, W.H., Blundell, J., Han, J., Greene, R.W., Powell, C.M., 2006. Postreactivation glu-
cocorticoids impair recall of established fear memory. J. Neurosci. 26 (37),
9560–9566.
Cheung, J., Garber, B., Bryant, R.A., 2015. The role of stress during memory reactivation
on intrusive memories. Neurobiol. Learn. Mem. 123, 28–34.
Christianson, S.A., 1992. Emotional stress and eyewitness memory, a critical review.
Psychol. Bull. 112 (2), 284–309.
Coccoz, V., Maldonado, H., Delorenzi, A., 2011. The enhancement of reconsolidation with
a naturalistic mild stressor improves the expression of a declarative memory in hu-
mans. Neuroscience 185, 61–72.
Cunha, C., Brambilla, R., Thomas, K.L., 2010. A simple role for BDNF in learning and
memory? Front. Mol. Neurosci. 3, 1.
de Quervain, Dominique, Schwabe, Lars, Roozendaal, Benno, 2017. Stress, glucocorti-
coids and memory: implications for treating fear-related disorders. Nat. Rev.
Neurosci. 18 (1), 7–19. http://dx.doi.org/10.1038/nrn.2016.155.
Devilly, G.J., Varker, T., Hansen, K., Gist, R., 2007. An analogue study of the effects of
psychological debriefing on eyewitness memory. Behav. Res. Ther. 45 (6),
1245–1254.
Drexler, S.M., Wolf, O.T., 2017. Stress disrupts the reconsolidation of fear memories in
men. Psychoneuroendocrinology 77, 95–104.
Drexler, S.M., Merz, C.J., Hamacher-Dang, T.C., Tegenthoff, M., Wolf, O.T., 2015. Effects
of cortisol on reconsolidation of reactivated fear memories.
Neuropsychopharmacology 40 (13), 3036–3043.
Elzinga, B., Bakker, M.A., Bremner, J.D., 2005. Stress-induced cortisol elevations are
associated with impaired delayed, but not immediate recall. Psychiatry Res. 134 (3),
211–223.
Forcato, C., Burgos, V.L., Argibay, P.F., Molina, V.A., Pedreira, M.E., Maldonado, H.,
2007. Reconsolidation of declarative memory in humans. Learn. Mem. 14 (4),
295–303.
Forcato, C., Argibay, P.F., Pedreira, M.E., Maldonado, H., 2009. Human reconsolidation
does not always occur when a memory is retrieved, the relevance of the reminder
structure. Neurobiol. Learn. Mem. 91 (1), 50–57.
Forcato, C., Fernandez, R.S., Pedreira, M.E., 2014. Strengthening a consolidated memory,
the key role of the reconsolidation process. J. Physiol. 108 (4–6), 323–333.
Giachero, M., Bustos, S.G., Calfa, G., Molina, V.A., 2013. A BDNF sensitive mechanismis
involved in the fear memory resulting from the interaction between stress and the
retrieval of an established trace. Learn. Mem. 20 (5), 245–255.
Godin, G., Shephard, R.J., 1997. Godin leisure-time exercise questionnaire. Med. Sci.
Sports Exerc. 29, 36–38.
Hötting, K., Röder, B., 2013. Beneficial effects of physical exercise on neuroplasticity and
cognition. Neurosci. Biobehav. Rev. 37 (9 Pt B), 2243–2257.
Hötting, K., Schickert, N., Kaiser, J., Röder, B., Schmidt-Kassow, M., 2016. The effects of
acute physical exercise on memory, peripheral BDNF, and cortisol in young adults.
Neural Plas. 1–12.
Hajisoltani, R., Rashidy-Pour, A., Vafaei, A.A., Ghaderdoost, B., Bandegi, A.R., Motamedi,
F., 2011. The glucocorticoid system is required for the voluntary exercise-induced
enhancement of learning and memory in rats. Behav. Brain Res. 219 (1), 75–81.
http://dx.doi.org/10.1016/j.bbr.2010.12.005.
Hill, E.E., Zack, E., Battaglini, C., Viru, M., Viru, A., Hackney, A.C., 2008. Exercise and
circulating cortisol levels: the intensity threshold effect. J. Endocrinol. Invest. 31 (7),
587–591. http://dx.doi.org/10.1007/bf03345606.
Holmes, E.A., Bourne, C., 2008. Inducing and modulating intrusive emotional memories,
a review of the trauma film paradigm. Acta Psychol. 127 (3), 553–566.
Horowitz, M.J., Wilner, N., Alvarez, W., 1979. Impact of events scale: a measure of
151
Psychoneuroendocrinology 86 (2017) 144–151
subjective stress. Psychosom. Med. 41 (3), 209–218.
Huang, T., Larsen, K.T., Ried-Larsen, M., Moller, N.C., Andersen, L.B., 2014. The effects of
physical activity and exercise on brain-derived neurotrophic factor in healthy hu-
mans, a review. Scand. J. Med. Sci. Sports 24 (1), 1–10.
Hupbach, A., Gomez, R., Hardt, O., Nadel, L., 2007. Reconsolidation of episodic mem-
ories, a subtle reminder triggers integration of new information. Learn. Mem. 14
(1–2), 47–53.
James, E.L., Bonsall, M.B., Hoppitt, L., Tunbridge, E.M., Geddes, J.R., Milton, A.L.,
Holmes, E.A., 2015. Computer game play reduces intrusive memories of experimental
trauma via reconsolidation-update mechanisms. Psychol. Sci. 26 (8), 1201–1215.
Joels, M., Pu, Z., Wiegert, O., Oitzl, M.S., Krugers, H.J., 2006. Learning under stress: how
does it work? Trends Cognit. Sci. 10 (4), 152–158.
Johnson, J.H., Prins, A., 1991. Prediction of maximal heart rate during a submaximal
work test. J. Sports Med. Phys. Fit. 31 (1), 44–47.
Keyan, D., Bryant, R.A., 2017a. Brief exercise enhances intrusive memories of traumatic
stimuli. Neurobiol. Learn. Mem. 141, 9–13.
Keyan, D., Bryant, R.A., 2017b. Role of BDNF val66met polymorphism in modulating
exercised-induced emotional memories. Psychoneuroendocrinology 77, 150–157.
Kindt, M., Soeter, M., Vervliet, B., 2009. Beyond extinction, erasing human fear responses
and preventing the return of fear. Nat. Neurosci. 12 (3), 256–258.
Lovibond, P.F., Lovibond, S.H., 1995. The structure of negative emotional states, com-
parison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and
Anxiety Inventories. Behav. Res. Ther. 33 (3), 335–343.
Marin, M.F., Pilgrim, K., Lupien, S.J., 2010. Modulatory effects of stress on reactivated
emotional memories. Psychoneuroendocrinology 35 (9), 1388–1396.
McGaugh, J.L., 2004. The amygdala modulates the consolidation of memories of emo-
tionally arousing experiences. Annu. Rev. Neurosci. 27, 1–28.
Nader, K., Schafe, G.E., LeDoux, J.E., 2000. The labile nature of consolidation theory. Nat.
Rev. Neurosci. 1 (3), 216–219.
Nater, U.M., Rohleder, N., Schlotz, W., Ehlert, U., Kirschbaum, C., 2007. Determinants of
the diurnal course of salivary alpha-amylase. Psychoneuroendocrinology 32 (4),
392–401.
Pitman, R.K., 1989. Post-traumatic stress disorder, hormones, and memory. Biol.
Psychiatry 26 (26), 221–223.
Roig, M., Nordbrandt, S., Geertsen, S.S., Nielsen, J.B., 2013. The effects of cardiovascular
exercise on human memory, a review with meta-analysis. Neurosci. Biobehav. Rev.
37 (8), 1645–1666.
Roig, M., Thomas, R., Mang, C.S., Snow, N.J., Ostadan, F., Boyd, L.A., Lundbye-Jensen, J.,
2016. Time-dependent effects of cardiovascular exercise on memory. Exerc. Sport Sci.
Rev. 44 (2), 81–88.
Roozendaal, B., Hahn, E.L., Nathan, S.V., de Quervain, D.J., McGaugh, J.L., 2004.
Glucocorticoid effects on memory retrieval require concurrent noradrenergic activity
in the hippocampus and basolateral amygdala. J. Neurosci. 24 (37), 8161–8169.
Roozendaal, B., Okuda, S., Van der Zee, E.A., McGaugh, J.L.1, 2006. Glucocorticoid en-
hancement of memory requires arousal-induced noradrenergic activation in the ba-
solateral amygdala. Proc. N. Y. Acad. Sci. 103, 6741–6746.
Roozendaal, B., 2000. Glucocorticoids and the regulation of memory consolidation.
Psychoneuroendocrinology 25 (3), 213–238.
Schiller, D., Monfils, M.H., Raio, C.M., Johnson, D.C., LeDoux, J.E., Phelps, E.A., 2010.
Preventing the return of fear in humans using reconsolidation update mechanisms.
Nature 463 (7277), 49–53.
Schulz-Klaus, B., Lessmann, V., Endres, T., 2013. BDNF-dependent consolidation of fear
memories in the perirhinal cortex. Front. Behav. Neurosci. 7, 205.
Schwabe, L., Nader, K., Wolf, O.T., Beaudry, T., Pruessner, J.C., 2012. Neural signature of
reconsolidation impairments by propranolol in humans. Biol. Psychiatry 71 (4),
380–386.
Seeman, T.E., Singer, B., Wilkinson, C.W., McEwen, B., 2001. Gender differences in age-
related changes in HPA axis reactivity. Psychoneuroendocrinology 26 (3), 225–240.
Segal, S.K., Cotman, C.W., Cahill, L.F., 2012. Exercise-induced noradrenergic activation
enhances memory consolidation in both normal aging and patients with amnestic
mild cognitive impairment. J. Alzheimers Dis. 32 (4), 1011–1018.
Sevenster, D., Beckers, T., Kindt, M., 2014. Prediction error demarcates the transition
from retrieval, to reconsolidation, to new learning. Learn. Mem. 21 (11), 580–584.
Shields, G.S., Sazma, M.A., McCullough, A.M., Yonelinas, A.P., 2017. The effects of acute
stress on episodic memory: a meta-analysis and integrative review. Psychol. Bull. 143
(6), 636–675.
Siette, J., Reichelt, A.C., Westbrook, R.F., 2014. A bout of voluntary running enhances
context conditioned fear, its extincition and its reconsolidation. Learn. Mem. 21 (2),
73–81.
Small, L., Kenny, L., Bryant, R.A., 2011. The cost in remembering of ruminating on ne-
gative memories. Emotion 6, 1434–1438.
Storer, T.W., Davis, J.A., Caizozzo, V.J., 1990. Accurate prediction of VO2max in cycle
ergometry. Med. Sci. Sports Exerc. 22, 704–712.
Tollenaar, M.S., Elzinga, B.M., Spinhoven, P., Everaerd, W.A., 2008. The effects of cortisol
increase on long-term memory retrieval during and after acute psychosocial stress.
Acta Psychol. 127 (3), 542–552.
van Stegeren, A.H., Wolf, O.T., Everaerd, W., Scheltens, P., Barkhof, F., Rombouts, S.A.,
2007. Endogenous cortisol level interacts with noradrenergic activation in the human
amygdala. Neurobiol. Learn. Mem. 87 (1), 57–66.
van Stegeren, A.H., 2008. The role of the noradrenergic system in emotional memory.
Acta Psychol. 127, 532–541.
Wahl, P., Zinner, C., Achtzehn, S., Bloch, W., Mester, J., 2010. Effect of high- and low
intensity exercise and metabolic acidosis on levels of GH, IGF-I, IGFBP-3 and cortisol.
Growth Horm. IGF Res. 20 (5), 380–385.
Winter, B., Breitenstein, C., Mooren, F.C., Voelker, K., Fobker, M., Lechtermann, A.,
Krueger, K., Fromme, A., Korsukewitz, C., Floel, A., Knecht, S., 2007. High impact
running improves learning. Neurobiol. Learn. Mem. 87 (4), 597–609.