BJMB
Brazilian Journal of Motor Behavior
Tutorials
!
Hill et al.
2023
VOL.17
N.6
270 of 281
How does motor performance change with increasing stress doses? A tutorial on dose-
response profiles applied to crew rowing
YANNICK HILL
1,2
| LAURA S. CUIJPERS
3
| PAULA L. SILVA
4
| RUUD J. R. DEN HARTIGH
3
| ADAM W. KIEFER
5
1
Department of Human Movement Sciences, Faculty of Behavioral and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam Movement Sciences, the
Netherlands
2
Lyda Hill Institute for Human Resilience, University of Colorado Colorado Springs, Colorado Springs, USA
3
Department of Psychology, University of Groningen, Netherlands
4
Center for Cognition, Action & Perception, Department of Psychology, University of Cincinnati, OH, USA
5
Department of Exercise and Sport Science, University of North Carolina at Chapel Hill, NC, USA
Correspondence to:!Yannick Hill, Department of Human Movement Sciences, Vrije Universiteit Amsterdam, Van der Boechorstraat 7, 1081 BT Amsterdam,
Netherlands.
email: y.hill@vu.nl
https://doi.org/10.20338/bjmb.v17i6.398
ABBREVIATIONS
HR Heart rate
HRmax
Appr
Approximated the maximum heart rate
HR
rest
Resting heart rate
PE Point-estimates
RPE Rate of Perceived Exertion
SD Standard deviations
SDDϕ Standardized the SD of the discrete relative
phase
SDϕ
catch
Standard deviations start of the drive phase
phase for catch
SDϕ
finish
Standard deviations of discrete relative
phase for finish
SDϕ
i
Standard deviation of relative phase in each
trial
SDϕ
lowest
Standard deviation of relative phase
corresponding to the trial with the lowest
dose
spm Stroke per minute
t
1,j
Time of the j
th
peak of the forward-backward
position of rower 1
t
2,j
Time of the j
th
peak of the forward-backward
position of rower 2
PUBLICATION DATA
Received 11 09 2023
Accepted 02 12 2023
Published 22 12 2023
ABSTRACT
Many models of motor performance acknowledge that adapting to stressors plays a major role
in how we move. However, most models lack a precise conceptualization of the way in which
stress-response dynamics unfold. To fill this void, we first present popular models from the
domain of biology and psychology which argue that the impact of a stressor depends on its
dose. Next, we provide a tutorial using the example of crew rowing to demonstrate how these
models can be scaled to human motor performance. In this example, the dose of the stressor
is varied by target times for several 500 m races and the response variable represents the
crew coordination. Specifically, we discuss how the necessary parameters can be determined
a priori and how the data can be analyzed to pinpoint the exact dose-response relationship.
These strides are necessary for developing more comprehensive theories of motor
performance and engage in cross-disciplinary research on the impact of stressors.
KEYWORDS: Complex Systems | Coordination Dynamics | Hormesis | Joint Action |
Phenotypic Plasticity | Resilience
INTRODUCTION
Researchers have long been interested in the impact of stressors on human behavior in general and human motor performance
in particular. A stressor reflects a stimulus in the environment that fosters an adaptation process by an organism exposed to it
1-3
.
Different models have been developed to explain the dynamic changes in human motor performance in response to stressors
4,5
. For
example, Nieuwenhuys and Oudejans
6,7
present a model on how a particular kind of stressor (anxiety) shapes performance through its
influences on our perceptions, decision-making, and related actions. Models differ in how they conceptualize the stressor-performance
relationship, yet they agree on at least one aspect: the stressor’s impact on performance varies with its dose
6-8
. However, a detailed
account of this relationship is lacking. Specifically, we rarely pinpoint how this relation can be measured, examined, and shaped by
interventions
2,9
. Therefore, the aim of this paper is to provide new avenues for determining the dose-response dynamics in human motor
performance. To this end, we first introduce three prominent dose-response models of stressors from the biological and psychological
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
271 of 281
literature. Second, we explain how these dose-response models can be scaled to motor performance. Finally, we provide an
accompanying empirical protocol that serves as a tutorial to detect different dose-response models for future studies.
Dose-Response Models in Biology and Psychology
In order to investigate the effect of varying doses of a stressor on an organism, biologists typically establish so-called dose-
response profiles
10-12
. The key ingredients for these profiles are: a) a stressor that can be systematically varied with regards to its dose,
and b) a relevant, measurable response (i.e., adaptation) of the organism
9
. The stressor may reflect any stimulus of interest that fosters
an adaptation process. The response variable can be scaled from an organism’s behavior to the functionality of a specific subsystem of
the organism (e.g., the immune system) and even to small-scaled cellular activities that represent a functional change as a result of the
exposure to the stressor. To create a dose-response profile, the response is plotted as a function of the increasing dose
3
. The shape of
the resulting curve depicts the dose-response model that describes the dynamics between the changes of the stressor and the observed
response
10
. This allows researchers to determine the dose of a specific substance, such as medication, that can be exposed to an
organism to trigger beneficial responses, while avoiding toxic effects from overdoses
11
.
Following the primarily negative connotation of the term stressor, some dose-response models only focus on the detrimental
effects of stressors regardless of their dose
10-12
. For example, in the domain of biology, the exposure to radiation has originally been
conceptualized as solely inducing negative outcomes for organisms even at the lowest dose (see Figure 1A
10,12
). This means that the
organism’s functioning decreases proportionally to the increasing doses of radiation, resulting in a linear dose-response model (see
Figure 1A). Translated to motor performance, this would mean that the performance would decline as soon as minimal exposure to any
stressor occurs. Consequently, ideal motor performance would only occur in the complete absence of stressors.
As exclusively and immediately negative effects of any stressor rarely occur, the linear dose-response model has been
extended by a threshold model (see Figure 1B). This model implies that the negative impact of a stressor can be resisted at low doses
(see Figure 1B
10,12
). Still, once the resistance threshold has been exceeded, increasing doses cause linear declines in the organism’s
function. These models are reminiscent of the diathesis-stress models common in clinical psychology. Specifically, Meehl
13
proposed
that schizophrenia originates from the interaction between a person and the stressors in the environment. When an individual has been
exposed to too high doses of (various) stressors and the threshold for the resistance has been exceeded, the psychopathological
patterns begin to emerge. The threshold varies for each person depending on the presence of certain risk and protective factors. The
more risk factors are present, the more the threshold moves to the left of the model (Figure 1B) and increasingly smaller doses suffice to
trigger the psychopathological responses. In contrast, the presence of protective factors that help an individual to buffer against stressors
moves the threshold to higher doses of stressors. From this perspective, motor tasks can be performed optimally until the dose of a
stressor exceeds the compensatory capabilities of a person. For example, Nieuwenhuys and Oudejans
6,7
postulate that extra mental
effort can be exerted to buffer a stressor and ensure high levels of motor performance. However, once the dose of a stressor exceeds the
compensatory limit, declines in motor performance start to occur.
Figure 1. Linear dose-response (A) and threshold model (B). The solid black line represents the organism’s response (e.g., motor performance) with
increasing dose. The grey dotted line represents the baseline response at 0 loading.
While the linear dose-response and the threshold model argue that a near absence of stressors is required for optimal motor
performance, there is a substantial body of work that shows that stressors can also have beneficial effects
11,14-18
. For example, the
seminal work by Lazarus and Folkman
19
illustrates that the perception of being able to successfully adapt to a stressor can elicit
facilitative behavioral responses (i.e., challenge appraisal, or eustress). Negative responses only emerge if the demands of the stressor
exceed the perceived ability to cope (i.e., threat appraisal or distress). This means that small doses of a stressor can trigger facilitative
psychological responses, while the negative impact only starts to emerge at larger doses
20-22
.
A similar idea can be found in the training literature. A training session may be regarded as a stressor
23
that aims to stimulate
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
272 of 281
functional adaptations over time
24,25
. For athletes, the training load is typically aimed to enhance performance without risking injuries
26
.
Thus, the athletes are exposed to a specific dose of a stressor that triggers adaptive responses while avoiding adverse effects (e.g.,
injuries or overtraining) from doses that are too high. The principle of low-dose stimulation and high-dose inhibition is commonly known in
the domain of biology as hormesis
1-3,10-12,14,15
. Hormesis relates to a J-shaped (i.e., bi-phasic) dose-response pattern with visible
improvements at low doses before the pattern is reversed and increasing doses begin to impair the organism’s response (see Figure 2).
Figure 2. Illustration of a dose-response profile resembling hormesis. The solid black line demonstrates the system’s response with an increasing dose
and the dotted grey line represents the baseline functioning for the system. The system’s response (e.g., motor performance) increases with increasing
doses before the pattern is reversed and the response begins to decline.
Scaling Dose-Response Profiles to Motor Performance
Following the example from biology, comprehensive dose-response profiles need to be established to identify the interaction
between an observed motor performance and a given stressor. The two key parameters consist of the stressor, which can be
systematically varied, and the relevant, measurable response (see for a detailed description of the variable selection, Kiefer et al.
3
and
Kiefer & Martin
9
). The order in which the key parameters are determined is not necessarily fixed. Whether the response variable or the
stressor is determined first depends on the main interest of the researchers. That is, if a researcher is interested in the specific effects of
a stressor on motor performance, then the stressor is defined first. If, however, the main interest is placed on a specific motor
performance, the relevant stressors to induce changes in this behavior are selected in the second step. In any case, the response
variable should represent a relevant adaptation to the stressor
9
. Note that many physiological and psychological responses likely scale
linearly with increasing doses due to homeostatic control mechanisms. For example, exercise intensity would induce linear increases of
heart rate, muscle activity, skin temperature, or perceived exertion. Conversely, the nonlinear patterns associated with hormesis are
thought to emerge from the interactions among these subsystems at a higher level of organization, such as behavioral performance
outputs
27-29
.
In an initial attempt to examine dose-response profiles for human performance, Hill and colleagues
2
instructed climbers to
complete bouldering routes with increasing levels of difficulty. The stressor was operationalized by the degree of difficulty for each route
and the response variable reflected the degree of route completion. However, this approach revealed two limitations to be addressed in
future dose-response profile research and protocols. First, the degree of route completion can quickly reach a ceiling effect. That is, if a
route was completed with the first attempt, the maximum score was obtained, which left no room for further improvement. Second, the
dose of the stressor was similar for everyone. Given that the skill level may vary, adaptative processes are likely to be idiosyncratic
30,31
.
That is, the ideal dose of a stressor varies among individuals. Therefore, the dose range of the stressor should be tailored to the
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
273 of 281
individual, and a suitable motor task should avoid ceiling effects of the performance output and allow for personalized doses of the
stressor.
EMPIRICAL EXAMPLE OF CREW ROWING
To demonstrate how dose-response profiles can be scaled to human motor performance, we provide a tutorial using the
example of crew rowing. Crew rowing is often quoted as an archetypical example of team performance, because of the high degree of
interpersonal coordination that rowing in the same boat requires. Rowers strive to row in perfect synchronization, as this is deemed to
minimize detrimental boat movements that cause additional drag, thereby optimally transferring the power that rowers produce into boat
velocity
32,33
. Rowing in perfect synchronization can be challenging, as the movements of each rowers set the boat, and consequently
their team members, in motion
34,35
. Moreover, the rowing movement consists of two phases: the drive, during which rowers push off with
their blades in the water, and the recovery, during which they balance their blades above the water, returning to their initial position. As
striving for perfect synchronization and balance at high stroke rates (e.g., 30 strokes per minute – spm) and maximal exertion is
challenging, even for elite rowers, coordination as a response variable is not likely to suffer ceiling effects.
In addition, crew rowing allows for individualized approaches of inducing varying stress doses. While the rowing crew ultimately
performs as a single unit, the unique performance and contribution of each member can be identified. For example, on the water, a
stronger rower may apply more power to their blade than a relatively weaker crew member. However, because the crewmembers
coordinate and mutually influence each others’s actions
34
, the total duration of the individual rowing strokes of each crewmember remain
approximately the same—similar to the reciprocal compensation of an interpersonal synergy
36
. In other words, individual variation in
terms of power still allows for synchronization of their strokes. This means that the power input of each crewmember can be
personalized, while allowing for an examination of the coordination as a superordinate performance of the crew.
Given the bidirectional coupling of the individual input and superordinate crew performance, the personalized variation for each
crewmember can be inferred from the crew performance. Specifically, each rowing crew achieves different personal bests, such as the
time for a specific distance. Because these target times emerge from unique contributions of each crewmember, these outcome variables
can be used to systematically vary stressors around them. The dose may then be either increased or decreased by reducing or
increasing the target time, respectively. Therefore, rowing represents a suitable motor performance task for assessing dose-response
relationships. Coordination as a response variable allows for sufficient variation under different doses even in elite performers, while
stressors can be personalized to both 1) different crews, and/or 2) the individual rowers in a crew.
Next, we will describe an experimental setup that can be used to create dose-response profiles for crew rowing. Specifically,
we will determine how we personalized the dose of the stressor (i.e., target times) to observe systematic changes in the response
variable (i.e., crew coordination) alongside some suitable manipulation checks. Furthermore, we outline the necessary steps to translate
the raw data into comprehensive dose-response profiles. Finally, we present possible interpretations of the data in line with the three
dominant dose-response models (i.e., linear, threshold, or hormesis
10,12
).
Participants
For this protocol study, we recruited three rowing crews from student rowing clubs. The crews consisted of two members who
had been competing together at the national level for at least one year. These high-level athletes were chosen to ensure that individual
rowing skill did not interfere with the quality of crew coordination, and to therefore minimize potential inferences of the dose-response
relationship. One crew consisted of a female dyad, while the other two consisted of male dyads (age 22 years, SD = 2.38). All
participants provided written consent to participate in the study.
Material & Setup
The dyads rowed on two ergometers (Type D, Concept2, USA), which were placed side-by-side facing a large computer screen
(see Figure 3). The drag factor of the ergometers’ flywheels was set to 120 (i.e., drag constant of 1.20*10
-4
kg*m2) for all dyads.
Combining the power output from each individual rower into a single team performance, the computer screen displayed how much
distance the dyad had covered together and a timer displayed the current duration of a given trial. We positioned a camera (GoPro7
Black Hero, GoPro, Inc., USA) on the ceiling above the ergometers to capture the rowers’ movements from a perpendicular perspective.
In order to minimize the radial distortion (i.e., fisheye effect) of the GoPro7 Black Hero, we chose for the “linear” recording mode, which
utilizes a 90-degree angle. Furthermore, we placed the camera exactly between the two ergometers to ensure that if distortions were still
present, they would be as equal as possible between the two rowers. During the trials, the rowers wore heart rate monitors (Polar H10
chest belt, Polar Electro, Finland) that were wirelessly connected to the ergometers.
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
274 of 281
Figure 3. Set up of the experiment. Two ergometers with individual monitors displaying heart rate and stroke rate stand next to each other in front of a
screen. Above them, a camera is placed to track the movement of the athletes.
Dose Personalization
As described above, we manipulated varied race times for a fixed distance (i.e., dose) to observe changes in the crew
coordination (i.e., response variable). We opted for race time at a fixed stroke rate (i.e., 30 spm). Note that a faster race time at the same
stroke rate implies a higher power output (i.e., more force over time, and thus more force per stroke as stroke duration remains the
same). The dose of the stressor (i.e., race time) was varied by either decreasing or increasing the time in which the dyads had to
complete a 500 m distance. This distance was chosen, so that five trials with different target times could be performed in one session.
The resulting overall physical demand of this study was close to a standard training session. Researchers should make sure that not only
single doses are within realistic (and ethical) boundaries, but also that the entire measurement protocol does not exceed these
boundaries. On the practical side, the number of trials may be increased to create even more detailed profiles. However, if the data is
supposed to be collected within one session, increasing the number of trials would force a reduction in the distances per trial, which may
lead to insufficient data to reliably determine the response variable during each trial.
The range of target times was personalized for each dyad based on the athletes’ resting and maximum heart rates. In case one
of these values were unknown to the athletes, we measured the resting heart rate before the experiment and approximated the maximum
heart rate by using the Karvonen formula:
!
where HRmax
Appr
represents the approximated maximum heart rate in beats per minute and age represents the rower’s age in years.
Note, however, that this approximation may lead to inaccurate values for individuals. Therefore, we recommend testing maximum
heartrate to determine accurate parameters. Based on the resting and maximum heart rate, we determined the 70% heart rate for each
athlete, given by:
where 70%HR represents the 70% heart rate, HR
rest
represents the resting heart rate, and HR
max
represents the (approximated)
maximum heart rate capacity. Training at the 70%HR can be classified as moderate intensity and therefore leaves room for both
decreases and increases to build a range from small to large doses
37
. To establish a suitable target time for the moderate dose, we
asked the dyads to complete a 500 m baseline trial where they were asked to achieve their individual 70%HR at a stroke rate of 30 spm.
The dyad’s time for the baseline trial was used to calculate the target times (i.e., dose) for each trial.
Specifically, two trials were conducted with smaller doses than the baseline trial, another two trials were conducted with larger
doses than the baseline trial, and one trial was equal to the baseline trial. The time interval for each dose variation was set to 5 seconds.
𝐻𝑅𝑚𝑎𝑥
𝐴𝑝𝑝𝑟
= *220 − 𝑎𝑔𝑒 1
70%𝐻𝑅 = ' 𝐻𝑅
𝑟𝑒𝑠𝑡
+'.7 ∗ (𝐻𝑅
𝑚𝑎𝑥
−'𝐻𝑅
𝑟𝑒𝑠𝑡
) 1
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
275 of 281
Thus, the smallest dose was 10 seconds slower, the medium-small dose was 5 seconds slower, the medium dose equaled the baseline,
the medium-high dose was 5 seconds faster, and the high dose was 10 seconds faster
a
. For example, if a dyad completed the baseline
trial in 1 minute and 50 seconds, the target times for five experimental trials, order from lowest to highest loading, would be 2 minutes 0
seconds (i.e., +10 seconds), 1 minute 55 seconds (i.e., +5 seconds), 1 minute 50 seconds (i.e., baseline time), 1 minute 45 seconds (i.e.,
-5 seconds), and 1 minute 40 seconds (-10 seconds). The sequence of the trials was counterbalanced based on two rules to avoid
ordering effects, such as fatigue. First, the two higher loading trials could not immediately follow each other and second, the third trial
always reflected the baseline trial. We recommend implementing manipulation checks to assess whether the order of the objectively
determined doses indeed aligns with the physiological responses and perceptions of the participants. Here, we assessed a physiological
indicator (i.e., heart rate) during each trial and a behavioral (i.e., the achieved time for a trial) as well as a psychological indicator (i.e.,
rate of perceived exertion, RPE
39
) following each trial.
Procedure
The study procedure was approved by a university-based ethical committee. Upon arriving at the lab facilities, the rowers were
informed about the study procedures, signed the informed consent, and were asked to change into their sportswear while putting on the
heart rate monitor in a separate room. Thereafter, the rowers were asked for their demographic information including their age as well as
the resting and maximum heart rate. During the athletes’ warm-up, the 70% HR was computed for the baseline trial and the specific order
of their trials was determined. Before the baseline trial, the rowers were informed that they should complete each trial with a constant
stroke rate of 30 spm. The stroke rate was always visible to each rower in the ergometer’s monitor. Furthermore, they were instructed to
synchronize their strokes as closely as possible with their respective crew member during each trial. However, during the baseline trial,
the major focus was placed on reaching the respective heart rate at 30 spm rather than coordinating with their crew member. Immediately
after each trial, the rowers were asked to indicate their RPE score. Then, the athletes were given sufficient rest and indicated themselves
when they would be ready for the following trial to begin. Once all experimental trials were completed, the rowers were given the
opportunity to cool down using the ergometers before they could change and return the heart rate monitors.
Measures
RPE. RPE was measured using the Borg
39
scale containing a single item ranging from 6 (no exertion; sitting, resting) to 20
(maximum exertion). The corresponding value reflects a psychological assessment of how intense a person rates a specific task.
Heart Rate. To obtain a physiological measure of the exertion, we measured the beats per minute (bpm) with a chest belt heart
rate monitor during each trial at a frequency of 1 Hz. Given that the heart rate changes during performance, we averaged the heart rate of
each individual per trial.
Position Data. The camera filmed the performance of the rowers at a frequency of 120 Hz at a resolution of 1080 p. Using a
motion tracking software (i.e., Kinovea, Version 0.8.51), we tracked the image-by-image handle position in pixels for both rowers in each
trial, yielding a two-dimensional movement profile. Particularly, we used the y-axis position data representing the forward-backward
oscillation on the ergometers. The oscillation movement of the participants translate into continuous cyclic movement data, which
unfolded over time resembling a sinusoid (wave-like) function.
Next to determining quality of crew coordination, the timeseries of the position data can be used to determine the actual time
that the dyads required to complete each trial. This information can serve as an additional manipulation check to identify whether there
were any discrepancies between the intended and the actual performance.
Data analysis
The first step of the data analysis was to assess whether the objectively determined dose variation was adhered to and whether
it induced expected physiological and psychological responses (i.e., manipulation checks). This means that we analyzed whether the
target time was achieved by the dyads and whether these objectively determined doses aligned with psychological (i.e., RPE) and
physiological (i.e., heart rate) measures of exertion. Specifically, we used Spearman’s rho to calculate whether the rank order of the RPE
and heart rate of each participant aligns with the increasing doses of the trials.
In the next step, the dose-response profiles were established for each crew. That is, the stability of the crew coordination was
mapped as a function of the target times for both the catch and the finish for each crew. As the rowing cycle deviates from perfect
harmonicity at lower stroke rates, we determined variability of crew coordination based on the discrete measure of relative phase that is
not sensitive to within-cycle harmonicities
34,35
. Accordingly, we created two dose-response profiles for each dyad corresponding to the
catch (i.e., start of the drive phase) and the finish (i.e., start of the recovery phase). This way, we assessed whether the resulting dose-
response model differs for different salient moments in the movement cycle.
The timeseries of the forward-backward position of the rowers were analyzed using customized procedures in MATLAB
(MathWorks, USA). The time-series were interpolated using a piecewise spline and were low-pass filtered using a bi-directional second
!
a
Alternative manipulations may represent force output, stroke frequencies, or psychological challenges, such as opponents catching up or overtaking
the team
32,38
.
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
276 of 281
order Butterworth filter with a cut-off frequency of 4 Hz
32
. For further analysis, steady state bins of 40 cycles were determined for each
time trial. This was done by selecting 40 cycles in which the rowers consistently achieved the requested stroke rate (note that it typically
takes the rowers a few strokes to get the flywheel in motion and reach the instructed stroke rate). As relative phase analysis results in a
number of data points per cycle instead of per time unit, it is important to compare an equal number of cycles when investigating
differences in coordination between stroke rates. For instance, at 30 spm, a trial of 1 minute 20 seconds (plus some extra strokes to
reach the intended stroke rate) would be sufficient to obtain 40 cycles, but for 20 spm, a minimum of 2 minutes 0 seconds would be
required. The discrete measure of relative phase based on point-estimates of oar angle extrema near the catch and finish of the stroke
was calculated for each full cycle as:
!
in which t
1,j
and t
2,j
indicate the time of the j
th
peak of the forward-backward position of rower 1 and 2. The instances of catch and finish
were determined as the minimum (i.e., catch) and maximum (i.e., finish) excursions of the signal using a custom-made peak-finding
algorithm. For each trial, standard deviations (SD) of discrete relative phase (SDϕ
catch
and SDϕ
finish
, for catch and finish, respectively)
were calculated as measures of coordinative performance.
Because the response variable is typically displayed in terms of the relative changes to the lowest observed dose (i.e.,
baseline, see Figures 1 and 2), we standardized the SD of the discrete relative phase scores per trial for each crew:
where SDϕ
i
refers to the standard deviation of relative phase in each trial and SDϕ
lowest
refers to the standard deviation of relative phase
corresponding to the trial with the lowest dose (i.e., +10 seconds). This means that the response variable at the lowest observed dose is
equal to 0 and the values corresponding to increasing doses correspond to relative changes of the value at the lowest dose. Note that
this formula also reverses the relative values so that a potential improvement in stability of coordination corresponding to a lower SDϕ is
displayed as an increase above 0 in the dose-response profile, while a decrease in stability of coordination falls below 0 (see Figure 4).
This standardization eases the interpretation of the emerging dose-response model. That is, if an improvement of the response variable
reaches 30-60% relative to the lowest observed dose, the dose-response relationship would reflect hormesis
1
. Minor variations (+/-10%)
around the baseline before a linear decline with increasing doses reflects a threshold response. Finally, an immediate, monotone decline
of the response variable suggests a linear decline model.
Figure 4. Example of a transformation from raw values of the response variable (i.e., SD relative phase) by trial to values standardized relative to the
lowest observed dose marked by the red dot. Note that the formula inverts the values to represent an improvement in coordination as a positive value on
the graph and a decline in coordination as a negative value.
RESULTS
The crews completed the 500 m races within the target time in 13 out of 15 trials (86.67%). One dyad completed the lowest
dose trial 10 seconds faster than intended, while another dyad needed 4 seconds longer for the highest loading trial (see Figure 5A).
Therefore, the crews seemed to be able to adhere to the intended doses for most trials. The rank-order correlation between the order of
the objectively determined doses (i.e., trials) and both RPE (r
s
= .89, p < .01) and heart rate (r
s
= .46, p = .01) yielded a significant result.
𝜙
𝑃𝐸𝑖
(𝑡) =
𝑡
2,𝑗
− 𝑡
1,𝑗
𝑡
2,𝑗 +1
− 𝑡
2,𝑗
360 1
𝑆𝑆𝐷ɸ = 1 −'
𝑆𝐷ɸ
𝑖
𝑆𝐷ɸ
𝑙𝑜𝑤𝑒𝑠𝑡
'' 1
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
277 of 281
This suggests that the sequence of the objectively determined doses, on average, corresponded to the desired psychological and
physiological responses. However, the graphical illustrations also point out that considerable inter- and intra-individual variability may
occur (see Figure 5B and 5C), which is in line with previous findings on stress-response dynamics
30,31
. For example, Rower 1 of Dyad 1
shows considerable variation in heart rate for the different trials. This variation may have occurred because the maximum heart rate was
estimated for this rower, reinforcing our suggestion to explicitly test the maximum heart rate to increase the accuracy of the personalized
for the dose variation.
Figure 5. The target (triangular markers) and achieved race times (solid lines) of the rowing crews (A), the heart rate of each rower (B), and the rate of
perceived exertion for each rower (C) across the trials ordered from lowest to highest loading. The solid line represents rower 1 and the dotted line
represents rower 2 in (B) and (C). Note that in (C) the values for Dyad 1 are exactly the same for each rower and the lines therefore overlap.
The resulting dose-response profiles are depicted in Figure 6. Dyads 1 and 3 display similar patterns. That is, during the catch
and the finish, the response variable seems to initially increase with the dose before the pattern is reversed and further increasing doses
cause a decline. This means that the coordination of these dyads improved with increasing doses (i.e., target times) before dropping
again with further dose increments. This low-dose stimulation and high-dose inhibition is indicative of hormesis. Additionally, the highest
response in both the catch and the finish is more than 30% larger than the baseline. This further supports the hormetic dose-response
models
1
even though the patterns of the catch and the finish do not follow the exact same trend per dyad. Note that this means that Dyad
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
278 of 281
1 interestingly demonstrates hormesis despite the large variation of one of the rowers in the average heart rate per trial. Thus, the
variation of the target time seems to be sufficient to create hormetic dose-response patterns even if the underlying homeostatic control
mechanisms do not change according to what may be expected.
Dyad 2 on the other hand does not follow a clear pattern that could be characterized according to the three dominant dose-
response models. The catch seems to display an immediate negative decline in the response variable with increasing dose, indicative of
a linear dose-response model. However, the response variable does not show any additional declines with increasing doses. Moreover,
the finish follows a similar initial pattern, but then begins to show signs of hormesis with an increase of the response variable of 25.05%
relative to the lowest observed dose at the baseline trial with steady decreases thereafter. All in all, this means that although crew
coordination shows signs of hormesis in response to race time variations, the exact relationship on the individual level may unfold in
idiosyncratic patterns.
Figure 6. The dose-response profiles for the catch and the finish of each dyad ordered from lowest to highest dose. The response variable (i.e., SD
relative phase) is standardized relative to the score at the lowest dose. This value is represented by the black dotted line.
DISCUSSION
Because of the ubiquitous presence of stressors, it is essential to understand the specific dose-response dynamics between
stressors and motor performance. That is, while several models acknowledge the dose-dependent impact of stressors on motor
performance, precise modeling of this relationship is often lacking. Developing a thorough understanding of the doses-response
dynamics at the level of motor performance may be leveraged for various practical applications, such as precision medicine and
rehabilitation
9,40
. Therefore, the aim of this paper was to provide specific tools which can help to fill this void. To do so, we provided an
overview of the most common dose-response models that exist across various disciplines as well as a detailed account on how to
translate the approach to identify these models (i.e., dose-response profiles) to motor performance. While this approach may help us to
develop research within the domain of motor performance, it also enables communicating findings from between various domains that
follow the same basic principles. Thereby, new cross-disciplinary research on dose-response dynamics may emerge.
Theoretical and Practical Implications
The dose-response profiles indicate that coordination of crew rowing likely follows the typical patterns of hormesis. This implies
that large deviations of target times at constant stroke rates in either direction (i.e., increases or decreases) can interfere with the desired
behavior. Because of the systematic variation of the stressor, it is possible to pinpoint the ideal dose for each individual dya
d 2,3,9
. These
insights can be translated to training regimes that build on this ideal pattern while avoiding training in regions that trigger undesirabale
behavior. Furthermore, dose-response profiles may be leveraged to develop and monitor tailored trainings that aim at expanding the
zone at which the maximally efficient coordination behavior can be exerted
3
.
Note that, while the dose-response profiles comprehensively depict the dynamics between a stressor and the response
variable, it may not be generalizable to related behaviors. For example, varying race time at constant stroke rates may cause different
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
279 of 281
patterns for different response variables. Similarly, systematically altering other stressors (e.g., changing stroke rates at constant speed)
may also elicit different dose-responses in crew coordination. Therefore, a response variable should be carefully selected to ensure that it
is meaningful in the context of performance, while the specific dynamics between two assumingly related variables also need to be tested
explicitly and should not be inferred from related dynamics. Accordingly, it is important to note that rowing on ergometers differs from
rowing on water
41
. While rowing on an ergometer is a suitable way of training rowing specific elements, it cannot replace on-water
training. One important difference is that on-water, rowers share the same boat, influencing each other movements not only perceptually
(e.g., seeing each other movements, hearing the sound of the blades entering the water/the sounds of the flywheel) but also
mechanically
34,35
. On the individual ergometers, the stroke rate of an individual rower depends on the force a rower applies to the
flywheel over time: the resistance of the individual flywheel remains at a constant. On the water, the pressure on the blade of an
individual rower is influenced by the pressure that their team members apply to their blades. If they apply more pressure, the individual
rower will move his/her blade faster through the water than if they apply less pressure, resulting in a higher stroke rate. In other words, on
the water the stroke rate is influenced by all rowers in the team, in contrast to rowing on individual ergometers
41
. Although there are ways
to mimic on-water rowing on an ergometer as much as possible (e.g., by connecting ergometers through slides to mimic the mechanical
coupling through the ‘boat’
34,35
), it remains important to account for the task constraints of the performance context when choosing a
response variable.
Due to the likely inter- and intra-individual variability of the dose-response dynamics
2,3,30,31
, it may not always be possible to
derive a single dose-response model that fits all participants and, in many instances, it may not be appropriate. As illustrated in the
accompanying empirical data, one rowing crew demonstrated a pattern that matched none of the popular models. While establishing
these profiles can be an important step towards acknowledging this idiosyncrasy, it may not be possible to derive uniform interventions or
training regimes. Furthermore, it should be considered whether a single session suffices for the assessment of dose-response dynamics
9
. That is, emerging patterns that coincide with specific dose-response models as well as diverting behavior may simply be due to natural
variations of behavior. Therefore, repeated assessments to increase reliable profiles is advised. This is especially important because
trainings may be specifically designed to alter the dose-response dynamics over time
3,9
.
Finally, the dose-response profiles are typically constructed relative to the lowest observed dose. In the example of medical
research, the lowest dose may simply represent the absence of a specific medication. However, motor performance by definition cannot
take place without any exertion
39
. Thus, a rowing crew cannot row without moving at all, even when they are not aiming at a specific
target time. Therefore, we first determined the medium dose and created both ascending and descending dose variations around it.
Future studies should carefully consider what level of exertion can be considered the lowest dose to create meaningful profiles.
CONCLUSION
Given the omnipresence and idiosyncrasy of stressors, establishing comprehensive behavior-based dose-response profiles are
essential to inform both theory-building and developing practical applications. We have provided a detailed instructional map for
translating common approaches from biology and medicine to motor performance research alongside initial empirical data from crew
rowing. This approach can not only improve research within the domain of motor performance, but also enable cross-disciplinary
research and communication between different fields under a unified approach.
REFERENCES
1. Calabrese EJ, Mattson, MP. Hormesis provides a generalized quantitative estimate of biological plasticity. J. Cell Commun. Signal. 2011;5:25-38.
doi: 10.1007/ s12079-011-0119-1
2. Hill Y, Kiefer AW, Silva PL, Van Yperen NW, Meijer RR, Den Hartigh RJR. Antifragility in climbing: Determining optimal stress loads for athletic
performance training. Front. Psychol., 2020;11:272. doi: 10.3389/fpsyg.2020.00272
3. Kiefer AW, Silva PL, Harrison HS, Araújo D. Antifragility in sport: Leveraging adversity to enhance performance. Sport Exerc. Perform. Psychol.
2018;7(4):342–350. doi: 10.1037/spy0000130
4. Eysenck MW, Derakshan N, Santos R, Calvo MG. Anxiety and cognitive performance: Attentional control theory. Emotion, 2007;7(2):336-353. doi:
10.1037/1528-3542.7.2.336
5. Moore LJ, Vine SJ, Wilson MR. Attention and visuomotor performance under pressure. In: Arnold R, Fletcher D, editors. Stress, well-being,
andperformance in sport. Routledge; 2021. p.113-130. doi: 10.4324/9780429295874-8
6. Nieuwenhuys A, Oudejans RRD. Anxiety and perceptual-motor performance: toward an integrated model of concepts, mechanisms, and processes.
Psychol. Res. 2012;76(6):747-759. doi: 10.1007/s00426-011-0384-x
7. Nieuwenhuys A, Oudejans RRD. Anxiety and performance: Perceptual-motor behavior in high-pressure contexts. Curr. Opin. Psychol. 2017;16:28-
33. doi: 10.1016/j.copsyc.2017.03.019
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
280 of 281
8. Blascovich J. Challenge, threat, and health. In: Shah JY, Gardner WL, editors. Handbook of motivation science. The Guilford Press; 2008. p. 481-
493
9. Kiefer AW, Martin DT. Phenomics in sport: Can emerging methodology drive advanced insights? Front. Net. Physiol. 2022;2:1060858. doi:
10.3389/fnetp.2022.1060858
10. Agathokleous E, Kitao M, Calabrese EJ. Environmental hormesis and its fundamental biological basis: rewriting the history of toxicology. Environ.
Res. 2018;165:274-278. doi: 10.1016/j.envres.2018.04.034
11. Calabrese EJ. Hormetic dose-response relationships in immunology: Occurrence, quantitative features of the dose response, mechanistic
foundations, and clinical implications. Crit. Rev. Toxicol. 2005;35:89-295. doi: 10.1080/10408440590917044
12. Costantini D, Metcalfe NB, Monaghan P. Ecological processes in a hormetic framework. Ecol. Lett. 2010;13:1435-1447. doi: 10.1111/j.1461-
0248.2010.01531.x
13. Meehl PE. Schizotaxia, schizotypy, schizophrenia. American Psychologist. 1962;17(12):827–838. doi: 10.1037/h0041029
14. Calabrese EJ. Evidence that hormesis represents an “overcompensation” response to a disruption in homeostasis. Ecotoxicol. Environ. Saf.
1999;42:135-137. doi: 10.1006/eesa.1998.1729
15. Calabrese EJ. Overcompensation stimulation: a mechanism for hormetic effects. Crit. Rev. Toxicol., 2001;31:425-470. doi:
10.1080/20014091111749
16. Calabrese EJ, Blain RB. The hormesis database: the occurrence of hormetic dose responses in the toxicological literature. Regul. Toxicol.
Pharmacol., 2011;61:73-81. doi: 10.1016/j.yrtph.2011.06.003
17. Mattson MP, Calabrese EJ. Hormesis: A revolution in biology, toxicology and medicine. Human Press; 2010. 213 p. doi: 10.1007/978-1-60761-495-1
18. Southam CM, Ehrlich J. Effects of exact of western red-dedar heartwood on certain wood-decaying fungi in culture. Phytopathol. 1943;33:517–524.
19. Lazarus RS, Folkman S. Stress, appraisal, and coping. Springer; 1984. 456 p.
20. Kerr JH. Rethinking aggression and violence in sport. Routledge; 2005. 176 p.
21. Seery MD, Holman EA, Silver RC. Whatever does not kill us: Cumulative lifetime adversity, vulnerability, and resilience. J. Pers. Soc. Psychol.,
2010;99(6):1025–1041. doi: 10.1037/a0021344
22. Voigt L, Hill Y, Frenkel MO. Testing the hormesis hypothesis on motor behavior under stress. Applied Ergonomics. 2024; 115:104161.
https://doi.org/10.1016/j.apergo.2023.104161
23. Den Hartigh RJR, Meerhoff LRA, Van Yperen NW, Neumann ND, Brauers JJ, Frencken WG, ... & Brink MS. Resilience in sports: a multidisciplinary,
dynamic, and personalized perspective. Int. Rev. Sport Exerc. Psychol. 2022;1-23. doi: 10.1080/1750984X.2022.2039749
24. Baetzner AS, Wespi R, Hill Y, Gyllencreutz L, Sauter TC, Saveman BI, Mohr S, Regal G, Wrzus C, Frenkel MO. Preparing medical first responders
for crises: a systematic literature review of disaster training programs and their effectiveness. Scand. J. Trauma Resusc. Emerg. Med. 2022;30:76.
doi: 10.1186/s13049-022-01056-8
25. Den Hartigh RJR, Hill Y. Conceptualizing and measuring psychological resilience: What can we learn from physics?. New Ideas in Psychol.
2022;66:100934. doi: 10.1016/j.newideapsych.2022.100934
26. Brink MS, Visscher C, Arends S, Zwerver J, Post WJ, Lemmink KA. Monitoring stress and recovery: New insights for the prevention of injuries and
illnesses in elite youth soccer players. Br. J. Sports Med. 2010;44(11):809–815. doi: 10.1136/bjsm.2009.069476
27. Balagué N, Hristovski R, Almarcha MDC, Garcia-Retortillo S, Ivanov PC. Network physiology of exercise: Vision and perspectives. Front. Physiol.
2020;11:611550. doi: 10.3389/fphys.2020.611550
28. Diniz A, Wijnants ML, Torre K, Barreiros J, Crato N, Bosman AM, ... Delignières D. Contemporary theories of 1/f noise in motor control. Hum. Mov.
Sci. 2011;30(5):889-905. doi: 10.1016/j.humov.2010.07.006
29. Van Orden GC, Kloos H, Wallot S. Living in the pink: Intentionality, wellbeing, and complexity. In: Hooker C, editor. Philosophy of complex systems:
Handbook of the philosophy of science (Vol. 10). Elsevier; 2011. p. 629-672. doi: 10.1016/B978-0-444-52076-0.50022-5
30. Hill Y, Meijer RR, Van Yperen NW, Michelakis G, Barisch S, Den Hartigh RJR. Nonergodicity in protective factors of resilience in athletes. Sport
Exerc. Perform. Psychol. 2021;10(2):217–223. doi: 10.1037/spy0000246
31. Neumann ND, Van Yperen NW, Brauers JJ, Frencken W, Brink MS, Lemmink KAPM, Meerhoff LA, Den Hartigh RJR. Non-ergodicity in load and
recovery: Group results do not generalize to individuals. Int. J. Sports Physiol. Perform. 2021;1–9. doi: 10.1123/ijspp.2021-0126
32. Cuijpers LS, Passos PJM, Murgia A, Hoogerheide A, Lemmink KAPM, De Poel HJ. Rocking the boat: does perfect rowing crew synchronization
reduce detrimental boat movements?. Scand. J. Med. Sci. Sports 2017;27(12):1697-1704. doi: 10.1111/sms.12800
33. Hill H, Fahrig S. The impact of fluctuations in boat velocity during the rowing cycle on race time. Scand. J. Med. Sci. Sports 2009;19(4):585-594. doi:
10.1111/j.1600-0838.2008.00819.x
34. Cuijpers LS, Den Hartigh RJR, Zaal FT, de Poel HJ. Rowing together: Interpersonal coordination dynamics with and without mechanical coupling.
Hum. Mov. Sci. 2019;64:38-46. doi: 10.1016/j.humov.2018.12.008
35. Cuijpers LS, Zaal FT, de Poel HJ. Rowing crew coordination dynamics at increasing stroke rates. PloS one, 2015;10(7):e0133527. doi:
10.1371/journal.pone.0133527
36. Riley MA, Richardson MJ, Shockley K, Ramenzoni VC. Interpersonal synergies. Front. Psychol., 2011;2:38. doi: 10.3389/fpsyg. 2011.00038
BJMB! ! ! ! ! ! ! ! ! ! ! !!!!!Tutorials
Brazilian(Journal(of(Motor(Behavior(
(
Hill et al.
2023
VOL.17
N.6
281 of 281
37. Seiler KS, Kjerland GØ. Quantifying training intensity distribution in elite endurance athletes: is there evidence for an “optimal” distribution?. Scand.
J. Med. Sci. Sports 2006;16(1):49-56. doi: 10.1111/j.1600-0838.2004.00418.x
38. Den Hartigh, RJR, Van Geert PL, Van Yperen NW, Cox RFA, Gernigon, C. Psychological Momentum During and Across Sports Matches: Evidence
for Interconnected Time Scales. J. Sport Exerc. Psychol. 2016;38:82-92. doi: 10.1123/jsep.2015-0162
39. Borg GA. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc., 1982;14:377–381.
40. Kiefer AW, Armitano-Lago CN, Sathyan A, MacPherson R, Cohen K, Silva PL. The Intelligent Phenotypic Plasticity Platform (IP3) for Precision
Medicine-Based Injury Prevention in Sport. In: Ossandon MR, Baker H, Rasooly A, editors. Biomedical Engineering Technologies. Methods in
Molecular Biology. Humana; 2022. p.877-903. doi: 10.1007/978-1-0716-1803-5_47
41. Cuijpers LS. (2019). Coordination dynamics in crew rowing. Chapter 7 Exploring the potential benefits of antiphase crew rowing on water. University
of Groningen. https://doi.org/10.33612/diss.94906482
ACKNOWLEDGMENTS
YH and LSC contributed equally to this work. We thank Mathijs Hofmijster for their detailed feedback on an early concept draft
of this manuscript and Marie Repgen for their support on the data collection.
Citation: Hill Y, Cuijpers LS, Silva PL, den Hartigh RJR, Kiefer AW. (2023).!How does motor performance change with increasing stress doses? A tutorial on dose-
response profiles applied to crew rowing. Brazilian Journal of Motor Behavior, 17(6):270-281.
Editor-in-chief: Dr Fabio Augusto Barbieri - São Paulo State University (UNESP), Bauru, SP, Brazil. !
Associate editors: Dr José Angelo Barela - São Paulo State University (UNESP), Rio Claro, SP, Brazil; Dr Natalia Madalena Rinaldi - Federal University of Espírito Santo
(UFES), Vitória, ES, Brazil; Dr Renato de Moraes – University of São Paulo (USP), Ribeirão Preto, SP, Brazil.!
Copyright:© 2023 Hill, Cuijpers, Silva, den Hartigh and Kiefer and BJMB. This is an open-access article distributed under the terms of the Creative Commons Attribution-
Non Commercial-No Derivatives 4.0 International License which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and
source are credited.
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interests: The authors have declared that no competing interests exist.
DOI:!https://doi.org/10.20338/bjmb.v17i6.398