Subjecting Elite Athletes to Inspiratory Breathing Load
Reveals Behavioral and Neural Signatures of Optimal
Performers in Extreme Environments
Martin P. Paulus1,2,3*, Taru Flagan1, Alan N. Simmons1,3, Kristine Gillis2, Sante Kotturi1, Nathaniel Thom2,
Douglas C. Johnson2, Karl F. Van Orden2, Paul W. Davenport4, Judith L. Swain2,5
1Department of Psychiatry, University of California San Diego, La Jolla, California, United States of America, 2OptiBrain Consortium, San Diego, California, United States of
America, 3Veterans Affairs San Diego Health Care System, San Diego, California, United States of America, 4Department of Physiological Sciences, University of Florida,
Gainesville, Florida, United States of America, 5Singapore Institute for Clinical Sciences-A*STAR and National University of Singapore, Singapore, Singapore
Background: It is unclear whether and how elite athletes process physiological or psychological challenges differently than
healthy comparison subjects. In general, individuals optimize exercise level as it relates to differences between expected
and experienced exertion, which can be conceptualized as a body prediction error. The process of computing a body
prediction error involves the insular cortex, which is important for interoception, i.e. the sense of the physiological condition
of the body. Thus, optimal performance may be related to efficient minimization of the body prediction error. We examined
the hypothesis that elite athletes, compared to control subjects, show attenuated insular cortex activation during an
aversive interoceptive challenge.
Methodology/Principal Findings: Elite adventure racers (n=10) and healthy volunteers (n=11) performed a continuous
performance task with varying degrees of a non-hypercapnic breathing load while undergoing functional magnetic
resonance imaging. The results indicate that (1) non-hypercapnic inspiratory breathing load is an aversive experience
associated with a profound activation of a distributed set of brain areas including bilateral insula, dorsolateral prefrontal
cortex and anterior cingulated; (2) adventure racers relative to comparison subjects show greater accuracy on the
continuous performance task during the aversive interoceptive condition; and (3) adventure racers show an attenuated
right insula cortex response during and following the aversive interoceptive condition of non-hypercapnic inspiratory
Conclusions/Significance: These findings support the hypothesis that elite athletes during an aversive interoceptive
condition show better performance and an attenuated insular cortex activation during the aversive experience.
Interestingly, differential modulation of the right insular cortex has been found previously in elite military personnel and
appears to be emerging as an important brain system for optimal performance in extreme environments.
Citation: Paulus MP, Flagan T, Simmons AN, Gillis K, Kotturi S, et al. (2012) Subjecting Elite Athletes to Inspiratory Breathing Load Reveals Behavioral and Neural
Signatures of Optimal Performers in Extreme Environments. PLoS ONE 7(1): e29394. doi:10.1371/journal.pone.0029394
Editor: Alejandro Lucia, Universidad Europea de Madrid, Spain
Received July 6, 2011; Accepted November 28, 2011; Published January 19, 2012
Copyright: ? 2012 Paulus et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by the Veterans Affairs Health Care System Center for Stress and Mental Health (http://cesamh.org/staff/index.html). The
funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
The neuroscience underlying optimal performance in extreme
environments is in its infancy . Nevertheless, there is a
burgeoning interest in understanding how the brain contributes
to optimizing performance . Altered cortical and subcortical
processing of tasks and external conditions has been proposed as
an important mechanism that differentiates elite performers from
comparison subjects . In a prior study, we examined neural
processing of elite military personnel (U.S. NAVY Sea, Air, and
Land Forces–SEALs) relative to comparison subjects during
emotion face processing, and showed relatively greater right-sided
insula, but attenuated left-sided insula, activation in the elite
performers. Moreover, the U.S. Navy SEALs showed selectively
greater activation to angry target faces relative to fearful or happy
target faces in both right and left insula . These individuals also
show greater insula activation when anticipating a change in
interoceptive state from the current state, but reduced insula
activation to aversive images relative to comparison subjects
(Simmons, in prep). Taken together, these results are consistent
with the hypothesis that elite performers deploy processing
resources that are more focused on specific task demands, and
they are better able to respond to external stimuli that perturb
Interoception comprises the sensing of the physiological
condition of the body , the representation of this internal state
 within the context of ongoing activities, and the initiation of
motivated action to homeostatically regulate the internal state .
PLoS ONE | www.plosone.org1 January 2012 | Volume 7 | Issue 1 | e29394
Interoception is an important process for optimal performance
because it links the perturbation of internal state as a result of
external demands to goal-directed action that maintain a
homeostatic balance . In particular, the interoceptive system
provides information about the internal state to neural systems
that monitor value and salience and are critical for cognitive
control processes. We recently proposed that maintaining an
interoceptive balance by generating body prediction errors in the
presence of significant perturbations may be a neural marker of
optimal performance . This notion is consistent with findings
that elite athletes pay close attention to bodily signals  and may
be particularly adept in generating anticipatory prediction errors
. Further, others  have proposed that individuals regulate
performance via perceived exertion through a ‘‘teleoanticipation’’
process  which is the combination of afferent and efferent
brain processes that attempt to couple the metabolic and
biomechanical limits of the body to the demands of the exercise
task. Specifically, an individual’s expectation of effort perception
during exercise is the basis for an ongoing interpretation of
perceived exertion, and is due to both efferent feed-forward and
afferent feedback signals . Thus, neural systems that process
the internal body state and are able to generate small body
prediction errors may be critical for optimal performance.
An extreme environment can be defined as an external context
that exposes individuals to demanding psychological and/or
physical conditions, and which may have profound effects on
cognitive and behavioral performance . Examples of these types
of environments include combat situations, Olympic-level compe-
tition, and expeditions in extreme cold, at high altitudes, or in
space. Adventure racing is a combination of two or more
endurance disciplines such as orienteering, navigation, cross-
country running, mountain biking, paddling, climbing, and related
rope skills. Individuals participating in adventure racing experi-
ence significant physical and psychological stress during these
competitions, which sometimes result in both significant injury and
in mood-state disruption . In this study we examine elite
adventure racers who are non-military elite performers and who
are often exposed to extreme environments .
The sensation of breathing is a complex process that is
modulated by numerous factors . In addition to chemorecep-
tors that form reflex feedback mechanisms for respiratory motor
activities , breathing is also influenced by internal and external
environmental changes, which is termed behavioral breathing.
Respiratory sensations are an essential interoceptive experience
because there is a profound evaluative component associated with
breathing sensation, and there is a strong motivational aspect to
adjust breathing via the respiratory motivation-to-action neural
system . These sensations are the result of both subcortical and
cortical processes , which include discriminative processing
(the awareness of the spatial, temporal and intensity components
of the respiratory input) and affective processing (the evaluative
and emotional components of the respiratory input).
Resistive load, i.e. restricted inspiration, was first introduced by
Lopata  and Gottfried , and is an airflow-dependent load
 and a simple but powerful experimental approach to induce
an altered interoceptive state. In contrast to expiratory breathing
load which affects C02, inspiratory breathing load results in
stable, unchanged C02levels . Inspiratory breathing loads can
be used to examine breathing difficulty and generate respiratory-
related evoked potentials with several peaks that indicate the
transition from an early sensory component to a later cognitive
aspect [23,24,25,26]. Moreover, resistive loads generate pre-motor
potentials that reflect the involvement of higher cortical motor
areas , they decrease systolic blood pressure , they differ
for males and females , they are perceived less intense in older
individuals , they generate load-dependent increases of
unpleasantness , and the subjective effects can be modified
by attentional distractions . Thus, inspiratory breathing load
provides a powerful experimental approach to examine how
optimal performers respond to the temporary perturbation of the
internal body state.
We have previously proposed that optimal performance may be
related to the ability to effectively minimize the body prediction
error , which allows individuals to better adjust to environ-
mental perturbations. Since the insula cortex is important in
generating body prediction errors , then one would hypoth-
esize that elite athletes show attenuated neural processing in the
insular cortex of afferent aversive interoceptive stimuli. Support for
this hypothesis would provide further evidence that elite
performers show a distinct brain signature that enables them to
adjust more quickly and appropriately to extreme environments.
This approach uses simple laboratory tasks to link neural and
cognitive processes that have been found to be important for elite
performance. As pointed out by others, this approach may help to
explain sporting skill at the highest levels of performance .
The University of California San Diego (UCSD) Institutional
Review Board approved this study and all subjects signed
informed consent. Ten adventure racers (6 males, 4 females) were
recruited by word of mouth and fulfilled the following criteria: (1)
participated in multi-day events on an international level; (2)
placed among the top 5 performing teams in at least 3 races; (3)
completed international races within the past 5 years; (4) were at
least 14 days out from their last race. The last criterion was used to
minimize acute effects related to physical and psychological
exhaustion. Eleven healthy control subjects (8 males, 3 females)
were recruited from other ongoing studies supported by the Center
of Excellence for Stress and Mental Health (CESAMH). All
twenty-one subjects completed the study. The mean age of the
adventure racers was 37.5+/26.0 years, and the controls 36.6+/
26.9 years. The adventure racers completed 16.3+/21.8 years of
education, and the control subjects 16.0+/21.4 years of
education. The groups did not differ in gender (x2=0.382,
p=0.537), age (t(19)=0.30, p=0.76), or years of education
(t(19)=0.29, p=0.77). All subjects were trained to perform a
non-hypercapnic breathing load task prior to fMRI scanning. No
restrictions were placed on the consumption of caffeinated
beverages prior to study, and none of the subjects were smokers.
Several personality and symptom assessment questionnaires
were administered including the Sensation Seeking Scale , the
Barratt Impulsiveness Scale [BIS-11, 35], the Brief Symptom
Inventory (BSI), a brief psychological self-report symptom scale
, and the Connor Davidson Resiliency Scale CD-RISC .
Aversive Interoceptive Stimulus: Non-Hypercapnic
Inspiratory Breathing Load
The subjects wore a nose clip and breathed through a
mouthpiece with a non-rebreathing valve (2600 series, Hans
Rudolph) that maintains an airtight seal. The apparatus was
attached to the scanner head coil to eliminate the need for the
subject to contract mouth muscles. The resistance loads consisted
of sintered bronze disks placed in series in a Plexiglas tube (loading
manifold), with stoppered ports inserted between disks. Loads were
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org2 January 2012 | Volume 7 | Issue 1 | e29394
selected by removing the stopper and allowing the subject to
inspire through the selected port. Each subject was given the
following instructions: ‘‘This task examines how people feel when
breathing becomes difficult. You will breathe through a hose,
which makes breathing-in more difficult. It is important for you to
know that this test is not physically harmful, but you may feel
uncomfortable when you breathe through the hose. You can stop
at any time if breathing becomes too difficult. You will be asked to
breathe through the hose several times. We would like you to
complete a one-page rating scale after each trial’’. Based on
preliminary data and previous experience, 40 cmH2O/L/sec was
selected as load, which alters subjective symptoms without
significantly affecting CO2or O2level. The subjects were asked
to rate their experience on a 10 cm Visual Analog Scale which was
anchored from ‘‘not at all’’ to ‘‘extremely’’ on the following 16
dimensions: pleasant; unpleasant; intense; tingling; fear of losing
control; faintness; fear of dying; unreality; hot/cold flushes;
trembling; choking; abdominal distress; chest pain; palpitations;
sweating; and dizziness, all of which correspond to items used in
one author’s (PWD) prior studies [17,38].
The basic experimental approach was similar to that of a
recently published study by our group involving human touch
. Specifically, individuals performed a simple continuous
performance task during the paradigm. Subjects were asked to
press a button corresponding to the direction pointed by an arrow
on the screen (left arrow=left button, right arrow=right button).
Both accuracy and response latency were recorded and analyzed
to determine effects of anticipation and stimulus presentation. At
the same time, the background color of the stimulus served as a
cue to the impending presentation of the breathing load, with gray
indicating that there will be no load, and yellow indicating a 25%
chance of load. Throughout the task, subjects experienced 3
conditions: (1) baseline condition: the individual performs the
continuous performance task; (2) anticipation condition: the
background color behind the arrow signals an impending
restricted breathing period; and (3) stimulus condition: during
the change in background color there is a 25% probability that the
subject experienced a 40 second period of restricted breathing. We
introduced this probability to maximize the opportunity to
measure the effect of anticipating an aversive interoceptive event.
The implementation of this paradigm utilized an event-related
fMRI design consisting of 2 scans with 256 repetitions
(TR=2 secs) yielding a total scan duration of 17 minutes and
4 seconds. During this period the individual was presented with 32
anticipation conditions during which 8 breathing-load episodes
occurred. The duration of each condition is ‘‘jittered’’ in time to
permit optimal resolution of the hemodynamic response function.
On average the baseline condition lasted 9 seconds, the anticipa-
tion condition 9 seconds, and post-stimulus condition 12 seconds.
Throughout the baseline condition a black arrow on a gray
background was presented on the screen every 3 seconds. For the
anticipation condition, subjects were informed that a blue
background on the screen predicted the subsequent restricted
breathing period; this phase lasted between 6 and 12 seconds. The
main behavioral variable was performance accuracy and latency
during the three different stimulus conditions, and the main
neuroimaging-dependent measure was the activation in function-
ally constrained regions of interest during the anticipation and
stimulus condition relative to the baseline condition.
performed on a 3T GE CXK4 Magnet at the UCSD Keck
Imaging Center, which is equipped with 8 high bandwidth
receivers that allow for shorter readout times and reduced signal
distortions and ventromedial signal dropout. Each one hour
session consisted of a three-plane scout scan (10 seconds), and a
standard anatomical protocol consisting of a sagittally acquired
spoiled gradient recalled (SPGR) sequence (FOV 25 cm; matrix:
1926256; 172 sagittally acquired slices thickness: 1 mm; TR:
8 ms; TE: 3 ms; flip angle=12). We used an 8-channel brain
array coil to axially acquire T2*-weighted echo-planar images
(EPI). The parameters for the EPI scans were: FOV 230 mm,
64664matrix; 40 2.6 mm
TR=2000 ms, TE=32 ms, flip
acquisition was obtained via GE’s ASSET scanning, a form of
sensitivity encoding (SENSE) which uses parallel imaging
reconstruction to allow for sub k-space sampling.
Image analysis pathway.
functional image processing was done with the Analysis of
Functional Neuroimages (AFNI) software package . The
multivariate regressor approach detailed below was used to
characteristics . EPI images were co-registered using a 3D-
coregistration algorithm  that was developed to minimize the
amount of image translation and rotation relative to all other
images. Six motion parameters were obtained across the time
series for each subject. Motion parameters were used as regressors
to adjust EPI intensity changes due to motion artifacts. This has
been shown to increase power in detecting task-related activation
. All slices of the EPI scans were temporally aligned following
registration to assure that different relationships with the regressors
are not due to the acquisition of different slices at different times
during the repetition interval.
Multiple regressor analyses.
generated to delineate the three conditions described above: (1)
anticipation, (2) breathing load, and (3) a post-breathing interval.
To that end, a 0–1 reference function of the particular time
interval was convolved with a gamma variate function 
modeling a prototypical hemodynamic response (6–8 second delay
) and to account for the temporal dynamics of the
hemodynamic response (typically 12–16 seconds) . The
convolved time series was normalized and used as a regressor of
interest. A series of regressors of interest and the motion regressors
were entered into the AFNI program 3dDeconvolve to determine
the height of each regressor for each subject. The main dependent
measure was the voxel-wise normalized relative signal change, or
% signal change for short. Subsequently spatial smoothing with
6 mm FWHM was applied to the % signal change data, which
were transformed into Talairach coordinates based on the
anatomical MR image for group or second-level analysis.
Group level analyses.
For the interoceptive fMRI paradigm
the dependent measure was the % signal changes during the
respectively. These dependent measures were entered into a
mixed effects model . We used the implementation of the linear
mixed effects models in R (www.cran.org), which estimates the
parameters of the mixed model using Maximum Likelihood
Estimation (MLE). These calculations were done within the R
computing environment using routines that read in AFNI data
sets. Specifically, the group (elite athletes versus comparison
subjects),and the experimental
stimulation, post-stimulus interval) were used as a fixed factor,
and subject was used as a random factor. The effects are estimated
using specific contrast matrices. Once these voxel-wise statistics
were calculated, we used a threshold adjustment method based on
Monte-Carlo simulations to guard against identifying false positive
areas of activation. Based on simulations implemented in the
All subject-level structural and
to differencesin task
Regressors of interest were
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org3 January 2012 | Volume 7 | Issue 1 | e29394
AFNI program AlphaSim, by using a constrained region of
interest analysis approach for the insular cortex it was determined
that the volume threshold for clusterwise probability of 0.05 was
512 uL. Only these clusters were considered for further analysis.
Finally, we conducted voxel-wise multiple linear regression
analyses with self-report measures as independent measures, and
the percent signal change during the breathing load condition as
the dependent measure using the robust Huber regressions based
on the rlm program of R.
Adventure racers relative to comparison subjects showed
elevated self-ratings of sensation seeking (Table 1). With the
exception of the thrill and adventure seeking subscale of the
Sensation Seeking Scale, adventure racers rated higher on all
other subscales. There were no significant overall differences
between adventure racers and controls on the Barratt Impulsivity
Scale, however, adventure racers rated themselves higher on the
perseverance subscale. Finally, there were no differences between
adventure racers and comparison subjects on the Brief Symptoms
Inventory or on the Connor Davidson Resiliency Scale.
Self-report during breathing load
There was an overall increase in VAS scale ratings of
unpleasantness when comparing baseline to 40 cm H2O/L/sec
load [F(1,15)=7.427, p=0.0156] (Figure 1). However there were
no significant group differences [F(1,17)=0.642, p=0.4339] or
group by condition interaction [F(1,15)=0.126, p=0.7273].
Although the breathing load resulted in an aversive experience,
there were no differences in the degree of unpleasantness between
adventure racers and comparison subjects.
Behavioral performance during breathing load
Individuals took longer to select a response during the
[F(2,95)=6.242, p=0.0028](Figure 2). Adventure racers did not
differ from comparison subjects [F(1,19)=1.034 p=0.322] and
there was no differential effect of anticipation or load condition for
the elite athletes relative to the controls [F(2,95)=2.441
p=0.0925]. In contrast, there was an overall trend for increased
response accuracy between groups during the anticipation and
breathing load conditions [F(2,95)=2.8 p=0.0630]. More
specifically, whereas healthy volunteers showed no clear difference
in accuracy during the different conditions, adventure racers
showed greater response accuracy during the anticipation and
breathing load condition, resulting in a significant group-by-
condition interaction [F(2,95)=4.5, p=0.0136]. Thus, the
aversive interoceptive perturbation improved performance in
adventure racers but not in healthy controls.
brain activation that varied across task condition (Table 2), and
which affected several areas of the brain as shown in Figure 3. In
general breathing load resulted in significant activation increases
in the bilateral insular cortex, anterior cingulate, and also in the
bilateral dorsolateral prefrontal cortex. In each of these areas
activation was significantly greater during the breathing load and
post-breathing load period relative to the anticipation period.
Moreover, both adventure racers and healthy comparison subjects
showed similarlystrong activation
interest analysis, the right insular cortex (Figure 4 and Table 3) was
the only brain area that showed a significant task-by-group
interaction, with adventure racers relative to comparison subjects
differentially activating this brain area as a function of task
condition. The right insular cortex was found to activate during
Loaded breathing induced a large change in
Using the constrained region of
Table 1. Personality and symptom assessment of elite
athletes and comparison subjects.
Sensation Seeking Scale 27.77**4.23 20.58 3.98
Thrill and Adventure Seeking
Experience Seeking (ES) 7.44* 1.665.911.62
Disinhibition (Dis)7** 1.934.41 2.35
Boredom Susceptibility (BS)4** 1.652.33 1.23
Attention 12.752.5411 2.18
Motor16.5 3.7 15.081.75
Self-Control 15.873.48 16.672.42
Cognitive Complexity12.752.3712.5 2.49
Perseverance9.62**1.3 7.58 1.48
Cognitive Instability7.125 2.416 1.56
CDRISC 30.6612.3231.42 10.07
Personality and symptom assessments show that elite athletes score higher on
sensation seeking and perseverance than comparison subjects.
Figure 1. Visual Analog Rating during Baseline (no load) and
40 cm H2O/L/sec inspiratory breathing load in comparison
subjects and elite athletes, respectively. Both groups showed
increased unpleasantness during the 40 cm H2O/L/sec load condition.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org4 January 2012 | Volume 7 | Issue 1 | e29394
the task in general [F(1,38)=11.295, p=0.0018]. Specifically,
relative to the anticipation condition, both groups showed greater
activation during the breathing load and post-breathing load
conditions [F(2,38)=5.890, p=0.0059]. Moreover, there were no
comparison subjects (F(1,19)=2.977, p=0.1007). Importantly,
whereas healthy volunteers showed an overall increase in
activation during the load condition, adventure racers showed
greater activation during the anticipation phase and attenuated
activation during the load phase, resulting in a significant
Finally, activation of the right insular cortex during the
breathing load condition correlated negatively with performance
accuracy (r=2048, p=0.02), with greater activation resulting in
lower performance accuracy. There was no significant correlation
of right insular cortex activation with latency or with the subjective
rating of unpleasantness (all ps.0.05).
Brain behavior relationships.
robust regression with both groups revealed that the degree of
brain activation during breathing load in two brain areas
correlated with the subjective ratings of unpleasantness due to
breathing load. Specifically, ventral anterior cingulate and left
anterior insula (including lateral inferior frontal gyrus) showed
greater activation in those subjects with higher unpleasantness
ratings (Figure 5). There were no differences across groups in these
areas. Further analysis of the brain area that correlated with self-
rated unpleasantness revealed two additional relationships. First,
higher impulsiveness ratings were associated with lower activation
in the left anterior insula during breathing load (r=20.46,
p=0.04). Second, greater activation during anticipation in the
right anterior insula was associated with less activation in the left
anterior insula during breathing load (r=20.53, p=0.01). There
were no correlations between either the ventral anterior cingulate
area or the insula area and measures of performance (accuracy or
latency) during the continuous performance task.
Whole brain analyses using
We examined the hypothesis that elite athletes show attenuated
neural processing of aversive interoceptive stimulation in the
insular cortex by testing the behavioral and neural processing
response of elite athletes during an aversive interoceptive non-
hypercapnic breathing load. The experiment yielded three main
results. First, non-hypercapnic inspiratory breathing load is an
aversive experience that results in a profound activation of a
distributed set of brain areas including bilateral insula, dorsolateral
prefrontal cortex and anterior cingulate. The degree of activation
in a subset of brain areas consisting of the ventral anterior
cingulate and left anterior insula was correlated with subjective
ratings of unpleasantness. Second, adventure racers, compared
with control subjects, show greater accuracy on the continuous
performance task during the aversive interoceptive stimulation.
Third, adventure racers show attenuated right insula cortex
response during the breathing load and the post-breathing
condition. Taken together, this experimental approach not only
shows that insular activation differentiates elite athletes (as an
example of optimal performers) from comparison subjects, but also
shows that these individuals perform better during aversive
interoceptive stimulation on a simple continuous performance
task. Thus, non-hypercapnic breathing load during functional
neuroimaging provides a laboratory approach to study elite
performers and identify behavioral and brain processes that are
important for optimal performance in extreme environments.
Adventure racing at an elite level comprises competitions that
last 100 hours or longer and can cause physical injury and
perturbation of mood states, which in turn can have profound
impact on optimal performance . Measures of mood states
have been used to predict athletic injury , but much less is
known about the central nervous system contribution to optimal
performance. There is a growing interest in understanding how
basic brain processes influence levels of performance, and several
investigators have begun to delineate which brain processes
Figure 2. Behavioral performance (latency and accuracy)
during the continuous performance task in both comparison
subjects and athletes (left) and separately for each group
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org5 January 2012 | Volume 7 | Issue 1 | e29394
contribute to athletic performance [49,50]. The finding in this
study identifies a brain area (the right insula) and a process (the
response to aversive interoceptive stimulation) that differ in elite
athletes. The finding is consistent with the proposed role of the
insular cortex as a component of the so-called central command,
i.e. the brain systems that are important for modulating the degree
to which individuals engage in demanding athletic performance
[49,51]. Several neuroimaging studies using Single Photon
Emission Tomography during physical exercise have demonstrat-
ed changes in activation in the insula cortex. For example
increased left insula regional cerebral blood flow (rCBF) was
observed during active, but not passive, cycling . Moreover,
greater insular rCBF was positively correlated with levels of
perceived cycling intensity  and with individual blood pressure
changes. Therefore, there is an emerging role of the insular cortex
in processing effort as the athlete perceives it during exercise and
Table 2. Main effect of breathing restriction on brain
activation in comparison subjects and elite athletes.
Volumexyz Brain Area BA
213 22 Bilateral Cingulate GyrusBA 23
74242936 33Right Superior Frontal Gyrus BA 9
23637 29 Left Middle Frontal GyrusBA 9
19843 4914 Right Medial Frontal GyrusBA 10
217 Right DecliveBA 18
Volume (mL), center of mass coordinate, and brain area based on the voxel-wise
mixed model main effect of breathing load. These areas showed brain
activation related to loaded breathing for both comparison subjects and elite
Figure 3. Main effect of task, i.e. brain changes as a consequence of inspiratory breathing load in both comparison subjects and
elite athletes. Activation increases primarily during the breathing load and post-breathing load condition.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org6 January 2012 | Volume 7 | Issue 1 | e29394
modulating physiological parameters that are critical for optimiz-
ing physical performance.
The relationship between self-rated unpleasantness and the
degree of activation in ventral anterior cingulate, left anterior
insular and lateral inferior frontal gyrus supports the notion that
brain structures which are important for regulating subjective
mood states [54,55] are also critical for modulating optimal
performance. These findings are consistent with those of William-
son and colleagues who have used false feedback of less than or
greater than actual physical demand to examine central nervous
system regulation of perceived exertions. These investigators found
that under these conditions, changes in rCBF in left and right
insular cortex as well as anterior cingulate cortex correlated with
perceived exertion, but not with changes in heart rate or blood
pressure . Furthermore, both insular and anterior cingulate
cortices were also found to activate during imagined exercise .
Taken together, the insula and anterior cingulate cortex are
important for processing levels of exertion with false feedback and
in an imaginary condition, i.e. can function as a central command
system without the necessity of peripheral feedback. The insula
and anterior cingulate also interact with thalamic and brainstem
structures which are important for cardiovascular integration.
Therefore, both insula and anterior cingulate may process the
individual’s sense of effort or exertion with and without the need
for peripheral afferents .
The differential activation in the right insular cortex in elite
adventure racers during breathing load and in the post-breathing
condition is similar to the differential activation during an emotion
processing task in NAVY SEALs relative to male comparison
subjects that we previously reported . Although there are a
number of caveats, e.g. there were different task and conditions,
different selective demand-dependent activations, and different
genders of subjects, there are some common findings between
these two studies that deserve to be highlighted. In both conditions
elite performers showed relatively less activation in conditions that
were ‘‘more challenging’’ to healthy volunteers. In extension of the
previous study that did not show performance accuracy or
response latency differences between elite war fighters and
comparison subjects , the current study shows that elite athletes
demonstrate greater accuracy under challenging conditions. The
combination of a continuous performance task and non-hyper-
capnic inspiratory breathing load may provide a simple behavioral
probe to examine both brain processing and behavioral perfor-
mance differences in individuals who are training to acquire elite
performance status. Moreover, the brain-behavior relationship
between the insular cortex and self-rated unpleasantness and task
performance may provide an initial step toward development of a
peripheral biomarker of optimal performance.
We have recently proposed a neuroanatomical processing
model as a heuristic guide to understand how interoceptive
processing may contribute to optimal performance. In this model
we propose that optimal performers generate more efficient body
prediction errors, i.e. the difference between the value of the
anticipated/predicted and value of the current interoceptive state,
as a way of adapting to extreme environments. However, body
prediction error differences may occur on several levels. For
example, optimal performers may receive different afferent
information via the C-fiber pathway that conveys spatially- and
time-integrated affective information . Alternatively, optimal
performers may generate centrally different interoceptive states
(e.g., via contextual associations from memory), which are
processed in insular cortex via connections to temporal and
parietal cortex to generate body states based on conditioned
associations . Consistent with this idea, Williamson and
colleagues suggest that the neural circuitry underlying central
regulation of performance includes the insular and anterior
cingulate cortex that interact with thalamic and brainstem
structures which are important for cardiovascular integration
 as well as for the central modulation of cardiovascular
Optimal performers may also differentially integrate interocep-
tive states within the insular cortex (which shows a clear gradient
from the dorsal-posterior to ventral-anterior part) to provide an
increasingly ‘‘contextualized’’ representation of the interoceptive
state . This integration may occur irrespective of whether it is
generated internally or via the periphery. The relative increase in
activation in the mid-insula in adventure racers prior to
experiencing the breathing load, and the relatively attenuated
activation after the load experience, support the notion that the
aversive interoceptive experience is less disruptive to these elite
athletes compared to control subjects, and may lead to relatively
fewer changes in the subjective response to this stressor.
Figure 4. Group6
significantly greater activation during breathing load and
post-breathing load condition in comparison subjects relative
to elite athletes.
6Task Interaction, right middle insula showed
Table 3. Task by group interaction: Elite athletes relative to
comparison subjects differentially activated the right insula
Volume (uL)xyz AreaBA
221 Right InsulaBA 13
Volume (mL), center of mass coordinate, and brain area based on the voxel-wise
mixed model task by group interaction. There were significant differences in the
right insula cortex between adventure racers and comparison subjects.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org7 January 2012 | Volume 7 | Issue 1 | e29394
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org8 January 2012 | Volume 7 | Issue 1 | e29394
Optimal performers may generate different context-dependent
valuation of the interoceptive states within the orbitofrontal cortex
 leading to altered error processing in the anterior cingulate
 and selection of different actions . The findings that both
ventral anterior cingulate and left anterior insula response are
important for the subjective effects of the breathing load support
the notion that optimal performers may show different integration
of aversive interoceptive stimuli. These results are consistent with
those of Hilty and colleagues  who reported that individuals
who perform a handgrip exercise prior to task failure show
increased activation in both the mid/anterior insular cortex and
the thalamus. Thus, greater activation and possibly a larger body
prediction error might predict sub-optimal performance.
Finally, it is also unclear whether optimal performers generate
different learning signals (similar to reward prediction error ),
as part of the interactions between the insula and the basolateral
amygdala  and the ventral striatum . The current results
are consistent with the notion that integration within different
parts of the insula cortex as well as top-down, feed-forward
information from other brain areas are important to optimize
performance. Optimal performers are able to more quickly adapt
to both bottom-up interoceptive afferents and top-down cognitive
control brain areas that modulate mood and anxiety  in
regulating one’s response to an aversive interoceptive perturba-
tion. Future investigation will need to examine at what stage of the
pathway elite athletes or optimal performers differ from compar-
ison subjects. This will require not only studies with more subjects
but also different paradigmatic approaches. Nevertheless, by
disentangling the processes that contribute to optimal performance
one can begin to develop brain-process specific interventions that
aim to improve performance.
The central governor model focused on perceived exertion
 (the subjective perception of exercise intensity) has been
used to explain performance differences in athletes .
Recently this model has been extended by Tucker and
colleagues  based on prior formulations by Hampson .
Specifically, a system of simultaneous efferent feed-forward and
afferent feedback signals are thought to optimize performance
by overcoming fatigue through permitting continuous compen-
sation for unexpected peripheral events . Afferent informa-
tion from various physiological systems and external or
environmental cues at the onset of exercise can be used to
forecast the duration of exercise within homeostatic regulatory
limits. This enables individuals to terminate the exercise when
the maximal tolerable perceived exertion is attained. In this
model the brain creates a dynamic representation of an
expected exertion against which the experienced exertion can
be continuously compared  to prevent exertion from
exceeding acceptable levels. The notion of a differential between
expected and experienced exertion parallels our model of the
body prediction error . However, the degree to which
peripheral input is necessary is still under debate. For example,
Marcora and colleagues have developed a psychobiological
model which proposes that perceived exertion is generated by a
top-down or feed-forward signal , i.e. the brain – not the
body – generates the sense of exertion. These investigators have
argued that the a centrally generated corollary discharge of the
brain is critical for optimal effort , and that mental fatigue
affects performance via altered perception of effort rather than
afferent and body originating cardiorespiratory and musculoe-
nergetic mechanisms . Nevertheless, whether it is a purely
central process, as suggested by Marcora, or an interaction
between afferent peripheral feedback and efferent central feed-
forward systems, the differential between the expected and
observed, i.e. the body prediction error, is the critical variable
that moderates performance. The implementation of this
process in the brain and its modulation by nature or nurture
will be central to understand optimal performance.
This investigation had several limitations. First, the group of
elite athletes we studied was relatively small and thus there may be
a lack of power to detect additional behavioral/functional
relationships. With larger number of subjects and different tasks,
other important relationships may become apparent. Second, this
cross-sectional study could not address the question of whether the
observed processing differences were part of the preexisting
characteristics of individuals who were selected and then trained
to become elite athletes, or whether these neural processing
differences were a consequence of training. Thus, future studies
will need to examine, in a within-subjects study design, individuals
prior to and again after elite athlete training.
This systems neuroscience approach to understanding optimal
performance in extreme environments has several advantages over
traditional descriptive approaches. First, identifying the role of
specific neural substrates in optimal performance is the first step to
develop more targeted interventions. For example, if attenuated
insular activation during aversive interoceptive experiences is
consistent with optimal performance, one may begin to target
insula modulation as a brain intervention to improve performance.
Second, studies of specific neural processing pathways involved in
performance in extreme environments can be used to determine
which processes are important for modulating optimal perfor-
mance. For example, it may be possible to use training of
anticipatory processing of aversive interoceptive events as a way of
improving the efficiency of deployment of effortful resources in
extreme environments. Third, quantitative assessment of the
contribution of different neural systems to performance in extreme
environments could be used as indicators of training status or
preparedness. These are just some of the possibilities for utilizing
neuroscience approaches to gain a better understanding of what
makes individuals perform differently when exposed to extreme
environments. As a consequence, one can begin to employ a
rational approach to develop strategies to improve performance in
The views expressed in this article are those of the authors and do not
reflect the official policy or position of the Navy, Department of Defense, or
the U.S. Government. This research has been conducted in compliance
with all applicable federal regulations governing the protection of human
subjects in research.
Conceived and designed the experiments: MPP ANS KFVO PWD JLS.
Performed the experiments: MPP TF KG ANS SK NT DCJ. Analyzed the
data: MPP TF ANS SK. Contributed reagents/materials/analysis tools:
MPP TF ANS. Wrote the paper: MPP ANS TF SK NT DCJ KFVO PWD
Figure 5. Brain activation during breathing load and self-rating of unpleasantness correlations in both comparison subjects and
elite athletes. Those individuals who rated the load as more unpleasant also showed greater activation in ventral ACC and left insula.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org9 January 2012 | Volume 7 | Issue 1 | e29394
1. Yarrow K, Brown P, Krakauer JW (2009) Inside the brain of an elite athlete: the
neural processes that support high achievement in sports. Nat Rev Neurosci 10:
2. Committee on Opportunities in Neuroscience for Future Army Applications
Board on Army S, Technology Division on E, Physical S (2009) Opportunities in
Neuroscience for Future Army Applications. Washington, D.C.: National
Academic Press. pp 1–120.
3. Nielsen JB, Cohen LG (2008) The Olympic brain. Does corticospinal plasticity
play a role in acquisition of skills required for high-performance sports? J Physiol
4. Paulus MP, Simmons AN, Fitzpatrick SN, Potterat EG, Van Orden KF, et al.
(2010) Differential brain activation to angry faces by elite warfighters: neural
processing evidence for enhanced threat detection. PLoS One 5: e10096.
5. Craig AD (2002) How do you feel? Interoception: the sense of the physiological
condition of the body. Nat Rev Neurosci 3: 655–666.
6. Craig AD (2009) How do you feel - now? The anterior insula and human
awareness. Nat Rev Neurosci 10: 59–70.
7. Craig AD (2007) Interoception and Emotion: a Neuroanatomical Perspective In:
Lewis M, Haviland-Jones JM, Feldman Barrett L, eds. Handbook of Emotions.
New York: Guilford Press. pp 272–290.
8. Paulus MP, Potterat EG, Taylor MK, Van Orden KF, Bauman J, et al. (2009) A
neuroscience approach to optimizing brain resources for human performance in
extreme environments. Neurosci Biobehav Rev 33: 1080–1088.
9. Philippe RA, Seiler R (2005) Sex differences on use of associative and
dissociative cognitive strategies among male and female athletes. Percept Mot
Skills 101: 440–444.
10. Aglioti SM, Cesari P, Romani M, Urgesi C (2008) Action anticipation and
motor resonance in elite basketball players. Nat Neurosci 11: 1109–1116.
11. Tucker R (2009) The anticipatory regulation of performance: the physiological
basis for pacing strategies and the development of a perception-based model for
exercise performance. Br J Sports Med 43: 392–400.
12. Hampson DB, St Clair Gibson A, Lambert MI, Noakes TD (2001) The
influence of sensory cues on the perception of exertion during exercise and
central regulation of exercise performance. Sports Med 31: 935–952.
13. Anglem N, Lucas SJ, Rose EA, Cotter JD (2008) Mood, illness and injury
responses and recovery with adventure racing. Wilderness Environ Med 19:
14. Lucas SJ, Anglem N, Roberts WS, Anson JG, Palmer CD, et al. (2008) Intensity
and physiological strain of competitive ultra-endurance exercise in humans.
J Sports Sci 26: 477–489.
15. Adriaensen D, Timmermans JP (2011) Breath-taking complexity of vagal C-fibre
nociceptors: implications for inflammatory pulmonary disease, dyspnoea and
cough. J Physiol 589: 3–4.
16. Homma I, Masaoka Y (2008) Breathing rhythms and emotions. Exp Physiol 93:
17. Davenport PW, Vovk A (2009) Cortical and subcortical central neural pathways
in respiratory sensations. Respir Physiol Neurobiol 167: 72–86.
18. Evans KC (2010) Cortico-limbic circuitry and the airways: insights from
functional neuroimaging of respiratory afferents and efferents. Biol Psychol 84:
19. Lopata M, La Fata J, Evanich MJ, Lourenco RV (1977) Effects of flow-resistive
loading on mouth occlusion pressure during CO2 rebreathing. Am Rev Respir
Dis 115: 73–81.
20. Gottfried SB, Altose MD, Kelsen SG, Fogarty CM, Cherniack NS (1978) The
perception of changes in airflow resistance in normal subjects and patients with
chronic airways obstruction. Chest 73: 286–288.
21. Kifle Y, Seng V, Davenport PW (1997) Magnitude estimation of inspiratory
resistive loads in children with life-threatening asthma. Am J Respir Crit Care
Med 156: 1530–1535.
22. Lofaso F, Isabey D, Lorino H, Harf A, Scheid P (1992) Respiratory response to
positive and negative inspiratory pressure in humans. Respir Physiol 89: 75–88.
23. Davenport PW, Chan PY, Zhang W, Chou YL (2007) Detection threshold for
inspiratory resistive loads and respiratory-related evoked potentials. J Appl
Physiol 102: 276–285.
24. Revelette WR, Davenport PW (1990) Effects of timing of inspiratory occlusion
on cerebral evoked potentials in humans. J Appl Physiol 68: 282–288.
25. Knafelc M, Davenport PW (1999) Relationship between magnitude estimation
of resistive loads, inspiratory pressures, and the RREP P(1) peak. J Appl Physiol
26. Davenport PW, Friedman WA, Thompson FJ, Franzen O (1986) Respiratory-
related cortical potentials evoked by inspiratory occlusion in humans. J Appl
Physiol 60: 1843–1848.
27. Raux M, Straus C, Redolfi S, Morelot-Panzini C, Couturier A, et al. (2007)
Electroencephalographic evidence for pre-motor cortex activation during
inspiratory loading in humans. J Physiol 578: 569–578.
28. Jones CU, Sangthong B, Pachirat O (2010) An inspiratory load enhances the
antihypertensive effects of home-based training with slow deep breathing: a
randomised trial. J Physiother 56: 179–186.
29. Alexander-Miller S, Davenport PW (2010) Perception of multiple-breath
inspiratory resistive loads in males and females. Biol Psychol 84: 147–149.
30. Allen SC, Vassallo M, Khattab A (2009) The threshold for sensing airflow
resistance during tidal breathing rises in old age: implications for elderly patients
with obstructive airways diseases. Age Ageing 38: 548–552.
31. von Leupoldt A, Dahme B (2005) Differentiation between the sensory and
affective dimension of dyspnea during resistive load breathing in normal
subjects. Chest 128: 3345–3349.
32. von Leupoldt A, Seemann N, Gugleva T, Dahme B (2007) Attentional
distraction reduces the affective but not the sensory dimension of perceived
dyspnea. Respir Med 101: 839–844.
33. Paulus MP, Stein MB (2006) An Insular View of Anxiety. BiolPsychiatry 60:
34. Zuckerman M (2007) The sensation seeking scale V (SSS-V): Still reliable and
valid. Personality and Individual Differences 43: 1303–1305.
35. Barratt ES, Patton JH (1983) Impulsivity: Cognitive, behavioral, and
psychophysiological correlates. In: Zuckerman M, ed. Biological Bases of
Sensation Seeking, Impulsivity, and Anxiety. Hillsdale, NJ: Erlbaum. pp 77–116.
36. Derogatis LR, Melisaratos N (1983) The Brief Symptom Inventory: an
introductory report. Psychol Med 13: 595–605.
37. Connor KM, Davidson JR (2003) Development of a new resilience scale: the
Connor-Davidson Resilience Scale (CD-RISC). Depress Anxiety 18: 76–82.
38. Chan PY, Davenport PW (2008) Respiratory-related evoked potential measures
of respiratory sensory gating. J Appl Physiol 105: 1106–1113.
39. Lovero KL, Simmons AN, Aron JL, Paulus MP (2009) Anterior insular cortex
anticipates impending stimulus significance. Neuroimage 45: 976–983.
40. Cox RW (1996) AFNI: software for analysis and visualization of functional
magnetic resonance neuroimages. Computers and Biomedical Research 29:
41. Haxby JV, Petit L, Ungerleider LG, Courtney SM (2000) Distinguishing the
functional roles of multiple regions in distributed neural systems for visual
working memory. Neuroimage 11: 145–156.
42. Eddy WF, Fitzgerald M, Noll DC (1996) Improved image registration by using
Fourier interpolation. Magn Reson Med 36: 923–931.
43. Skudlarski P, Constable RT, Gore JC (1999) ROC analysis of statistical methods
used in functional MRI: individual subjects. Neuroimage 9: 311–329.
44. Boynton GM, Engel SA, Glover GH, Heeger DJ (1996) Linear systems analysis
of functional magnetic resonance imaging in human V1. J Neurosci 16:
45. Friston KJ, Frith CD, Turner R, Frackowiak RS (1995) Characterizing evoked
hemodynamics with fMRI. Neuroimage 2: 157–165.
46. Cohen MS (1997) Parametric analysis of fMRI data using linear systems
methods. Neuroimage 6: 93–103.
47. Littell RC, Pendergast J, Natarajan R (2000) Modelling covariance structure in
the analysis of repeated measures data. Stat Med 19: 1793–1819.
48. Galambos SA, Terry PC, Moyle GM, Locke SA, Lane AM (2005) Psychological
predictors of injury among elite athletes. Br J Sports Med 39: 351–354.
49. Williamson JW, Fadel PJ, Mitchell JH (2006) New insights into central
cardiovascular control during exercise in humans: a central command update.
Exp Physiol 91: 51–58.
50. Marcora SM (2008) Do we really need a central governor to explain brain
regulation of exercise performance? Eur J Appl Physiol 104: 929–931.
51. Noakes TD, Peltonen JE, Rusko HK (2001) Evidence that a central governor
regulates exercise performance during acute hypoxia and hyperoxia. J Exp Biol
52. Williamson JW, Nobrega AC, McColl R, Mathews D, Winchester P, et al.
(1997) Activation of the insular cortex during dynamic exercise in humans.
J Physiol 503(Pt 2): 277–283.
53. Williamson JW, McColl R, Mathews D, Ginsburg M, Mitchell JH (1999)
Activation of the insular cortex is affected by the intensity of exercise. J Appl
Physiol 87: 1213–1219.
54. Drevets WC, Price JL, Simpson JR, Todd RD, Reich T, et al. (1997) Subgenual
prefrontal cortex abnormalities in mood disorders. Nature 386: 824–827.
55. Mayberg HS (1997) Limbic-cortical dysregulation: a proposed model of
depression. J Neuropsychiatry Clin Neurosci 9: 471–481.
56. Williamson JW, McColl R, Mathews D, Mitchell JH, Raven PB, et al. (2001)
Hypnotic manipulation of effort sense during dynamic exercise: cardiovascular
responses and brain activation. J Appl Physiol 90: 1392–1399.
57. Williamson JW, McColl R, Mathews D, Mitchell JH, Raven PB, et al. (2002)
Brain activation by central command during actual and imagined handgrip
under hypnosis. J Appl Physiol 92: 1317–1324.
58. Gray MA, Critchley HD (2007) Interoceptive basis to craving. Neuron 54:
59. Shipp S (2005) The importance of being agranular: a comparative account of
visual and motor cortex. Philos Trans R Soc Lond B Biol Sci 360: 797–814.
60. Rolls ET (2004) Convergence of sensory systems in the orbitofrontal cortex in
primates and brain design for emotion. Anat Rec A Discov Mol Cell Evol Biol
61. Carter CS, Braver TS, Barch DM, Botvinick MM, Noll D, et al. (1998) Anterior
cingulate cortex, error detection, and the online monitoring of performance.
Science 280: 747–749.
62. Rushworth MF, Behrens TE (2008) Choice, uncertainty and value in prefrontal
and cingulate cortex. Nat Neurosci 11: 389–397.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org10 January 2012 | Volume 7 | Issue 1 | e29394
63. Hilty L, Jancke L, Luechinger R, Boutellier U, Lutz K (2010) Limitation of
physical performance in a muscle fatiguing handgrip exercise is mediated by
thalamo-insular activity. Hum Brain Mapp.
64. Schultz W, Dickinson A (2000) Neuronal coding of prediction errors. Annu Rev
Neurosci 23: 473–500.
65. Jasmin L, Burkey AR, Granato A, Ohara PT (2004) Rostral agranular insular
cortex and pain areas of the central nervous system: a tract-tracing study in the
rat. J Comp Neurol 468: 425–440.
66. Fudge JL, Breitbart MA, Danish M, Pannoni V (2005) Insular and gustatory
inputs to the caudal ventral striatum in primates. J Comp Neurol 490: 101–118.
67. Paulus MP, Stein MB (2010) Interoception in anxiety and depression. Brain
Struct Funct 214: 451–463.
68. Borg G, Dahlstrom H (1962) A case study of perceived exertion during a work
test. Acta Soc Med Ups 67: 91–93.
69. St Clair GA, Noakes TD (2004) Evidence for complex system integration and
dynamic neural regulation of skeletal muscle recruitment during exercise in
humans. Br J Sports Med 38: 797–806.
70. Marcora S (2010) Counterpoint: Afferent feedback from fatigued locomotor
muscles is not an important determinant of endurance exercise performance.
J Appl Physiol 108: 454–456.
71. Marcora SM, Staiano W, Manning V (2009) Mental fatigue impairs physical
performance in humans. J Appl Physiol 106: 857–864.
Optimal Performer’s Brain Signature
PLoS ONE | www.plosone.org11 January 2012 | Volume 7 | Issue 1 | e29394