BJMB
Brazilian Journal of Motor Behavior
Special issue:
The role of practice in motor learning
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
162 of 178
Learning and performing: What can theory offer high performance sports practitioners?
IAN RENSHAW
1
| KEITH DAVIDS
2
| MARK O’SULLIVAN
2,3
1
School of Exercise & Nutrition Sciences, Faculty of Health, Queensland University of Technology.
2
Sport & Physical Activity Research Centre, Sheffield Hallam University, Sheffield, UK.
3
AIK Football, Research and Development Department, Stockholm, Sweden.
Correspondence to: Ian Renshaw.
School of Exercise & Nutrition Sciences
Queensland University of Technology
Victoria Park Road
Kelvin Grove
4059
email: i.renshaw@qut.edu.au
https://doi.org/10.20338/bjmb.v16i2.280
ABBREVIATIONS
RT Retention test
PUBLICATION DATA
Received 20 12 2021
Accepted 04 05 2022
Published 01 06 2022
ABSTRACT
Currently, the most prominent motor control theories that underpin the pedagogy of coaches in high
performance sport are derived from the discipline of psychology with a dominant focus on internalised control
processes for learning and performance. In contrast, ecological dynamics is a contemporary meta-theory
focused on the person-environment scale of analysis for understanding human behavior, exemplified by
strengthening the relations between each learner and their environment. In this tutorial, we outline key concepts
in ecological dynamics that considers learning and performance as being distinct, yet inextricably linked. In our
considerations, we raise questions on long-held assumptions about control process theories on learning and
performance for practice designs in high performance sports. For example, how useful is inferring learning by
describing improved performance as showing more relative permanence, greater stability and consistency, with
commensurate lower levels of attention and movement variability? How relevant are traditional ways of
measuring learning using retention and transfer tests in high performance sports? What is actually attained in an
ecological view of learning, focussed on education of attention and calibration of actions to specifying
information present in performance environments? An implication of these issues for high performance sport is
that learning needs to be assessed by how well a learner adapts to the specific constraints and demands of a
performance context. This key idea has important implications for performance analysis and evaluation in sport.
KEYWORDS: Learning | Performance | Ecological dynamics | Motor learning theory | Skill adaptability
INTRODUCTION
Currently, the most prominent motor control theories underpinning coaching
pedagogies in high performance sport focus on internalised control processes for learning
and performance
1,2
. These models of control invite coaches to organise their session
intentions around developing ‘robust’ internalised motor programmes and schema,
purported to result in alignment with an optimal movement model. Practising a technique
is, therefore, of paramount importance and generally undertaken via isolated drills with
feedback directed at reducing the gap between what the movement looks like and the
putative ‘ideal’ technical model
3
. However, a somewhat puzzling observation when
watching high performance athletes is that successful performers do not always (in fact
rarely) have what may be considered to be the most biomechanically ‘optimal’ techniques.
In contrast, success in high performance sport appears to be predicated on the ability of
highly skilled performers to excel at learning in performance to quickly exploit opportunities
to coordinate their actions to adapt to what the competitive context offers them, to function
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
163 of 178
more effectively and efficiently. For example, the fastest sailors in a regatta continually
(re)organise their actions, highly attuned to immediate changes in prevailing currents or
winds at any moment; the skillful footballer adapts the weight of their pass to match it to
the demands of a wet surface where the ball ‘skids’ or a dry surface which has greater
friction; the expert cricket spin bowler quickly ‘finds’ the most optimal pace when bowling
on a new pitch, and the skilled ice climber explores and perceives properties of a frozen
waterfall when traversing a route on a rocky surface. It is clear from these examples that a
key part of performing functionally (effectively and efficiently) is learning (quickly) in
competitive performance. Put another way: performing competitively provides an
invaluable context for motor learning through skill adaptation in athletes
4
.
A key question for those involved in coach education and, in particular, increasing
practitioners’ awareness of contemporary models of skill learning, is why coaches continue
to stick with outdated models? Indeed, whilst theories of skill acquisition have been
contemporized in the last 40 years, approaches to practice design across educational and
sporting settings have largely remained unchanged. For example, informal yearly surveys
of our students since our first formal survey in 2014
3
continue to reveal that 95% of
practice in schools is still being delivered using either an atheoretical approach or teaching
methods aligned (if aligned at all) with motor control theories based on cognitive models
published between 1960-2000 (see
5, 6
).
In short, the teaching of skill acquisition on coaching courses appears to be failing
to influence pedagogical practice. It has been suggested that a path dependence, an
ideological inertia, has prolonged the shelf life of some inherited beliefs
7,8
associated with
‘traditional’ approaches to skill acquisition. This inertia to contemporization implies that
they are somewhat sticky with practitioners, who as ‘successful’ products of these
approaches
3
have a strong attachment to pedagogical methods promoted by them. An
important question is how can we help practitioners to develop the skills and knowledge to
change outdated practice ideologies? Alexander
9
suggests that practitioners need to
identify ‘what one needs to know, and the skills one needs to command’ to make and
justify the many different kinds of decisions of which teaching is constituted’ (p.47). The
role of skill acquisition specialists in coach education courses could, therefore, be viewed
as providing practitioners with a ‘familiarity and ease’ with the key ideas underpinning
contemporary theories to use the associated pedagogical methodologies appropriately,
effectively and efficiently’
10
. This tutorial seeks to contribute to that goal. We propose that
in line with ideas of Smith discussed in Ovens et al.
11
, an ecological dynamics approach
offers a ‘plausible’ theoretical lens for supporting context-specific skill-learning practices.
In this tutorial, we explain how some of the ‘givens’ in ‘traditional’ approaches to
skill acquisition are so sticky with practitioners and discuss alternatives moving forward.
Our main vehicle to explore the key issues will be to consider what we understand as
constituting learning in the contrasting approaches and its relationship to performance. We
consider the importance of re-framing the concepts of learning and performing to
encourage sport practitioners to move away from long-held beliefs that underpin current
coaching philosophies, predicated on the dominant view that skill learning is fundamentally
founded on internalised control processes. This ideology is not aligned with contemporary
theoretical models of movement coordination and the acquisition of coordination
12,13,14
or,
even, more contemporary embodied cognitive approaches
15
. We begin by explaining how
traditional beliefs shape practice designs, highlighting the limitations of such approaches
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
164 of 178
given current understanding of skill learning. We next propose an alternate approach
adopting an individual-environment scale of analysis to reframe explanations of skilled
behaviour away from internalised processes
16
.
Teaching skill acquisition in coach education programmes
Despite the central place of skill acquisition in high performance sports coaching, it
receives limited attention in many coach education programmes, typically being delivered
in one two-hour ‘module’ by an academic skill acquisition lecturer from ‘local’ higher
education institutions. Consequently, the key concepts covered tend to revolve around
fundamental “givens” in sports coaching, such as the idea that movement techniques need
to be taught first
5
. This accepted wisdom is predicated on cognitive, enrichment-based
models of skill acquisition with an information-processing lens
17
with the optimization of
motor programs and the development of schema as the major goals. An important concern
for sport practitioners in high performance organisations, therefore, is how these ideas
frame our understanding of learning and the implications for coaching practice design (i.e.,
what coaches’ say and do). Most fundamental of all is to consider whether these ideas are
still valid today.
Learning from a Cognitive Viewpoint
Uncontroversially, Schmidt & Lee (1999, p.264)
18
defined learning as “a process of
acquiring the capability for producing and controlling skilled actions”. Perhaps, more
controversially, they also define this process as a purely internal one
17
. Motor control is
considered to emerge from the skillful implementation of internalised control processes
19
.
The most prominent and popular internal motor control theory is Schema Theory
1
,
countering initial ideas of Jack Adams
20
on closed loop control. Schmidt
1
further explored
the emphasis in Adams' theory on learners re-producing movements in line with an
established reference of correctness
19
. Each movement made in pursuit of this goal during
learning, results in a gradual adaptation of a perceptual trace of that movement in the brain
21
. Learning is, therefore, proposed as a feedback-driven process of internalised trace
strengthening, with skilled performers accumulating a greater number of ‘correct’ traces for
controlling movements. The strengthening of a memory trace is directly proportionate to
the number of practice trials undertaken and the quality of feedback available to learners.
This key idea served to shape Schmidt’s thinking about the memorization processes (recall
and recognition) of a generalised motor program
21
. These dominant conceptualisations of
motor control and learning underpinned ideas around deliberate practice
22
. Variable
practice was proposed as a way to strengthen the schema built on the assumption that the
generalized motor program, which controlled the sequencing and timing of muscle activity,
was already acquired a priori.
The characteristics of learning from a cognitive perspective
In the dominant cognitive paradigm described above, learning is predicated on the
elaboration of ‘internal’ control processes and, therefore, not directly observable, except
through analysing performance repeated over time. Learning has to be indirectly inferred
from characteristics of improved performance, such as showing more relative permanence,
greater stability and consistency, with commensurate lower levels of movement variability
and attention
23
. Relative permanence through greater stability and consistency is assumed
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
165 of 178
to come from lower levels of variability as the motor program and parameterisation of the
schema becomes more aligned to the target movement template. As movements become
more ‘automatic’, less attention needs to be paid to cognitively controlling them
24
. Why
reduced cognitive attention here results in better performance is not quite clear, but ideas
put forward include: (i) no attention needs to be paid to the ‘verbal-cognitive’ aspects of the
skill as now “we know” which pattern to use rather than still having to decide (i.e., in the
cognitive stage of learning). In line with this idea is that control is hierarchical and learning
leads to control moving ‘from ‘higher’ decision-making levels to ‘lower’ motor program
control level. (ii); the sub-routines (smaller composite programs developed in early
learning) are combined as a single more extensive motor program (see Schmidt
25
citing
Keele’s ‘gearshift analogy’).
One other characteristic that learning is thought to be judged upon is the learners’
ability to adapt ‘to a variety of performance context characteristics’
23
. The process or
mechanisms for this increased adaptability is not explained, but interestingly, is said to be
needed because “we never really perform a skill when everything in the performance
context is exactly the same each time” and, therefore, “successful performance requires
adaptability to changes in personal, task and/or environmental characteristics”
23
(p.258).
Note the all-inclusive use of the term ‘we’. This use of language is useful in highlighting
how traditional theories on motor control processes and mechanisms adopt a task-based
approach in which representation formation and feedback implementation seems to be
universal, stage-based processes for all individuals. An individualised approach to
performance and learning is eschewed. What does this signify? This idea does not appear
to fit well with a mode of motor control residing in the construction of common
representational programs. There is little attention paid to consider the influence of specific
situations or contexts, ignoring the mutuality of the individual and the environment
26
. As
such, a cognitive perspective on motor learning and control that adopts an asymmetric
focus on the internal representations of an individual fails to consider how the relationship
between that athlete and a performance environment is strengthened during their
individualised learning and development experiences. Indeed, the suggestion is that key
individual experiences and ‘specific performance characteristics’ such as emotions,
changes in task constraints and the environment, should be removed when learning and
when attempting to assess and analyse learning.
The distinction between learning and performance
A key feature proposed in cognitive models is that, while learning and performing
are distinct processes, they are inter-dependent. Learning cannot emerge without
performance and the latter needs to be assessed over time to evaluate the former. This
results in a failure to consider how the relationship between that athlete and a performance
environment is strengthened during their individualised learning and development
experiences. Learning can lead to behavioural changes, but refers to the set of processes
that support such changes
25
. Research in the field, therefore, became focused on
identifying the variables that best supported learning, especially in the context that learning
could only have said to have taken place if it led to a permanent, rather than a temporary
change in behaviour. Accordingly, any variable that was seen to interfere with learning, or
mediate the effect of practice was to be removed. Unfortunately, this meant that variables
such as emotions, fatigue, or competition and the uniqueness of the setting are said to
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
166 of 178
have the potential to affect a person’s performance, but not the degree of learning the
person has achieved
23
. Some variables (although Schmidt
25
highlights that sometimes it is
hard to distinguish between a learning variable and a performance variable) are, therefore,
irrelevant or detrimental to learning and need to be removed from practice contexts. In
Schmidt’s
25
own words; “If I know a certain variable only affects performance, and it is not
important for learning, then I will not have to worry about adjusting the level of that variable
in the learning situation (p.456-7). For us, this idea is problematic as it fails to take into
account the necessity for learning experiences that emerge when performing in
competitive performance environments where it is likely that individuals will experience
intense emotions or have to ‘deliver’ a performance when fatigued, distracted or
emotionally perturbed. In this definition, performance in competition is not seen as being a
part of the ‘appropriate conditions’ to assess how much learning has taken place
23
. We
argue that learning and performing are symbiotic, with both processes emerging
simultaneously when individuals engage in motor behaviours in specific contexts e.g., a
competition environment
17
. Consequently, when attempting to work with sports
practitioners to prepare athletes to perform in competitive environments, performance
variables cannot be ignored and removed because of the important relationship between
performance and learning. In fact, we suggest that these variables should be sampled from
competitive performance environments and embraced in practice designs as their
presence can result in deeply significant changes in intentions, perceptions and actions to
which an athlete needs to adapt
10
. We will delve further into this later after briefly
considering how we have traditionally attempted to measure motor learning.
Measuring learning: retention and transfer tests
Excluding performance variables may seriously misrepresent the amount of
learning that has taken place for each individual, given the over-reliance on experimental
investigations of learning in motor learning. Also, in experimental investigations there is an
over-reliance on specific methods and data to verify learning such as the need for a
delayed test, to accurately assess learning
25
. A Retention Test (RT) generally requires
learners to repeat the same activities acquired in learning to observe whether changes
inferred from performance, undertaken during practice, have become relatively permanent.
In experiments, cognitive motor control theorists emphasise the need to ‘sanitise’ the
learning environment significantly by controlling potentially intervening variables in the RT
environment to remove perceived contamination that could influence performance
temporarily and reduce effectiveness of the pedagogical setting or the researcher’s
confidence to evaluate the impact of learning
25
. An additional way to test learning is to use
a transfer test to examine the ability of the learner to adapt the skill they have learned to
novel situations, or environments and may include requiring the production of novel
variations of the learned skill. Tests of skill transfer also examine how participants cope
with changes in performance characteristics such as dealing with stressful competition
23
.
Linking back to earlier comments, there is significant research evidence to suggest
that removing performance variables from practice until a skill has been well established
27
would be detrimental to learning. For example, suppressing or removing emotions from
learning designs may actually be impossible, since all practice tasks, irrespective of being
drill based or games based, still create emotions, feelings and thoughts and impact on
participant motivation
28
. Indeed, research has shown that cognitions and emotions are
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
167 of 178
intertwined with actions and consequently act to shape co-ordination
29
. For example, many
studies (e.g.,
30,31,32
) have revealed how differing anxiety levels change intentions of
performers, what they perceive and, consequently, how they move. Importantly,
improvements in movement analysis technology, combined with more advanced methods
of interpreting data, have enabled researchers to capture how so-called performance
variables result in changes in movement co-ordination. For example, as aligned with
Bernstein’s
33
early evidence and insights on challenges of repeating simple aiming
movements in a stereotyped way, Bauer and Schöllhorn (1997)
34
showed how movement
patterns of high performance athletes changed from session to session across
competitions and practice sessions. Further evidence has been found in studies of gait,
which may be considered as an ideal task vehicle to demonstrate stable repeatable motor
programs, given the significant number of trials undertaken over an individual’s life span.
However, the evidence suggests the opposite. While gait patterns are found to be unique,
yet persistent in each individual, remarkably, gait has been found to change by 85-95%
within and between days
35,36
. Further, performance variables such as fatigue and emotions
showed that individual’s gait patterns alter in fatigued conditions
37
Emotional states such
as being happy, sad or angry
38
could also be differentiated through gait analysis. Notably,
a key finding in these studies highlights that optimal movement co-ordination was highly
individual and emphasized that group-based studies have limited value for assessing
learning in high performance sport. In summary, these studies highlight the triadic
relationship between intentions, emotions and actions and suggest there is a need for
practitioners to embrace performance variables rather than attempting to cleanse the
practice environments by removing them. We begin to address how these ideas can inform
the work of practitioners next.
What does all of this mean for the practitioner?
For the sports practitioner who interrogates the motor learning literature, the
separation of learning and performing may, at first, appear somewhat confusing, if not
faintly ridiculous. For coaches and those invested in an individual or team’s performance,
the most important measure of learning concerns how a participant performs in
competition. Performance is a reflection of the coaches’ ability to ‘prepare’ performers for
competition by ‘teaching’ or ‘coaching’ them, so they are capable of demonstrating the
necessary skills in a performance setting. This requirement suggests that non-motor
control specialists may require practice effectiveness to be judged by performance under
the constraints of a competitive environment. Under these specific competitive constraints,
the ability to perform when nervous, fatigued or required to adapt to unique contexts, such
as when facing novel opponents or dynamic defensive formations, or when performing in a
variety of weather conditions is paramount for understanding effects of learning.
To summarise so far, we have highlighted that any model of skill learning to be
utilised to underpin practice design for high performance athletes needs to embrace the
dynamic nature of competitive environments. Successful performance in competition
requires athletes to adapt to dynamic task constraints, often when performing under
intense emotional states induced by contextual events and surroundings, continually
influencing their cognitions, perceptions and actions
39,40
. In the next section we provide an
alternate viewpoint where learning and performing are seen as being tightly interconnected
and in fact, performing requires direct learning and learning take place through performing
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
168 of 178
in representative environments
41
. In this approach, motor learning is predicated on a
transdisciplinary focus on movements in action rather than an isolated disciplinary view
42
.
An ecological dynamics approach to learning and performing
‘The approach [the ecological approach], while yet in its infancy, may
provide the next big shift in emphasis in attempting to understand motor
behavior (p.18).’
This epigram, taken from the excellent chapter that plots the history of motor
behaviour as an independent discipline was written by R.A Schmidt in 1982
25
. Setting up
this comment, Schmidt cited the early work of important influencers in ecological
dynamics, including Bernstein
33
, Turvey
43
and Gibson
26
. In this tutorial we recognise
Schmidt’s significant insights on ‘the ecological approach’
26
. Schmidt emphasized the
basic premise that “our motor system was created through evolution and interaction with
the physical characteristics of our environment and that we should, therefore, attempt to
understand the structure and function of the motor system by using more natural research
settings” p.18). Although we support this viewpoint, it is somewhat limited in recognising
the potential value of an ecological approach as being restricted to a greater emphasis on
use of ‘natural research settings’. In this tutorial we seek to support these early ideas by
setting learning and performing as symbiotic processes within an ecological dynamics
framework. Over decades, an ecological dynamics rationale has advocated the deep
integration of learning and performing in a model of skill acquisition based on the key
principle of the mutuality of the learner and the environments in which they are required to
perform
44,45
. An ecological model of learning and performance also takes into account
system complexity and variability
47
, amidst the non-linear, dynamical interactions of
individuals with environmental and task constraints. Specifically, we highlight the
importance of adopting ideas of representative learning design
47
to maximize the potential
for learning in practice tasks.
ECOLOGICAL DYNAMICS AND SKILL ADAPTATION
An ecological dynamics approach to skill acquisition adopts a systems orientation,
focusing on coordination of actions with a dynamic performance environment. This
systemic approach moves practitioners away from the asymmetric focus on internal control
processes within an individual, and more specifically brain activity, to one that focuses on
skill adaptation or skilled attunement to a performance environment
4
. Learning is,
therefore, reframed as a process of continually improving the fit (functional relationship)
between an individual and the environment by using surrounding perceptual information to
continuously regulate actions. The aim of learning designs are, therefore, to render
perception-and action more and more tightly coupled
13,26
. However, given the dynamic
nature of the relations between an individual and a performance environment, the
functionality of this fit is ‘a work in progress’ involving continuous adaptations and learning.
It is a nonlinear, dynamic relationship which can regress, stabilize, or progress, depending
on an individual’s experiences (practice, age, injury etc) over the life course. Davids and
Araújo
48
highlighted how functional behaviours emerge in performance environments and
BJMB
Brazilian Journal of Motor Behavior
Renshaw,
Davids,
O’Sullivan
2022
VOL.16
N.2
Special issue:
The role of practice in motor learning
169 of 178
that an individual’s performance solutions may vary over different timescales. These
variations are exemplified within individuals (e.g., changes in capacities and skills,
changes through growth and maturation) over macro timescales of years as the sport
evolves (e.g., performance surfaces, changes in rules and regulations, new technologies
and equipment, tactical trends).
When learning is conceptualised as an ongoing process throughout the lifespan, it
is seen as part of experience and a function of development
49
. It is important to note that
learning does not only occur in formalized, structured teaching and coaching sessions.
Learning opportunities also emerge in sport performance, providing contextualized
experiences of engaging with constraints of competition environments. Over short, medium
and long time frames, the skills and abilities that each individual develops are shaped by
all the environments or landscapes of affordances, providing opportunities for actions, to
which athletes are exposed
26,50
. For example, a tennis player who is brought up in a talent
development program on clay courts is more likely to become a strong baseline player,
whereas one brought up on grass courts, would more likely develop a strong serve and
volley game as the affordances of these surfaces invite the player to develop these skills
51
.
A cricket spin bowler who is brought up in Australia on hard pitches where the ball has lots
of bounce, will typically bowl faster and with less variability in delivery speed than those
spin bowlers brought up on pitches in the Indian sub-continent
52
. These experiences result
in bowlers performing better in the environments to which they are most adapted or
attuned to, based on their development
52
.
Of course, development continues throughout the life span. The previous
examples highlight the key idea that in an ecological dynamics approach, learning can be
re-framed as “an ongoing dynamic <