www.nature.com/scientificreports
OPEN
Received: 3 August 2018
Accepted: 28 June 2019
Published: xx xx xxxx
Caffeine intake modulates the
functioning of the attentional
networks depending on
consumption habits and acute
exercise demands
Florentino Huertas
1
, Esther Blasco1, Consuelo Moratal1 & Juan Lupiañez
2
Consume of stimulants (as caffeine) is very usual in different contexts where the performers have to
take quick and accurate decisions during physical effort. Decision-making processes are mediated by
the attentional networks. An experiment was carried out to examine the effect of caffeine intake on
attention (alerting, orienting, and executive control) as a function of consumption habit under two
physical exertion conditions (rest vs. aerobic exercise). Two groups of participants with different caffeine
consumption profiles (moderate consumers vs. low consumers) performed the Attention Network Test–
Interactions under four different conditions regarding activity (rest vs. exercise) and intake (caffeine vs.
placebo). Results showed that whereas exercise led to faster reaction times (RT) in all cases, caffeine
intake accelerated RT but only at rest and in moderate caffeine consumers. More importantly, caffeine
intake reduced the alertness effect in moderate consumers only at the rest condition. No interactions
between Intake and Activity were observed in the other attentional networks, with exercise reducing
orienting independently of caffeine intake, which suggests that physical exercise and caffeine are
different modulators of attention but can interact. Caffeine intake had differential effects on reaction
speed at rest and during physical exercise depending on the individual consumption habit. On the
basis of these finding it seems that mainly alertness is modulated differently by internal and external
“arousing” conditions.
At different environmental situations (as practicing sports, driving vehicles, surgery interventions…) people
must often perform physical activities and respond quickly to external challenges. Decision-making processes
are mediated by the three attentional networks (alerting, orienting, and executive control) defined by Petersen &
Posner1.
The executive control network is involved in situations that require planning, decision making, error detection, execution of novel responses, or overcoming habitual actions2. The orienting network is responsible for
allocating attention to a particular object or region of space in a voluntary-endogenous or reflexive-exogenous
way, enhancing its processing while ignoring irrelevant objects/locations3. Finally, the alerting network participates in the general activation of cortical and thalamic areas, thus preparing the perceptual-motor system for fast
reactions through changes in the norepinephrine system.
During different situations of daily living like working and sport tasks, these attentional networks have to
work in cooperation to maintain an efficient psychomotor performance during the physical-cognitive dual task
challenges inherent to those activities.
The relationship between exercise and two of the main core components of executive control (inhibitory processes and cognitive flexibility) has been widely studied by researchers. Inhibitory control refers to the ability to
attend to a relevant stimulus while ignoring the others that are not relevant to the goal or task at hand. Cognitive
1
Department of Physical Education & Sport Sciences, Catholic University of Valencia “San Vicente Mártir”, Valencia,
Spain. 2Department of Experimental Psychology, Mind, Brain, and Behavior Research Center (CIMCYC), University of
Granada, Granada, Spain. Correspondence and requests for materials should be addressed to F.H. (email: florentino.
huertas@ucv.es)
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
1
www.nature.com/scientificreports/
www.nature.com/scientificreports
flexibility refers to the ability to rapidly switch between different tasks. Previous research reported that the abovementioned component of executive control were improved (enhancing inhibitory processes4 and reducing switch
cost5) when participants were exercising under an aerobic workload. Yet, and according to the hypofrontality
hypothesis6, neuroelectric7 and behavioral8 findings showed impaired conflict resolution during exercise conditions as compared to rest conditions. In addition, other studies failed to show any differential effects of acute
aerobic exercise on participants’ performance9, thus adding more contradictory findings regarding the nature
and direction of this relationship. More importantly for the purposes of the present study, scarce studies have
analyzed the effect of concomitant physical effort on the three attentional networks assessed at the same time.
Faster RT during the exercise condition were reported by Chang and colleagues10 using the Attentional Network
test (ANT)11, but only on trials with incongruent flankers, thus leading to reduced interference10. Another study,
closer to the present one, using a version of the ANT named Attentional Network Test Interaction (ANT-I)12
showed no effect of exercise on executive control13.
Regarding spatial attentional orienting, in a series of studies Pesce and colleagues found that aerobic exercise
enhanced the flexibility in modulating the spatial extent of visual attentional focus at different ages and sport
modalities14–18. Regarding the shifting of spatial attention in the visual space, the ability we will analyze in our
study, previous findings have shown that the effect of exercise on cueing effects in visuospatial attention is modulated by gender19 and sport expertise20. However, to the best of our knowledge, few studies have explored the
effects of acute bouts of aerobic exercise on the deployment of exogenous visual spatial attention. Previous studies
found that an acute bout of aerobic exercise performed during or even immediately before a spatial orienting
task eliminated the typical spatial cueing effect21. Later studies established that spatial orienting in overt attentional capture was influenced by high intensity exercise, and depended on participants’ fitness level showing more
reduced attentional effects in low-fit participants than in high-fit participants22. However, contradictory results
were found when attentional networks were assessed simultaneously, with no effects of exercise on exogenous
spatial orienting10,13.
On the other hand, the modulation of alertness by acute exercise has been widely studied but also shows
mixed results. There are two types of alertness. Phasic extrinsic alertness is associated with an abrupt increase
of nonspecific activation when a warning cue is presented preceding the target (this is related to the orienting response23). By contrast, tonic alertness or vigilance refers to sustained activation over a period of time24.
Regarding phasic alertness, previous studies reported that participants’ accuracy was lower when the task was
performed under an intense physical exertion condition than at rest25. However, contradictory results emerge
again when attentional networks are assessed simultaneously using the ANT task. Neuroelectric findings have
showed larger P3 amplitude (a positive-going component associated to the amount of attentional resources allocated to environmental events), on alerting trials in the exercise condition10. However, another behavioral study
has found a reduction of the size of the phasic alerting effect during aerobic exercise as compared to the rest condition13. Regarding the tonic alertness (sustained attention), there is a broad consensus that moderate to intense
exercising enhance the state of tonic vigilance26, though poorer vigilance performance was observed under heavier exercising conditions27.
Previous research has shown that the effect of acute bout of exercise on attentional networks seems to be
moderated by different individual and contextual constraints (e.g., participant physical fitness, sport modality,
exercise intensity, or others as consume of ergogenic or stimulant substances). Therefore these variables need to
be considered in the study of the acute exercise-attention relationship.
During challenging environmental conditions, people often decide to consume stimulants to improve performance. Caffeine (1,3,7-trimethylxanthine) has become one of the most popularly used ergogenic substances
to improve both physical and cognitive performance28,29. However, though improvement in general attentional
performance induced by caffeine intake is well accepted, there is scarce and controversial evidence about its effect
on each attentional network and their interactions.
Regarding the effect of caffeine intake on executive control, caffeine has been shown to have a positive influence on the ability to switch attention between tasks (task switching) and anticipatory control processes30,31.
However caffeine consumption has failed to modulate other facets of inhibitory function like the ability to control
interference from distracting stimuli (using Ericksen Flanker Test) and the ability to suppress responses selectively (e.g., by using a Stroop task)32. Previous studies exploring the interactions between exercise and caffeine on
executive control reported a positive effect of caffeine intake after exercise33, during aerobic34 and intermittent
exercises35.
Few studies have been nevertheless conducted on the effect of caffeine on attentional orienting36,37. Such
studies, performed under resting conditions, have only shown marginal effects of caffeine intake on orienting
when low caffeine consumers were tested. However, controversial findings have been reported when interaction
between caffeine intake and orienting was studied manipulating exercise conditions. While some studies did
not find modulations of caffeine on object tracking or covert spatial orienting respectively after aerobic acute
exercise38,39, others showed positive effects on shifting attention after longer medium to high intensity aerobic
exercises33.
Considering alertness, systematic reviews have reported significant enhancements of various doses of caffeine on vigilance40,41. Concerning the interaction between caffeine and exercise on alertness, sustained vigilance
appears to be improved by caffeine intake after33 and during34 an acute exercise bout. However, caffeine specific
effects on phasic alertness during exercise hasn’t been studied yet.
Scarce studies have investigated the effects of caffeine on the functioning of the three attentional networks
concurrently. To our knowledge, Brunyé and colleagues are the only ones who have investigated the effects of caffeine intake on the three attentional networks simultaneously by means of the ANT task36,37. They have reported
that caffeine-induced physiological arousal amplifies global spatial processing, vigilance and executive control,
but these effects are at least partially driven by the administered dose and habitual caffeine consumption. These
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
2
www.nature.com/scientificreports/
www.nature.com/scientificreports
findings are in line with those showing that caffeine increases arousal on every type (low and habitual) of caffeine consumers42. However, most studies in this topic have involved caffeine–deprived habitual caffeine consumers, which makes it difficult to know whether the observed findings are due to the effects of caffeine intake
or the alleviation of caffeine withdrawal. On this vein, another study investigated acute effects of caffeine in both
habitual and non-habitual caffeine consumers observing a more positive effect of caffeine intake on consumers’
mood state while improving cognitive performance more in the non-consumers43. Thus, previous findings suggest that the effects of caffeine intake on attentional performance could be explained, not only by a withdrawal
reduction model, but also by other mechanisms involving the brain biochemistry and the effect of caffeine on
dopamine-rich areas of the brain.
Considering that in many daily life activities caffeine intake is combined with a physical effort, we have to
take into account the evidence from previous literature reviewing the effects of acute exercise on cognition from
a neurochemical perspective. A recent review by McMorris44 has shown that long duration, moderate intensity,
and heavy exercise generate excessive concentrations of catecholamines and cortisol inhibiting working memory.
Nevertheless, heavy exercise has a positive effect on long-term memory due to activation of β-adrenoreceptors
and increased exercise-induced brain-derived neurotrophic factor (BDNF) levels. Since it is widely known that
attentional functioning is also modulated by the exercise- induced physiological response facilitating or inhibiting
attentional performance, it would be worth to study the joint effects of physical exercise and caffeine as both take
part in similar mechanism underlying the brain biochemistry.
However, and importantly for our research, the exercise- and caffeine-induced changes on the attentional
networks are somewhat equivocal44, as described above, and therefore future research is necessary. Furthermore,
understanding the interplay between exercise, caffeine intake and consumption habits on the functioning of
the three attentional networks is crucial for adapting caffeine consumption to optimize its expected benefits.
Nevertheless, to the best of our knowledge, no study so far has examined this interaction. In order to fill this
gap, the goal of the present study was to explore the individual and interactive effects of caffeine intake and acute
physical exercise on the functioning of the attentional networks in the context of a multi-functional task (ANT-I
task12) under four different conditions resulting from the combination of activity (i.e., rest vs. sustained moderate
aerobic exercise) and intake (caffeine vs. placebo).
Following the above reviewed literature, behavioral and neurophysiological findings lead to argue that
the efficiency of the central nervous and cognitive function (reaction time and attention) is modulated by
exercise-induced44 and caffeine-induced37 neurobiochemical mechanisms regulating arousal. It is relevant to
highlight that caffeine consumption41 and exercise45 modulate the baseline arousal level (lower under rest and placebo conditions), mainly in participants more habituated to caffeine intake. Based on this evidence we expected
to observe an interaction between caffeine intake, consumption habit and exercise condition on cognitive performance. This interaction was expected for overall RT and alertness, as these variables seems to be affected by
opposite manipulations reducing arousal46–48. In our case, we expected faster RT (tonic alertness) and reduced
phasic alertness effects in the conditions in which the combination of exercise and caffeine intake would lead to
stronger activation. We also wanted to explore whether caffeine intake and exercise also affect the functioning
of the other attentional networks (orienting and executive control). Although some effects have been previously
reported in the above reviewed literature, no clear predictions were anticipated.
Method
Participants. Twenty-four male undergraduate sport sciences students (age range: 21–25 years; M = 22
years) were selected to participate in this study. We used the usual sample size with the ANTI task12,49, without
using a priori power analyses. Therefore, a sensitivity analysis was conducted using G*power50 which showed that
with our sample size (N = 24) and 48 repeated measures, the minimum effect size that could have been detected
for α = 0.5, and 1 − β = 0.80, for 2 groups, is f = 0.11 (minimum detectable effect).
Participants were assigned to one of two groups according to their habitual caffeine consumption based on
their responses to a self-report food survey. Following previous recommendations51 we used a between-group
design, allocating participants to one of two groups depending on their caffeine consumption habit and dispensing an individual caffeine dosage depending on participants’ body weight. According to participants’ estimated
daily caffeine intake52, they were assigned to either the low or moderate caffeine consumers group using a cut-off
point criterion (Low consumers ≤ 50 mg/day vs. Moderate consumers > 50 mg/day)43. Half of participants were
categorized as low caffeine consumers (M = 14 mg/day, range = 0–28 mg/day) and half were considered moderate
caffeine consumers (M = 101 mg/day, range = 57–228 mg/day). All participants were non-smokers who were in
good health and were not taking any medication. They all completed both an ethical clearance form and a statement regarding informed consent.
Procedure. Present research followed accepted ethical, scientific and medical standards and was conducted in
compliance with recognized international standards, in accordance with the revised version (2013) of the Helsinki
Declaration. Informed consent was obtained from all study participants. The study protocol was approved by the
Ethics Committee of the Catholic University of Valencia, Spain (UCV/2016-2017/02).
Participants visited the laboratory on five separate days with a minimum interval of 48 hours and a maximum
interval of 96 hours, approximately at the same time of the day. They were instructed to abstain from consuming
caffeine or any stimulant substance and to avoid physical exercise for 24 h prior to each session. On the first visit,
participants performed a graded submaximal test in order to determine their Lactate Threshold (LT). The next
four visits were experimental sessions in which participants had to complete the attentional tasks corresponding
to the Intake (caffeine vs. placebo) and Activity (rest vs. exercise) conditions, whose order was counterbalanced
across participants using a Latin square design. The experimental timeline is depicted in Fig. 1. At the end of the
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
3
www.nature.com/scientificreports/
www.nature.com/scientificreports
Figure 1. Schematic representation of the experimental protocol timing.
last experimental session, participants were debriefed on the purposes of the study and received a description of
their physical and cardiovascular fitness and attentional performance.
During cycling parameter estimation test and attentional evaluation we have employed the same attentional
tasks, and similar resources and procedures previously described by the authors in a similar research about the
effect of acute bouts of exercise on functioning of the attentional networks13.
Cycling parameter estimation test. On the first session, participants were informed about the details of the study
and gave written informed consent. After that, participants practiced a reduced version of the ANT-I. Next,
participants were fitted with a Polar RS800 heart rate monitor (Polar Electro Ltd., Kempele, Finland). A magnetically braked cycle ergometer (Cardgirus Medical, G&G Innovación, La Bastida, Alava, Spain) was adjusted
to accommodate them to perform a graded submaximal test according to previous recommendations53. After a
7-min warm-up, the test started at an initial workload of 100 W, with increments of 30 W every 4 min. The test was
terminated when participants acknowledged voluntary exhaustion, could not maintain the minimum cadence
of 60 revolutions per minute (rev · min−1) or when their heart rate (HR) reached 95% of their maximum HR
(HRmax) obtained from the age-predicted equation54: HRmax = 208 − 0.7 · Age. Finally, participants pedaled until
their HR was under 120 beats per minute (bpm) before getting off the cycle ergometer. Earlobe capillary blood
samples were collected in the last 15 s of each stage and were analyzed by a blood lactate test meter (Lactate Pro
LT-1710, Arkay KDK, Japan). Power output and HR were continuously monitored using a sampling rate of 1 Hz.
Determination of the LT was based on the criteria established by previous studies55. Workload at LT was defined
as the power output elicited at the stage before LT. HR at LT was estimated as the statistical mode value of HR at
the stage before LT. These results were used to set the individual exercise workloads in the exercising experimental
conditions.
Materials and procedure used in the experimental sessions. Prior to each experimental session and upon their
arrival in the lab, participants were weighed and given a capsule containing their previously assigned intake substance. We used a double-blind design to manipulate this variable. Caffeine (4 mg · kg −1 pure anhydrous powder)
and placebo (pregelatinized starch powder) were administered in identical color, size, weight, and shape capsules with ad libitum water. Dosage was selected based on a previous review51 which suggested doses from 3 to
6 mg · kg−1, and considering the recommendations by European Food Safety Authority56. Next, participants were
asked to take a 45-minute break to allow for sufficient plasma concentrations of caffeine after consumption57.
Participants completed the ANT-I task at approximately 60 cm from the computer monitor in a dimly-lit laboratory. A headphone set was used to deliver the acoustic alerting signal. Two computers were used simultaneously
in the experimental set-up. One was used to run E-Prime software58 presenting stimuli and collecting the participants’ responses during the attentional task in all experimental conditions. The other computer was used to run
Cardgirus software adjusting and collecting power output and HR data during exercise conditions. In the rest
condition, participants completed ANT-I while sitting on a chair.
Previous findings have shown that the efficiency of some brain areas and systems involved in cognitive arousal
and cognitive resources is related with physiological exercise workload-induced changes44,45. During the exercise
conditions, participants performed the attentional task while cycling on the ergometer at 80% of their workload
at LT. This intensity has been used in previous studies investigating the effect of acute exercise on attentional
networks13, in which physiological variables (e.g., HR blood flow and lactate) were maintained in a moderate
aerobic steady-state zone during the whole session. With the aim of keeping physiological response (HR) in a
steady-state zone, a 10-min incremental warm-up was used to reach the initial predetermined exercise workload.
Workload could be regulated lightly by the experimenter manually throughout the exercise sessions to maintain
each individuals’ HR in the target steady-state zone. To confirm that participants performed the cognitive task
under a steady state metabolic during both rest and exercise sessions, blood was drawn from their earlobes for
lactate analysis at the beginning and immediately after finishing the attentional task. Moreover, HR was monitored during the whole sessions.
®
Attentional task (ANT-I). The participants’ task was to respond as quickly and accurately as possible by pressing
the left or right key placed on the handlebar. Participants responded to the direction of the target stimulus (a central target arrow 0.55° long pointing either left or right), which was flanked by two identical to the target irrelevant
arrows on each side (0.06° away from each other). On each trial, an acoustic alerting tone (2000 Hz and 50 ms)
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
4
www.nature.com/scientificreports/
www.nature.com/scientificreports
Figure 2. Procedure of the ANT-I task.
and/or spatial orienting visual cue (an asterisk 0.6° long and 50 ms) preceded the target arrow. Participants were
strongly encouraged to keep their eyes fixed on the fixation point (variable duration of 400–1600 ms) throughout
the trial. The sequence of events for each trial is shown in Fig. 2. The interference variable was defined according
to the congruency of the direction of the flankers and target arrows: congruent trials (50% of trials), when the
target was flanked by arrows pointing in the same direction, and incongruent trials (the other 50% of trials), when
the flanking arrows and the target pointed in opposite directions. The orienting signal was presented in two thirds
of the trials above or below the fixation point. Three orienting conditions were thus established according to the
presence of the cue; cued location trials, when the cue was presented at the same location as the target; uncued
location trials, when the cue was presented at the opposite location to the target, or absence of cue, no-cue trials,
when the cue was not presented. The alerting signal was presented before the onset of the target in only half of the
trials. The alerting variable was established according to the presence (tone) or absence (no tone) of the alerting
sound. The target was presented until participants responded or for 1800 ms. After the response, or after the
maximum time had elapsed, the fixation point was presented for a variable duration (depending on the RT of the
preceding trial and the duration of the initial display for that trial) so that all trials were equally long. Initially, participants completed a practice block of 48 trials, followed by five experimental blocks of 48 trials each, with resting
intervals of about 1 min between them, but maintaining constant cycling parameters in the exercise sessions.
Thus, participants performed the ANT-I for approximately 25 min in each experimental session. The exercise
sessions required longer time (about 35 min) because the warm-up stage was necessary to reach the physiological
steady state to perform the 25 min of continuous cycling at the pre-established HR and exercise workload. Both
the workload intensities and the duration made it possible to obtain the required amount of data from the attentional task during exercise, maintaining a physiological steady state while avoiding fatigue in the participants.
Results regarding physiological response to each Activity and Intake condition confirm the high level of stability
of the variables used for measuring physiological response to exercise.
Design and statistical analysis. We conducted a mixed repeated-measures MANOVA on the data of each
dependent variable (HR, RT, and accuracy), with caffeine Consumption Habit (low, moderate) as a between-group
factor, and the other independent variables (Intake: caffeine, placebo; Activity: rest, exercise; Alerting: tone, no
tone; Orienting: uncued, no cue, cued; and Congruency: congruent, incongruent) as within- participant factors.
Subsequent ANOVAs were performed for mean RT and overall error percentages to analyze the general effects
of Consumption Habit, Intake, and Activity on each attentional function. Post hoc analyses (paired t-tests) were
conducted to further explore significant interactions.
Results
Preliminary analysis. Descriptive statistics of anthropometric and physiological characteristics of the whole
sample and separately for low and moderate caffeine consumers are displayed in Table 1. No significant differences were observed between groups in none of these variables (all ps > 0.2).
The Mean (M) and Standard Deviation (SD) of HR and lactate values per experimental session are displayed
in Table 2.
In order to confirm whether the applied exercise workload and caffeine intake induced different changes in
physiological state in each intake group of participants, a mixed 2 (Habit) × 2 (Intake) × 2 (Activity) MANOVA
was performed on participants’ average HR values, with Habit as a between-group factor. Results showed a significant main effect of Activity, F(1, 22) = 4158.19, p < 0.001, η2 = 0.99, indicating that HR increased by 56 bpm
from the rest to the exercise condition. However, caffeine intake did not modulate HR (p = 0.321). Similarly,
Consumption Habit did not modulate HR response under either Activity (p = 0.238) or Intake (p = 0.841)
conditions.
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
5
www.nature.com/scientificreports
www.nature.com/scientificreports/
Mean ± SD
Low
consumers
Moderate
consumers
Age (years)
21.9 ± 1.9
22.9 ± 3.9
Weight (kg)
74.6 ± 6.5
72.9 ± 11.1
Height (cm)
178.2 ± 4.6
Variables
HR baseline (bpm)
178.8 ± 5.7
All
22.4 ± 3.0
73.7 ± 9.0
178.5 ± 5.1
64 ± 6
67 ± 8
66 ± 7
Predicted HRmax (bpm)
177 ± 3
183 ± 2
180 ± 2
HR at LT (bpm)
150 ± 13
155 ± 9
155 ± 11
LT (mmol · L−1)
3.6 ± 0.8
3.6 ± 0.7
3,6 ± 0.8
173.6 ± 28.0
160.0 ± 22.1
166.5 ± 25.5
Workload at LT (W)
Table 1. Anthropometric and physiological characteristics according to participant group. Note. SD = standard
deviation; HR = heart rate; LT = lactate threshold; bpm = beats per minute; W = watts; mmol · L−1: millimoles
per liter; HRmax = maximum predicted heart rate.
Attentional networks functioning. Incorrect responses (1.63%), and those with RT faster than 200 ms
(considered as anticipations) or slower than 3 SD above the overall mean (1.28%; considered as lapses) were
discarded from the analysis of RT. Mean RTs were computed with the remaining RTs and submitted to the corresponding analyses. Mean RTs and error percentages per experimental condition are displayed in Table 3.
A first analysis was carried out on the Mean RTs and Error percentages to test the general functioning of the
ANTI task in this population. The 2 (Alerting) × 3 (Orienting) × 2 (Congruency) repeated measures ANOVA
performed on Mean RTs showed the usual results, i.e., main effects for each variable, F(1, 23) = 55.41, p < 0.001,
η2 = 0.71, F(2, 46) = 165.36, p < 0.001, η2 = 0.88, and F(1, 23) = 628.54, p < 0.001, η2 = 0.96 (for Alerting,
Orienting and Congruency respectively), and the usual Alerting × Orienting, F(2, 46) = 52.97, p < 0.001,
η2 = 0.70 (larger cueing effect with the alerting signal), Alerting × Congruency, F(1, 23) = 25.33, p < 0.001,
η2 = 0.52 (larger flanker-congruency effect with the alerting signal) and Orienting × Congruency interactions,
F(2, 46) = 35.86, p < 0.001, η2 = 0.61 (smaller flanker-congruency effect for cued than uncued trials).
The corresponding analysis performed on the error percentages showed similar results with F(1, 23) = 3.98,
p < 0.058, η2 = 0.15, F(2, 46) = 11.96, p < 0.001, η2 = 0.34, and F(1, 23) = 32.57, p < 0.001, η2 = 0.59 for the
main effects of Alerting, Orienting and Congruency respectively. Similarly, the usual Alerting × Congruency
and Orienting × Congruency interactions were also significant, F(1, 23) = 7.10, p < 0.014, η2 = 0.24, and F(2,
46) = 13.06, p < 0.001, η2 = 0.36, respectively (the Alerting × Orienting did not reach statistical significance, F(2,
46) = 1.92, p < 0.158, η2 = 0.08).
Importantly, when Consumption habit, Intake, and Activity were added to the analysis, these variables modulated the main effect of some of the attentional variables, but no interaction between attentional variables was
modulated. Therefore, in order to have a closer look to the data, 4 different ANOVAs were performed, one for
Mean RT (and overall error percentages) to analyze the general effects of Consumption habit, Intake, and Activity,
and one for each attentional network to analyze how they affect attentional performance.
Overall effects of Consumption habit, Intake, and Activity. A 2 (Consumption habit) × 2 (Intake) × 2
(Activity) mixed ANOVA was performed on Mean RT and percentage of errors, with Consumption habit as a
between-participants factor and the other variables as within participants variables. RT results revealed a significant main effect of Activity, F(1, 22) = 91.70, p < 0.001, η2 = 0.81, with 50 ms faster RT in the exercise than the
rest condition. Although Intake did not modulate significantly RT, both the Activity × Intake and the three-way
Activity × Intake × Consumption habit interactions were significant, F(1, 22) = 4.86, p = 0.038, η2 = 0.18, and
F(1, 22) = 7.74, p = 0.011, η2 = 0.26, respectively. Partial ANOVAs showed that the effect of Intake was only
observed at rest in moderate caffeine consumers, F(1, 11) = 16.68, p = 0.002, η2 = 0.60 (F < 1 in the remaining
cases). As can be observed in Fig. 3, whereas exercise led to faster RT in all cases, caffeine intake led to 37 ms faster
RT but only at rest and in moderate caffeine consumers.
The corresponding analysis on the error percentages only revealed a main effect of Activity, F(1, 22) = 19.65,
p < 0.001, η2 = 0.47, showing that participants made slightly more errors (0.73%) in the exercise than the rest
condition.
Alertness. A 2 (Consumption habit) × 2 (Intake) × 2 (Activity) × 2 (Alertness; No Tone vs. Tone) mixed
MANOVA was performed on Mean RT. Following previous studies12, only data from the no cue condition was
considered for this analysis. A main effect of Alertness was observed, F(1, 22) = 109.75, p < 0.001, η2 = 0.83, with
31 ms faster RT for the Tone than the No Tone condition. More importantly, this alertness effect was modulated
by both Intake and Activity, F(1, 22) = 4.91, p = 0.037, η2 = 0.18, and F(1, 22) = 4.99, p = 0.036, η2 = 0.18, respectively. The four-way interaction was also significant, F(1, 22) = 5.13, p = 0.034, η2 = 0.19, showing a marginal
reduction of alertness effect with caffeine intake (compared with no intake), in low consumers at the exercise
condition (see top panel of Fig. 4), F(1, 11) = 4.47, p = 0.058, η2 = 0.29, but a significant reduction in moderate
consumers at the rest condition (see bottom panel of Fig. 4), F(1, 11) = 5.15, p = 0.044, η2 = 0.32.
The corresponding analysis of the error percentages only revealed as significant the Intake × Alertness interaction, F(1, 22) = 5.74, p = 0.026, η2 = 0.21. Whereas after caffeine intake slightly more errors (0.61%) were made
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
6
www.nature.com/scientificreports
www.nature.com/scientificreports/
Rest
Exercise
Caffeine
Placebo
Caffeine
Placebo
Variables
Low
Moderate
All
Low
Moderate
All
Low
Moderate
All
Low
Moderate
All
Workload (W)
—
—
—
—
—
—
106.3 ± 19.3
104.7 ± 19.6
105.5 ± 19.0
106.7 ± 23.6
101.5 ± 18.1
104.0 ± 20.6
Initial [La−]
(mmol · L−1)
—
—
—
—
—
—
2.5 ± 0.9
2.4 ± 1.1
2.4 ± 1.0
2.2 ± 0.7
2.3 ± 0.8
2.3 ± 0.7
Final [La−]
(mmol · L−1)
—
—
—
—
—
—
2.0 ± 0.9
2.4 ± 0.8
2.2 ± 0.8
1.9 ± 0.8
2.2 ± 0.7
2.1 ± 0.7
HR (bpm)
67 ± 8
70 ± 11
68 ± 9
64 ± 6
67 ± 8
65 ± 7
120 ± 9
126 ± 8
122 ± 8
121 ± 7
126 ± 9
123 ± 8
% predicted
HRmax
38.5 ± 5.2
38.2 ± 6.0
38.3 ± 5.6
36.2 ± 4.1
37.7 ± 5.4
36.5 ± 4.0
68.6 ± 4.0
69.2 ± 3.7
68.7 ± 4.5
68.6 ± 4.8
69.3 ± 3.3
68.8 ± 4.1
% HR at LT
44.1 ± 5.2
45.5 ± 4.0
44.7 ± 5.3
43.6 ± 5.4
43.8 ± 4.6
43.7 ± 5.3
80.2 ± 4.0
85.3 ± 4.3
82.3 ± 4.1
80.3 ± 4.1
85.2 ± 4.3
82.6 ± 4.1
Table 2. Mean (M) ± SD values of workload and physiological response values to each Activity and Intake
experimental condition according to participants’ caffeine consumption habit. Note. Note that under Rest
condition the Workload and Lactate values were -, because no work was developed and the instrument used
to measure Lactate concentration does not give values below 0.8. SD = standard deviation; HR = heart rate;
HRmax = maximum predicted heart rate; LT = lactate threshold; bpm = beats per minute; W = watts; mmol · L−1:
millimoles per liter.
Low Consumers
Moderate Consumers
No Tone
Congruent
Caffeine Intake
Incongruent
Exercise
Congruent
Placebo Intake
Incongruent
Congruent
Caffeine Intake
Incongruent
Rest
Congruent
Placebo Intake
Incongruent
Tone
No Tone
Tone
Uncued
No Cue
Cued
Uncued
No Cue
Cued
Uncued
No Cue
Cued
Uncued
No Cue
Cued
451
467
428
445
427
422
455
470
440
441
429
414
(0.00%)
(0.69%)
(0.17%)
(0.35%)
(0.69%)
(0.00%)
(0.00%)
(0.35%)
(0.00%)
(0.35%)
(0.35%)
(0.00%)
520
490
480
533
497
474
532
530
505
540
508
482
(6.42%)
(1.39%)
(1.74%)
(5.90%)
(2.78%)
(2.43%)
(5.56%)
(2.08%)
(1.04%)
(5.56%)
(2.43%)
(1.04%)
445
464
426
446
429
410
456
468
432
441
432
414
(0.69%)
(0.69%)
(0.69%)
(0.52%)
(0.69%)
(0.17%)
(1.04%)
(0.69%)
(0.00%)
(0.35%)
(0.35%)
(0.00%)
524
517
495
535
495
481
540
523
502
532
500
481
(3.30%)
(3.30%)
(2.26%)
(6.77%)
(1.91%)
(4.34%)
(2.43%)
(4.51%)
(2.43%)
(7.29%)
(2.43%)
(3.13%)
505
509
460
496
472
456
496
500
468
484
462
447
(0.17%)
(0.52%)
(0.00%)
(0.35%)
(0.00%)
(0.00%)
(0.00%)
(0.69%)
(0.00%)
(0.35%)
(0.00%)
(0.00%)
581
559
520
585
542
510
586
565
530
591
546
510
(3.30%)
(0.87%)
(0.52%)
(5.03%)
(2.43%)
(0.87%)
(2.43%)
(1.04%)
(0.35%)
(5.21%)
(2.43%)
(1.04%)
493
510
466
488
473
456
536
547
508
520
490
486
(0.17%)
(0.17%)
(0.17%)
(0.00%)
(0.35%)
(0.35%)
(0.00%)
(0.00%)
(0.00%)
(0.00%)
(0.00%)
(0.35%)
576
555
528
580
536
511
631
620
565
615
567
544
(2.26%)
(2.26%)
(1.39%)
(4.51%)
(2.26%)
(2.43%)
(1.74%)
(2.78%)
(1.04%)
(2.78%)
(2.43%)
(2.43%)
Table 3. Mean Reaction Time and error percentage (in parenthesis) for each experimental condition and for
each consumption habit group.
Figure 3. Mean RT as a function of Activity and Caffeine intake for each group of Consumption habit. Note:
***p < 0.005.
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
7
www.nature.com/scientificreports/
www.nature.com/scientificreports
Figure 4. Attentional index of the functioning of each attentional network, as a function of Activity, Caffeine
intake and Consum Habit (top panel: Low caffeine consumers; bottom panel: Moderate caffeine consumers).
Alertness index = RT No Tone – RT Tone trials (exclusively on No cue trials). Orienting index = RT Uncued –
RT Cued trials. Control index (interference) = RT Incongruent – RT Congruent trials. Note: *p < 0.05.
in the presence of the tone than when it was absent, less errors (0.30%) were made when the tone was present in
the placebo intake condition.
Orienting. A 2 (Consumption habit) × 2 (Intake) × 2 (Activity) × 2 (Orienting; uncued, no cue, cued) mixed
MANOVA was performed on Mean RT. The expected main effect of Orienting was highly significant, F(2,
44) = 153.27, p < 0.001, η2 = 0.87. More interestingly, Orienting was modulated by Activity, F(2, 44) = 13.48,
p < 0.001, η2 = 0.38, leading to reduced orienting effects in the exercise condition in both groups, as shown in
Fig. 4 (top and bottom panels). Both the costs (uncued minus no cue), F(1, 22) = 6.46, p < 0.019, η2 = 0.23, and
the benefits (no cue minus cued), F(1, 22) = 10.78, p = 0.003, η2 = 0.33, were reduced in the exercise (30 and 19,
respectively) compared to the rest (22 and 11 ms, respectively) condition.
The corresponding analysis or error percentages also showed a significant mean effect of orienting, F(2,
44) = 12.01, p < 0.001, η2 = 0.35. Additionally, a significant Intake × Orienting interaction was found observing
larger orienting effect in the caffeine intake (1.98%) than in the placebo intake (0.80%) condition, F(2, 44) = 3.39,
p = 0.043, η2 = 0.13.
Control. A 2 (Consumption habit) × 2 (Intake) × 2 (Activity) × 2 (Congruency; Congruent vs. Incongruent)
mixed MANOVA was performed on Mean RT. The main effect of congruency was highly significant, F(1,
22) = 672.79, p < 0.001, η2 = 0.97, with 70 ms faster RT for the Congruent than the Incongruent condition. This
effect was not modulated by any variable.
The analogous analysis of error percentages revealed a significant main effect of congruency, F(1, 22) = 31.22,
p < 0.001, η2 = 0.59. Interestingly, this effect was modulated by Activity, F(1, 22) = 8.31, p = 0.009, η2 = 0.27, with
larger interference being observed in the exercise (3.10%) than in the rest (2.16%) condition.
Figure 4 depicts the attentional index of the functioning of each attentional network, as a function of Activity,
Caffeine intake and consumer habits.
Discussion
The present study is the first to measure the effect of caffeine on attention considering its multifunctional nature
(alerting, orienting, and executive control) under two different physical effort conditions and using two groups of
participants selected according to their caffeine consumption habit. Attentional functioning was assessed under
caffeine and placebo intake conditions at rest and during moderate aerobic steady-state exercise.
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
8
www.nature.com/scientificreports/
www.nature.com/scientificreports
Previous studies attempting to explore the interactions between exercise and caffeine and their effects on cognition have shown controversial results due to methodological issues: different caffeine dosage and participant’s
consumption habit and withdrawal40, or moment of cognitive evaluation (before, during or after completion of
physical effort34,59). Importantly, and considering the recommendations by previous authors, here we tried to
control the modulation of different variables that could affect the caffeine effect (individual caffeine dosage and
participant’s consumption habits) during exercising.
Our results replicated the general facilitating effect of exercise on speed of response, previously described
by similar studies13,21, probably related to a shift in the speed-accuracy trade-off toward response speed44. More
importantly for the aims of the present study, our results have shown rather specific effects of caffeine intake
on both reaction speed and the efficiency of the alerting network. Caffeine intake reduced the effect of alerting
signals on response speed in low caffeine consumers when they were exercising. Contrarily, for more habitual
caffeine consumers caffeine intake only reduced alertness at rest.
Regarding exercise modulations of RT, our results indicated that the applied exercise workload induced a
significant enhancement of physiological state (HR and lactate accumulation) in both low and moderate caffeine consumers and, more importantly, that exercise shortened response speed in all experimental conditions,
while reduced only slightly the response accuracy. In line with previous accounts, our results replicated the finding about the benefit on response speed during aerobic exercise (below the LT)5,13,21,60,61. These results could be
partially explained by the arousal theory and increases of adrenaline-noradrenaline secretion induced by exercise45. McMorris44 has recently described a more comprehensive framework to explain how acute exercise affects
cognition from a neurochemical perspective. The author reviewed the litarature showing that intrinsic exercise
variables (as duration or intensity) modulate the secretion of different neurochemical factors (catecholamines,
cortisol and BDNF). In our study, the slight reduction of response accuracy in the flankers task is in agreement
with previous studies10,44,62, pointing to the stress level induced by moderate aerobic exercise, which mildly modulates in the sympathetic-adrenal and vagal/nucleus tractus solitarii (NTS) pathway to impair accuracy compared
to higher intensities.
Concerning caffeine intake, and similarly to findings included in a recent meta-analysis63, our results did
not show any main effect of caffeine on HR, lactate accumulation and RT during exercise in any of the controlled experimental conditions. These results are partially incongruent with the resounding evidence of significantly faster RT under caffeine intake conditions due to increase of adrenaline-noradrenaline secretion64 using
a broad variety of methodological approaches. However, most of these studies underscored the importance of
exploring these effects in a range of participant consumption profiles and under different exercising conditions.
Importantly, and according to our expectations, the effect of caffeine intake on performance was moderated by
both variables (consum habit and exercise), observing that caffeine shortened RT only in moderate caffeine consumers at rest. Yet, importantly, moderate consumers were significantly slower under the placebo-rest condition
than low consumers. Our results could be explained by the onset of a withdrawal effect associated with caffeine
deprivation in participants more habituated to caffeine consumption (moderate consumers) in conditions where
tonic arousal was lower (i.e., the placebo and rest condition), but they equate to non consumers when external arousing stimuli appear (i.e., exercising, caffeine intake or alerting signals). These finding are in agreement
with the poorer performance observed under caffeine deprivation of habitual coffee consumers65,66. Increases of
adrenaline-noradrenaline secretion induced by exercise45,67 and caffeine64 could justify the greater effect of caffeine on RT under the rest condition, when baseline arousal level is lower.
According to caffeine and exercise modulations of attentional networks, our results replicated the typical
pattern of results about the general functioning of the ANTI task in this population (main effects and interactions between attentional networks) on mean RTs and error both at rest12 and during exercise conditions13. More
importantly for the purpose of the current study, our results revealed that the effect of caffeine on phasic alertness
was moderated by the consum habit and exercise condition simultaneously. Caffeine intake reduced the effect of
alerting signals on RT in low consumers when they were exercising, but contrarily, caffeine induced a lower alertness effect in moderate consumers at the rest condition. In line with previous accounts using a similar attentional
tasks, we observed a reduction of alerting acoustic signals effect under exercising13,25. This interaction could be
expected by taking into account that caffeine interacts with the dopaminergic systems64, and that the alertness
attentional network activates dense dopaminergic innervated areas68. However, and contrarily to Brunyé et al.
using the ANT, who observed a lower alerting size effect under placebo compared to both the 200 mg and 400 mg
caffeine intake conditions in low consumers37, our data showed a larger alerting effect under placebo than under
caffeine intake conditions. Interestingly, the differences between our results and those obtained by Brunyé and
colleagues36,37 could be related to the different measure of the alerting function as assessed with the Attentional
Network Test (ANT) vs. the ANT-I used here. The ANT uses a less arousing visual cue to measure alertness,
whereas the ANT-I uses an auditory cue, which seems to be more effective in producing alertness69. Thus, a more
arousing alerting cue in our study produced larger alertness in the reduced tonic activation condition (placebo
intake), whereas a less arousing cue lead to larger alertness only when added by cafeine intake arousal. This idea
is supported by previous findings36 suggesting that, given that caffeine upregulates dopaminergic availability,
more habituated caffeine consumers may need higher doses to reach a similar “baseline” state of vigilance than
less habitual consumers. Taken altogether, our results seem to suggest that higher phasic alertness (due to slower
responses in no tone conditions) is usually observed under conditions of reduced tonic vigilance, as in rest and
placebo conditions in the present study or morning-type individuals when tested in the evening46. According
to the observed interaction between all “arousing” variables (alerting stimulus, activity and caffeine intake), our
results are coherent with the hypothesis whereby the locus coeruleus–norepinephrine system is activated in a
similar manner by these variables70.
Concerning the orienting network, results revealed that, compared to the rest condition, only exercise modulated this function inducing a decrease in spatial attentional orienting, both in attentional benefits and attentional
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
9
www.nature.com/scientificreports/
www.nature.com/scientificreports
costs. Our results are partially consistent with others showing a reduction of the spatial orienting effect21, and
revealing a deployment of exogenous spatial attention after a bout of intense aerobic exercise22. Present results
may indicate that exercise-induced physiological changes activate the attentional system, making it more reactive
to relevant peripheral stimuli, and therefore being less affected by irrelevant cues. However, our finding is somewhat in contrast with other study using the ANT-I task during an aerobic exercise in a group of highly skilled
cyclists and did not observe an interaction between exercising and exogenous visual spatial attention13. These
controversial results could be explained by previous findings showing that exercising affect spatial visual orienting
differently depending on the participant’s physical fitness level22 and motor expertise71. Further research investigating the exercise-attentional networks functioning interactions should consider the role of both moderators
(physical fitness and/ or motor skills).
Regarding the effect of caffeine on the orienting function, we replicated previous results observed in high consumers by Brunyé and colleagues37 showing that caffeine did not modulate the orienting network function when
response speed was considered, although they found a marginal effect of caffeine on orienting in low consumers
when using larger doses than the ones used in the present study. Overall, these results support the idea that
caffeine-induced changes in orienting may depend on the caffeine doses used and individuals’ consumption habit.
The absence of a significant interaction between caffeine and visual orienting has been explained by arguing that
the brain areas involved in attentional orienting (superior parietal lobe brain areas)72 have sparse dopaminergic
innervation, while dopamine is the neurotransmitter affected by caffeine.
In regard to the executive control network our results are in line with previous findings9,13 showing no evidence of any modulation of executive control performance by exercise. However, regarding the effect of caffeine
on executive functioning, our results challenge those previously described by Brunyé and colleagues demonstrating that the group of low-caffeine consumers exhibited dose-dependent increases in the executive control
function (at doses from 0 to 400 mg)36,37. Yet, the daily caffeine intake of participants almost doubled the intake
of the low consumers in our study.
Conclusions
Our results support the assumption that exercise speeds up response ability by modulating exogenous
stimulus-driven attentional functions (i.e., alerting and exogenous spatial orienting). These findings could be
explained in terms of exercise-induced increases in physiological arousal and consequently a better efficiency
of peripheral motor processes during exercise conditions. Nevertheless, this hypothesis should be verified by
further research, for instance, by exploring the relationship between electromyographic response and attentional
networks functioning during exercise.
Importantly, our results show that caffeine and exercise have a similar effect on performance, speeding up
motor responses, but this effect is moderated by the caffeine consume habit. In this regard, habitual caffeine intake
induces a reduction of tonic arousal, which turns comparable to the level in non-consumers when other arousing
stimulus appear (exercise, caffeine intake or alerting signals).
Given the interaction between these two variables (activity and caffeine), we can assume that they both
increase general activation in a similar way. However, the absence of interactions between Intake and Activity
with regard to orienting and executive control suggests that exercise and caffeine are different modulators of
attention. Only in the case of the modulation over the general activation indexed by alertness these two variables
seems to interact. It seems that each attentional network, depending on its behavioral significance, is modulated
differently by internal and external “arousing” conditions.
Overall, the results of this study suggest that caffeine intake seems to be necessary in more habituated caffeine consumers to reach an optimal level of general arousal and respond in low “arousing” situations (i.e., rest
conditions). Caffeine deprivation periods in more habitual consumers may impair RT performance due to lower
general activation at baseline related to dopamine availability conditions.
The systematic control and administration of acute and habitual personalized caffeine doses is relevant in
order to optimize the balance between physiological and cognitive cost and benefits. These recommendations
could be extended to different contexts where caffeine consumption is habitual to cope with the high demands
of vigilance and accurate decision making during prolonged time, as in professional drivers, pilots or air traffic
controllers.
Data Availability
The datasets generated during and/or analysed during the current study are available in the OSFHOME repository, DOI 10.17605/OSF.IO/VZ6NH, https://osf.io/vz6nh/?view_only=a7ee034bb2cf47ba8f4b3f0bf6a139e9.
References
1.
2.
3.
4.
5.
6.
7.
8.
SCIENTIFIC REPORTS |
Petersen, S. E. & Posner, M. I. The attention system of the human brain: 20 years after. Annu. Rev. Neurosci. 35, 73–89 (2012).
Miller, E. K. The prefrontal cortex and cognitive control. Nat. Rev. Neurosci. 1, 59–65 (2000).
Posner, M. I. Orienting of attention. Q. J. Exp. Psychol. 32, 3–25 (1980).
Audiffren, M., Tomporowski, P. D. & Zagrodnik, J. Acute aerobic exercise and information processing: modulation of executive
control in a Random Number Generation task. Acta Psychol. (Amst.) 132, 85–95 (2009).
Pesce, C. & Audiffren, M. Does acute exercise switch off switch costs? A study with younger and older athletes. J. Sport Exerc.
Psychol. 33, 609–626 (2011).
Dietrich, A. Transient hypofrontality as a mechanism for the psychological effects of exercise. Psychiatry Res. 145, 79–83 (2006).
Pontifex, M. B. & Hillman, C. H. Neuroelectric and behavioral indices of interference control during acute cycling. Clin.
Neurophysiol. 118, 570–580 (2007).
Del Giorno, J. M., Hall, E. E., O’Leary, K. C., Bixby, W. R. & Miller, P. C. Cognitive function during acute exercise: a test of the
transient hypofrontality theory. J. Sport Exerc. Psychol. 32, 312–323 (2010).
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
10
www.nature.com/scientificreports/
www.nature.com/scientificreports
9. Davranche, K., Hall, B. & McMorris, T. Effect of acute exercise on cognitive control required during an Eriksen flanker task. J. Sport
Exerc. Psychol. 31, 628–639 (2009).
10. Chang, Y.-K., Pesce, C., Chiang, Y.-T., Kuo, C.-Y. & Fong, D.-Y. Antecedent acute cycling exercise affects attention control: an ERP
study using attention network test. Front. Hum. Neurosci. 9, 156 (2015).
11. Posner, M. I. & Petersen, S. E. The attention system of the human brain. Annu. Rev. Neurosci. 13, 25–42 (1990).
12. Callejas, A., Lupianez, J. & Tudela, P. The three attentional networks: On their independence and interactions. Brain Cogn. 54,
225–227 (2004).
13. Huertas, F., Zahonero, J., Sanabria, D. & Lupiáñez, J. Functioning of the attentional networks at rest vs. during acute bouts of aerobic
exercise. J. Sport Exerc. Psychol. 33, 649–665 (2011).
14. Pesce, C., Capranica, L., Tessitore, A. & Figura, F. Focusing of visual attention under submaximal physical load. Int. J. Sport Exerc.
Psychol. 1, 275–292 (2003).
15. Pesce, C., Cereatti, L., Casella, R., Baldari, C. & Capranica, L. Preservation of visual attention in older expert orienteers at rest and
under physical effort. J. Sport Exerc. Psychol. 29, 78–99 (2007).
16. Pesce, C., Cereatti, L., Forte, R., Crova, C. & Casella, R. Acute and chronic exercise effects on attentional control in older road
cyclists. Gerontology 57, 121–128 (2011).
17. Pesce, C., Tessitore, A., Casella, R., Pirritano, M. & Capranica, L. Focusing of visual attention at rest and during physical exercise in
soccer players. J. Sports Sci. 25, 1259–1270 (2007).
18. Condello, G. et al. Steps to health in cognitive aging: Effects of physical activity on spatial attention and executive control in the
elderly. Front. Hum. Neurosci. 11 (2017).
19. Pesce, C., Casella, R. & Caprancia, L. Modulation of visuospatial attention at rest and during physical exercise: Gender differences.
Int. J. Sport Psychol. 35, 328–341 (2004).
20. Cereatti, L., Casella, R., Manganelli, M. & Pesce, C. Visual attention in adolescents: Facilitating effects of sport expertise and acute
physical exercise. Psychol. Sport Exerc. 10, 136–145 (2009).
21. Sanabria, D. et al. Effects of acute aerobic exercise on exogenous spatial attention. Psychol. Sport Exerc. 12, 570–574 (2011).
22. Llorens, F., Sanabria, D. & Huertas, F. The influence of acute intense exercise on exogenous spatial attention depends on physical
fitness level. Exp. Psychol. 62, 20–29 (2015).
23. Oken, B. S., Salinsky, M. C. & Elsas, S. M. Vigilance, alertness, or sustained attention: physiological basis and measurement. Clin.
Neurophysiol. 117, 1885–1901 (2006).
24. Posner, M. I. Measuring alertness. Ann. N. Y. Acad. Sci. 1129, 193–199 (2008).
25. Mahoney, C. R., Hirsch, E., Hasselquist, L., Lesher, L. L. & Lieberman, H. R. The effects of movement and physical exertion on
soldier vigilance. Aviat. Space Environ. Med. 78, B51–B57 (2007).
26. Smit, A. S., Eling, P. A. T. M., Hopman, M. T. & Coenen, A. M. L. Mental and physical effort affect vigilance differently. Int. J.
Psychophysiol. 57, 211–217 (2005).
27. Eddy, M. D. et al. The effects of load carriage and physical fatigue on cognitive performance. PloS One 10, e0130817 (2015).
28. Astorino, T. A. & Roberson, D. W. Efficacy of acute caffeine ingestion for short-term high-intensity exercise performance: a
systematic review. J. Strength Cond. Res. 24, 257–265 (2010).
29. McLellan, T. M., Caldwell, J. A. & Lieberman, H. R. A review of caffeine’s effects on cognitive, physical and occupational
performance. Neurosci. Biobehav. Rev. 71, 294–312 (2016).
30. Tieges, Z. et al. Caffeine improves anticipatory processes in task switching. Biol. Psychol. 73, 101–113 (2006).
31. Tieges, Z., Snel, J., Kok, A., Plat, N. & Ridderinkhof, R. Effects of caffeine on anticipatory control processes: evidence from a cued
task-switch paradigm. Psychophysiology 44, 561–578 (2007).
32. Tieges, Z., Snel, J., Kok, A. & Richard Ridderinkhof, K. Caffeine does not modulate inhibitory control. Brain Cogn. 69, 316–327
(2009).
33. Hogervorst, E., Riedel, W., Kovacs, E., Brouns, F. & Jolles, J. Caffeine improves cognitive performance after strenuous physical
exercise. Int. J. Sports Med. 20, 354–361 (1999).
34. Hogervorst, E. et al. Caffeine improves physical and cognitive performance during exhaustive exercise. Med. Sci. Sports Exerc. 40,
1841–1851 (2008).
35. Ali, A. et al. Caffeine ingestion enhances perceptual responses during intermittent exercise in female team-game players. J. Sports
Sci. 34, 330–341 (2016).
36. Brunyé, T. T., Mahoney, C. R., Lieberman, H. R., Giles, G. E. & Taylor, H. A. Acute caffeine consumption enhances the executive
control of visual attention in habitual consumers. Brain Cogn. 74, 186–192 (2010).
37. Brunyé, T. T., Mahoney, C. R., Lieberman, H. R. & Taylor, H. A. Caffeine modulates attention network function. Brain Cogn. 72,
181–188 (2010).
38. Church, D. D. et al. The effect of an acute ingestion of Turkish coffee on reaction time and time trial performance. J. Int. Soc. Sports
Nutr. 12, 37 (2015).
39. Connell, C. J. W., Thompson, B., Kuhn, G. & Gant, N. Exercise-induced fatigue and caffeine supplementation affect psychomotor
performance but not covert visuo-spatial attention. PloS One 11, e0165318 (2016).
40. Einöther, S. J. L. & Giesbrecht, T. Caffeine as an attention enhancer: reviewing existing assumptions. Psychopharmacology (Berl.) 225,
251–274 (2013).
41. Snel, J. & Lorist, M. M. Effects of caffeine on sleep and cognition. Prog. Brain Res. 190, 105–117 (2011).
42. Childs, E. & De Wit, H. Subjective, behavioral, and physiological effects of acute caffeine in light, nondependent caffeine users.
Psychopharmacology (Berl.) 185, 514–523 (2006).
43. Haskell, C. F., Kennedy, D. O., Wesnes, K. A. & Scholey, A. B. Cognitive and mood improvements of caffeine in habitual consumers
and habitual non-consumers of caffeine. Psychopharmacology (Berl.) 179, 813–825 (2005).
44. McMorris, T. Exercise-cognition interaction: Neuroscience perspectives. (Elsevier Academic Press, 2016).
45. Chmura, J., Krysztofiak, H., Ziemba, A. W., Nazar, K. & Kaciuba-Usciko, H. Psychomotor performance during prolonged exercise
above and below the blood lactate threshold. Eur. J. Appl. Physiol. 77, 77–80 (1998).
46. Matchock, R. L. & Mordkoff, J. T. Chronotype and time-of-day influences on the alerting, orienting, and executive components of
attention. Exp. Brain Res. 192, 189–198 (2009).
47. Miró, E. et al. Attentional deficits in fibromyalgia and its relationships with pain, emotional distress and sleep dysfunction
complaints. Psychol. Health 26, 765–780 (2011).
48. Roca, J. et al. The effects of sleep deprivation on the attentional functions and vigilance. Acta Psychol. (Amst.) 140, 164–176 (2012).
49. Callejas, A., Lupiàñez, J., Funes, M. J. & Tudela, P. Modulations among the alerting, orienting and executive control networks. Exp.
Brain Res. 167, 27–37 (2005).
50. Faul, F., Erdfelder, E., Lang, A. G. & Buchner, A. G*Power 3: A flexible statistical power analysis program for the social, behavioral,
and biomedical sciences. Behav. Res. Methods 39, 175–191 (2007).
51. Paluska, S. A. Caffeine and exercise. Curr. Sports Med. Rep. 2, 213–219 (2003).
52. James, J. E. Caffeine and health. (Academic Press London, 1991).
53. Craig, N. et al. Protocols for the physiological assessment of high performance track, road and mountain cyclists. In Physiological
tests for elite athletes (Human Kinetics, 2000).
54. Tanaka, H., Monahan, K. & Seals, D. Age-predicted maximal heart rate revisited. J. Am. Coll. Cardiol. 37, 153–156 (2001).
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
11
www.nature.com/scientificreports/
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
www.nature.com/scientificreports
López Chicharro, J. & López Mojares, L. M. Fisiología clínica del ejercicio. (Ed. Médica Panamericana, 2008).
EFSA. Scientific Opinion on the safety of caffeine. EFSA J. 13, 4102 (2015).
Arnaud, M. J. The pharmacology of caffeine. Prog. Drug Res. 31, 273–313 (1987).
Schneider, W., Eschman, A. & Zucolotto, A. E-Prime user’s guide. (Psychology Software Tools Inc., 2002).
Crowe, M. J., Leicht, A. S. & Spinks, W. L. Physiological and cognitive responses to caffeine during repeated, high-intensity exercise.
Int. J. Sport Nutr. Exerc. Metab. 16, 528–544 (2006).
Davranche, K. & Audiffren, M. Facilitating effects of exercise on information processing. J. Sports Sci. 22, 419–428 (2004).
Fernandes, A., Medeiros, A. R., Del Rosso, S., Stults-Kolehmainen, M. & Boullosa, D. A. The influence of exercise and physical
fitness status on attention: a systematic review. Int. Rev. Sport Exerc. Psychol. 1–33, https://doi.org/10.1080/1750984X.2018.1455889
(2018).
McMorris, T. & Hale, B. J. Differential effects of differing intensities of acute exercise on speed and accuracy of cognition: a metaanalytical investigation. Brain Cogn. 80, 338–351 (2012).
Glaister, M. & Gissane, C. Caffeine and physiological responses to submaximal exercise: A meta-analysis. Int. J. Sports Physiol.
Perform. 13, 402–411 (2018).
Ferré, S. An update on the mechanisms of the psychostimulant effects of caffeine. J. Neurochem. 105, 1067–1079 (2008).
Lane, J. D. & Phillips-Bute, B. G. Caffeine deprivation affects vigilance performance and mood. Physiol. Behav. 65, 171–175 (1998).
Juliano, L. M. & Griffiths, R. R. A critical review of caffeine withdrawal: empirical validation of symptoms and signs, incidence,
severity, and associated features. Psychopharmacology (Berl.) 176, 1–29 (2004).
Chmura, J., Nazar, K. & Kaciuba-Uscilko, H. Choice reaction time during graded exercise in relation to blood lactate and plasma
catecholamine thresholds. Int. J. Sports Med. 15, 172–176 (1994).
Lumme, V., Aalto, S., Ilonen, T., Någren, K. & Hietala, J. Dopamine D2/D3 receptor binding in the anterior cingulate cortex and
executive functioning. Psychiatry Res. 156, 69–74 (2007).
Ishigami, Y. & Klein, R. M. Repeated measurement of the components of attention using two versions of the Attention Network Test
(ANT): Stability, isolability, robustness, and reliability. J. Neurosci. Methods 190, 117–128 (2010).
Aston-Jones, G. & Cohen, J. D. An integrative theory of locus coeruleus norepinephrine function: Adaptive gain and optimal
performance. Annu. Rev. Neurosci. 28, 403–450 (2005).
Abernethy, B. Visual search in sport and ergonomics: Its relationship to selective attention and performer expertise. Hum. Perform.
1, 205 (1988).
Fan, J., McCandliss, B. D., Fossella, J., Flombaum, J. I. & Posner, M. I. The activation of attentional networks. Neuroimage 26,
471–479 (2005).
Acknowledgements
We thank Prof. Bruce Milliken for his stimulating discussions and proofreading of a first version of the manuscript
during a research stay of F.H. at McMaster University. This research was supported by Universidad Católica de
Valencia “San Vicente Mártir” grant (2019-158-003) to F.H. and by the Spanish Ministerio de Economía, Industria
y Competitividad (PSI2017-84926-P) to J.L. & F.H. The funders had no role in study design, data collection and
analysis, decision to publish, or preparation of the manuscript.
Author Contributions
Conceived and designed the experiments: F.H., E.B. and J.L. Performed the experiments: F.H, E.B. and C.M.
Performed the initial processing of the data: F.H., E.B. and C.M. Analysed the data: F.H., E.B. and J.L. Wrote
the manuscript: F.H. and J.L. All authors contributed to the scientific discussion and reviewed the manuscript.
Coordinated the research: F.H and J.L.
Additional Information
Competing Interests: The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2019
SCIENTIFIC REPORTS |
(2019) 9:10043 | https://doi.org/10.1038/s41598-019-46524-x
12