Movement learning

As motor learning (also English motor learning ) refers to the relatively permanent change of a motion sequence (coordination pattern of muscles) of a living being, if this is done by the intention that you could not reach by then to achieve through this movement a specific goal ( e.g. jumping over an obstacle, catching a ball, learning to speak a foreign language or learning to walk again after a stroke). You don't have to be aware of this process. The improvement of a movement sequence ( economization , faster, more fluid execution) is also movement learning.

Movement learning takes place constantly, as it also serves to adapt to new environmental situations. In this respect, movement learning is an essential part of the evolutionary development of living things. Since movement learning is controlled by the sensory organs (searching for and assessing the target) and regulated (constant online checking of whether one is on the right path to the target), one also speaks of sensorimotor learning . Professions that require a good knowledge of movement learning are those of sports teacher, coach, occupational therapist and physiotherapist.

Object of movement learning

In general, we associate the concept of movement learning with the idea of ​​learning movement sequences from the field of sport . We are also familiar with learning or improving movement sequences in physiotherapy . This includes B. learning (or relearning) to walk after an injury to the musculoskeletal system (bone fracture, joint or tendon, ligament injury, amputation) or the nervous system ( stroke , paraplegia ).

Movement learning in occupational therapy is a bit different because it is less important to practice the physiological process than to perform an action through movement that always has a purpose, a specific meaning.

Walking is an example of everyday movement such as speaking, writing, carrying loads, etc., which are first learned, but then have to be constantly adapted to our current environmental conditions and changes in our body (e.g. growth, aging). Another area of ​​physical activity learning relates to professional activities. These include B. Movements when doing handicrafts , when playing a musical instrument, when painting , picture cutting or acrobatics . Singing (mouth, throat and body movements) or learning (motor skills of pronunciation) a foreign language make special demands on learning to move.

In today's world of mechanization, robots play an increasing role. Their movements must be programmed. The knowledge about this is derived from the knowledge of natural movements. The current development is that the robots should be able to initiate and carry out learning processes and optimization of motion sequences themselves. The research group around Daniel M. Wolpert in Cambridge (UK) is working on this.

Function and biological basis

Movement learning takes place through imitation , trying out (similar to trial and error ) or instruction . Most of the time, all of these approaches are involved.

The ability of living organisms to learn new movements and to adapt to these new situations is based on the plasticity of the nervous system and the biochemical properties of nerve cells. The muscles are activated by the nerve cells, which means that the beginning and end as well as the strength (intensity) of their tension are determined.

Understanding the interplay between nerve cells and muscles requires knowledge of the structure and function of muscles and nerves.

Musculature

construction

There are three different types of muscles (see muscles ) in the human organism: the smooth muscles , the heart muscle and the skeletal muscles . They differ slightly in their structure. The skeletal muscle is the functional unit in the human organism that is responsible for the movements of the body. Two of its characteristics are responsible for this:

• It is contractile ; H. it can shorten (contraction, active) and slacken (passive).
• It is connected to more than one bone. Because of this, it is able to move the bones to which it is attached.

The skeletal muscle is also referred to as a striated muscle because its fine structure shows striations under the microscope. The structure that leads to this horizontal striation is responsible for allowing the muscle to contract.

A transversely striated muscle consists of several (the number depends on its size) parallel muscle fiber bundles . It is surrounded by a sheath of connective tissue, the muscle fascia . The muscle fiber bundles in turn consist of muscle fibers , these are the muscle cells. These in turn consist of a larger number of parallel structures, the myofibrils , which are responsible for the contraction of the muscle. These are also closely surrounded by connective tissue and form a functional unit with it.

function

At least 2 muscles are necessary for every movement of the body. Someone, the agonist (Greek: αγω = to do, to set in motion), who contracts and thereby actively induces movement in a joint, another muscle usually located on the "opposite side" of the joint, must allow this movement, by letting it stretch. This muscle is called the antagonist (Greek αντι, opposite '). When learning a new sequence of movements, the work of these muscles must be coordinated so that a successful and smooth movement is created. When a movement is made for the first time, the agonist and antagonist are usually activated together and contract. It's not very economical. In the course of the learning process, i.e. the exercise or the repetitions, the antagonist must “learn” to “retract” the activation and the agonist only to apply enough contraction for the movement to be successful and economical. This has to be done for all muscles involved in the movement and takes place through internal feedback (regulation) processes.

The activation or non-activation of each muscle is carried out (as the last final common pathway of the motor) via the motor nerve cells, which motor neurons in the spinal cord. From there, nerve lines pull to the muscle fibers in the muscles. There they end at special structures, the motor end plates . If an electrical signal ( action potential ) is sent from the motor neuron via this nerve to the motor end plates on the muscle, this signal is transferred to the muscle while being converted into a chemical process. There, specific chemical, electrical and mechanical processes cause the muscle fibers to contract. Several, but different numbers, muscle fibers are innervated by a motor neuron at the same time. This group together with the relevant motor neuron is called the motor unit .

A muscle contains many - often over 1000 - muscle fibers that are grouped together to form various motor units. When executing a sequence of movements, all motor units of a muscle are never innervated - except in absolute emergencies. The selection of the activated motor units follows a certain rule (the size principle : This means that small motor units are innervated first, then larger and larger ones), but they do not always have to be the same for every process.

Change through learning

When learning new or modifying motion sequences, many more motor units are activated for the sequence at the beginning of the learning process than are absolutely necessary. This is to ensure the security of the execution - to safeguard the execution. However, it often also means that the sequence of movements looks a bit rough and angular . In the course of improvement (through repetition ≈ exercise), the number of activated motor units is reduced more and more (this means that fewer and fewer nerve cells from the set - the pool - of motor neurons that are responsible for the muscle in question send out an action potential ) until an optimum is reached with which the goal of movement can still be safely achieved. The motion sequence looks easy and smooth at this stage. The duration of the muscle contraction is also adapted in the course of the learning (exercise) process for the movement in question. These processes are processes of economization.

The selection of the activated motor units by neurons at a higher level, the motor neurons to activate. This selection can be adapted to a new task (objective) through learning processes. But even the motor neurons themselves can be adjusted to new tasks through higher influences in such a way that they release more or fewer transmitters (here acetylcholine ). This leads to a change in the strength of the contraction. However, this does not mean that the contraction strength is not adapted to the current requirements (regulation of each individual sequence) for each individual movement.

Since coping with larger loads is also an adaptation (i.e. learning), the multiplication of mitochondria and the thickening of muscle cells, which occurs through endurance training and strength training , are also part of the changes in the muscle through learning.

Nervous system

construction

The nervous system consists of the central nervous system : cerebrum , cerebellum , brain stem and spinal cord and the peripheral nervous system (nerve lines to the organs of success). By learning to move, changes occur in all parts of this system.

The central building block of the nervous system is the nerve cell . The shape of the nerve cells can be very different depending on their task, but their structure is always the same. It consists of the nucleus, which is located in the body ( soma ) of the nerve cell. This is filled with the cell plasma and is surrounded by a membrane. There are numerous organelles in the plasma, some of which are necessary for the transport of substances, for example proteins , between the nucleus and the membrane, for the renewal of the nerve cells or for the support of information processing.

The individual nerve cells exchange information via their processes ( dendrites and axons ). Each nerve cell has a large number of dendrites, but only one axon at a time. Both types of processes are very strongly branched at their ends. The axons end on the dendrites, on the body of other nerve cells or on axons, mostly other nerve cells. At the end points they have specific thickenings, the synapses .

Motor neurons and interneurons are located in the spinal cord. The target areas of the axons of the motor neurons are not other nerve cells, but the fibers of the muscles . The axons of the interneurons terminate at motor neurons or other interneurons and can in this way serve to form networks.

function

The task of the nerve cells is to take in information, process it and forward the results. They receive their information via the synapses of the dendrites. They pass on the result of the information processing via the axon. The carrier of the information is the action potential . Since an action potential always has the same shape, the information lies in the frequency of the action potential. This can, for example, be uniform or occur in specific groupings.

Information processing takes place at the plasma membrane. After specific receptors at a synapse ensure that ion channels are opened by the transmitter of the sending nerve cell on the recipient cell (postsynaptic membrane), the influx of these ions changes the transmembrane potential. This applies to ionotropic receptors, which ensure rapid information processing, as is necessary for movement.

Neural control

Before a sequence of movements can be executed, many parts of the brain were already involved in preparing and planning a correct execution. There are so-called readiness potentials that precede every action and create a general disposition and activation of the organism. If a specific goal is to be achieved with the movement, a plan must be drawn up and the individual parts must be put together from known movement sequences. This happens in different parts of the cerebrum . The partial steps carried out are constantly compared with the goal and adapted to the conditions of the current situation (e.g. physical and emotional state) of the organism. In this respect, hardly any process is completely identical to another, although it often looks that way.

The "output area" of a movement begins in the motor cortex . From there, the motor neurons (the one hand motor neurons ) in the spinal cord innervating (directly) on the other hand, pulses to control centers such. B. sent to the cerebellum , which monitor the correct execution - in which one cannot consciously intervene. For the learning of new or the modification of already skilled movements, it is therefore important that these control structures are also adapted to the new requirements. This happens via feedback and the following modifications in nerve cells .

Nerve cells in the central nervous system (see: Central nervous system ) are responsible for the innervation of the motor neurons in the spinal cord. So if more or fewer motor neurons are required in the spinal cord for execution, the “decision” to do so is made at a higher level. This means that in the nerve cells of the higher levels there are changes (modifications) that are necessary for learning. If functions of the organism are to be changed in the long term, as is the case with learning , these changes must take place right up to the modification of DNA and genes . Such modifications take place in the cell nuclei . For movement learning, these changes must take place in the nerve cells that are responsible for the process and control of movements. If this change is not to be accidental, but targeted and repeated, the corresponding information must reach the cell nucleus. There are various ways for signal processing within the nerve cell - from the plasma membrane to the executing structures. All are initiated by the fact that specific ions flow into the cell from the extracellular side through specific channels (ion channels). To do this, these channels must open up. This in turn depends on their receptors, which react to specific substances on the outside of the cell membrane , either a transmitter released by another neuron or a hormone. A distinction is made between different types of receptors - the ionotropic and the metabotropic. The ionotropes usually respond to transmitters. The channels then open very quickly - z. B. Sodium ions into the cell, which change the membrane potential and thereby lead to an action potential in sufficient quantities, through which the signal is passed on. These channels ensure that signals are passed on quickly and directly, as required for the quick execution of actions such as B. Movements are necessary. For processes that do not have to take place as acutely as quickly as learning, the receptors are more diverse and the mechanisms for opening the channels more complex and they usually take place via so-called G proteins. You react z. B. on hormones (metabotropic). Information processing then takes place within the nerve cell. So-called second messengers play an important role here. In this way, necessary information can reach all components of the nerve cell, including the nerve cell nucleus. If they repeat themselves frequently in a similar way, they can make changes (modifications) e.g. B. to cause DNA and gene modification.

Change through learning

Changes in the function of nerve cells through learning occur primarily in two places: at the synapses and in the cell nucleus. If these changes are not to be accidental but rather permanent, as is the case with learning, the corresponding information must get into the cell nucleus and the DNA must be modified accordingly.

Changes can also be seen in the motor cortex (MI = primary motor cortex - this is where the nerve cells are located , from which the muscles can be directly innervated). As soon as movements are carried out repeatedly, the areas of the neurons that can trigger these movements enlarge (neural plasticity). This happens after just a few repetitions. A distinction can be made between 2 levels. An early learning stage in which there is a very rapid improvement in the execution of movements in a single exercise unit. This is followed by a later, slower learning stage, in which further, but slower learning progress can be observed over several exercise units that can take a few weeks.

Research history

The interest in how a new movement can be learned is very old, because especially in the times when no or only a few machines did the physical work for people, this work had to be done by people and as well as possible . The instructions for their execution were mostly about instructions and descriptions from the visible change in the movements of the learner. In the arts, techniques for mediation through experience and tradition had always been optimized. The general theories about learning were also based on the experiences of teachers and their records. This resulted in the methodological information. In scientific research, the theories of learning - including movement learning - run parallel and approximately at the same time as research into the nervous system and its structure and functions.

Reflexes and reflex chains

Systematic research into the learning processes of movements has only been possible since the beginning of the 20th century. However, they have not been integrated into general learning theories for a long time because they were not considered sufficient (see behaviorism ) to describe and explain primarily cognitive learning. Research into the nervous system since its inception has provided sufficient indications for the formation of movements, since the physiological causes are more easily accessible to the investigations. For a long time they were only obtained by studying movements - of animals.

The outstanding discoverers and researchers of this nervous system were Charles Scott Sherrington (1857-1952) and Ivan Petrovich Pavlov (1849-1936). Sherrington, for example, discovered that the nervous system consists of individual nerve cells that are physiologically separate from one another, but can exchange signals with one another at certain points and this exchange of signals is the cause of the movements. Both dealt with the reflexes. While Sherrington mainly examined the anatomy of the individual nerve cells and their function, Pavlov developed important models based on the reflex stimulus-response theory (see stimulus-response model ). Pavlov developed, among other things, the theory of the conditioned reflex (see Pavlov's dog experiment), which also became the basis of behaviorism.

Both came to the conclusion that the movement of a living being arises from a chain of reflexes. The contraction of a muscle serves as a reaction to a stimulus. This contraction, in turn, serves as a stimulus for the contraction of the next muscle and so on. Such a process was then called a reflex chain (see reflex chain theory ). The idea of ​​these reflex chains was that they would run in a very stereotypical manner, similar to a slot machine. This idea was supported by the fact that the observed movements of animals and people, which were mostly perfectly controlled executions - such as everyday movements - always looked uniform.

Discovery of the control loops in living organisms

When examining the behavior and movement of animals, e.g. B. in Seewiesen (see Konrad Lorenz ) were carried out around 1930, it was found that the movements of the animals could compensate for disturbances in their movement, for example due to obstacles - the movement sequences are not completely rigid.

In a special way, Erich von Holst (1908–1962) dealt with the question of how animals can compensate for such disturbances by modifying their movements. In the course of these investigations he developed the principle of reactivity , which explains these processes. However, it took a few years before the knowledge reached motor skills research that the reactivity principle is an important building block for understanding the movement of living organisms and learning to move.

In principle, the reactivity principle is a nested control loop . It fits in with the discoveries of the time when biologists and engineers found out that the subject of their respective research has much in common, only takes place on different matter and is usually described differently. It has been accepted since that time that living systems are also described using technical means.

This collaboration also develops the idea and description of human actions as a process of information processing. Since the entire action and learning process is divided into different, manageable sub-processes, it became clear at which points in the learning or teaching process, interventions in the learning process are useful and how they must be designed in order to be effective. These findings were also incorporated into motor skills research.

When the reactivity principle was recognized as an important building block for understanding movement and movement learning in the 1980s, a new, increased interest in researching movement and its control - in performing movement and learning new movements - began.

Movement learning and neuroscience

The new findings from the neurosciences, which had also become possible through technical advances in the study of brain structures, brought great progress.

One direction of research is concerned with determining in which areas of the brain the individual actions take place during human activities - following the flow of information in the brain . There were models put forward the track this flow and deepen this way understanding in a different direction studied to the activities of nerve cells and tried in these cells look inside to find out how they provide their services and influence the modifications of movements can have.

With the development of new technical processes (EEG, PET), with the help of which processes in the brain can be represented even more precisely, it was possible to observe the activation of brain areas while executing and learning movement sequences. This allowed the ideas to be checked. For example, it was found that many more cortical areas are involved in the creation of movement than just the motor cortices and some subcortical regions such as the basal ganglia and cerebellum.

Functional magnetic resonance imaging (fMRI) enables the flow of activity in the brain to be observed even more precisely, so that today it is possible to localize the stations of control loops when executing movements and to follow how they are built up when learning movements. It is hoped that through these investigations one can develop measures with which one can influence learning processes. The problem with this, however, is that large-scale movements cannot be carried out in the tomograph. It was also discovered that changes in learning processes can be observed, especially in the synapses, which are now being intensively researched.

Movement learning and robotics

For the use of robots in industry (construction of industrial goods), in medicine, agriculture and in the military sector, but also in the newer applications to help in many areas of life, it has proven to be important that robots not only programmed them in a prescribed, necessary way Can fulfill functions. Rather, it is also necessary that they can further develop their skills to adapt (adapt) - i.e. learn - to new situations.

For the control of the movements of robots, engineers had already dealt intensively with the movements of living beings and also advanced the findings for the neuroscientists . It was a logical step from studying control processes to examining learning processes, because control of movements must be learned, but it also leads beyond this.

However, since it is very difficult to investigate control and thus the learning of complex movements, simple movements were sought in which only a few (1 or 2) joints and only a limited number of muscles are involved. This is the case, for example, of the flexion of a forearm when viewed in isolation. This movement is diverse and has been examined under very different conditions of the disturbance of the process (control also means to compensate for a disturbance).

Eye movements to certain predetermined goals can also be viewed as such partial movements - this has been investigated in monkeys, for example. These can then be viewed as movement modules, which are then also referred to as "primitives". (Primitives in software engineering are the smallest elements of a larger (complex) computer program ).

These “primitives” can be used to examine very precisely how certain movements, which have been given new tasks, can be learned. The learning laws found correspond to those that were also described by the behaviorists - but they were called differently. For example: A new behavior is learned through “rewards” (= adaptation, behaviorism : acquisition). If it is no longer rewarded, it comes to a “washout” (behaviorism: extinction, deletion), if it is then “relearned” after a while, learning takes place more quickly through “savings” (behaviorism: “spontaneous recovery”). With these researchers today, however, in contrast to the behaviorists, the processes in the learning processes inside the organism are also examined. It was found that these terms of learning processes mainly in the cerebellum ( cerebellum play).

However, the researchers of these new studies do not make any statements about how later, after learning the individual "primitives" (partial movements), these can be combined into a "fluid, entire movement sequence". However, this is always the goal of human learning to move ( the movements are all known as individual exercises. The only difficulty lies in their coupling .)

Movement learning in the learning theories

There are various models that are intended to explain movement learning. These are mostly closely related to a general learning theory. They are changed when new insights or views on (movement) learning make this appear necessary.

In contrast to general learning theory, the theory of movement learning mostly serves to experimentally test and confirm or refute a procedure that is often based on tradition and / or experience and is controversial.

The theories for movement learning were and are essentially determined by the theories prevailing in the USA - there is also a learning theory for movement learning from the Soviet scientific field, which is, however, only poorly received in the western states. While research in the GDR made strong reference to this strand of research, the dissolution of numerous institutes in the course of German reunification prevented these concepts from being adopted in German research.

The Gestalt or wholeness theory, (≈ cognitivism in the USA), the first scientific learning theory, does not deal with movement learning because movement learning is not cognitive. In the middle of the 20th century, however, an offshoot of cognitivism, in the form of Austrian school gymnastics, gained at least limited importance in Germany (it was incorporated into the didactics of physical education by the Cologne Sports University and was included in the first post-war guidelines for physical education for the state of North Rhine-Westphalia). Karl Gaulhofer (1885–1941) and Margarete Streicher (1891–1985) contrasted Pehr Henrik Ling's physiological movement structure with natural gymnastics, which always assumes the entirety of a sequence of movements, so a learner starts with the entire sequence - often simplified - performs so that he can achieve the goal of movement - i.e. understand the whole process. Then individual parts of the movement can be practiced separately. This concept was mainly intended for elementary school gymnastics and is justified there.

behaviorism

→ Main article : Behaviorism

The most influential learning theory to date is behaviorism (behavior, also: behavior = behavior). It originated at the beginning of the 20th century in the USA and goes back to ideas of John Broadus Watson , although Edward Lee Thorndike (1874-1949) or Burrhus Frederic Skinner (1904-1990) are usually considered to be their founders. Watson assumed that learning can only be analyzed in a scientifically responsible manner by precisely describing and evaluating what is objectively observable, namely the behavior of the learner. Introspection - the means of cognitive learning theory - was rejected by him as unscientific.

The behaviorists were the first to systematically study and describe the processes of learning. They are often accused of having carried out their studies mainly on animals (for example pigeons and cats) and therefore not, as has happened, simply allowed to transfer them to humans. It was argued that their theory ignores consciousness and emotional states, neglects that all behavior is acquired during and on the basis of the individual life story, cannot explain achievements in the arts (for example in music, literature and also in the exact sciences) and much more . However, the behaviors described are elementary processes that can be triggered in every living being. With animals, it may be the only ones, humans can learn in more diverse ways.

In the simple learning mechanisms that have described the behaviorists, it is essentially the effect of reward ( reinforcement , reward ) and punishment ( punishment ), acquisition ( acquisition ), deletion ( extinction ), the spontaneous recovery ( spontanious recovery ) of behaviors, to name the most important ones.

Findings for movement learning

In behaviorism, movement learning is a more or less complex chain of simple SR connections, in which an SR element that has already been learned serves as an initial stimulus for a subsequent SR connection - this can be based on either classic or operant conditioning .

Unlocking a door consists of: B. from the individual compounds

1. Take the key in your hand (S: see the key - R: take the key in your hand)
2. Grasp the key correctly (S: hold the key in your hand - R. turn the key in the right direction)
3. Insert the key into the keyhole
4. Turn the key
5. open door

All of these individual connections had to be skillful so that, when the entire task was offered, it could be performed as a whole in a single process (i.e., spontaneously, without practice).

This does not always have to be successful for more complex processes. For example: Turn right by car when the traffic light turns green.

Chain:

1. See the green light → shift into gear and / or accelerate
2. Look to the right (because of cyclists)
3. Turn the steering wheel → check the turning process
4. Turn the steering wheel back
5. Check your journey : (stay in lane, reset direction indicator ?, speed)

The learner driver has mastered all of these elements. If they are challenged together for the first time, they often fail.

The amplifier plans by Burrhus Frederic Skinner are significant . You will e.g. B. in physiotherapy (PT) mostly used unconsciously: In PT practice, continuous reinforcement , i. H. Each execution is reinforced (correctness confirmed) the process is most securely consolidated. Later, at home, every execution is no longer reinforced; there is a reinforcement at irregular intervals ( intermittent , variable ratio ; the intervals can become longer and longer). This leads to the safest form of retention.

Skinner also developed the method of a higher form of operant conditioning: Shaping , in which the target behavior is systematically developed. It is used in animal training. The behavior that comes closest to the target behavior is reinforced from the respective behavior repertoire that the animal shows in the test situation. This changes the repertoire of the animal's response behavior and brings it closer to the target behavior. In this way - in sometimes long processes - very precise and alien behavior can be conveyed - not only in animals.

The relationship between the spatial constellation of stimuli and learning success was investigated, especially in the field of work sciences. It was found that if the elements to be operated are arranged in an unfavorable manner, learning success is delayed and errors can easily occur even when the process is mastered, e.g. B. Placement of switches for the plates of a stove or when an element to be manipulated with the right hand is placed on the left side of the body ( incompatibility of stimulus and response). These compatibility rules play a role in the design of objects, especially in industrial design for optimizing work processes.

Motor Control Theory

→ Main article : Control of movement

Control means constant monitoring of a process so that it can be successfully completed (by: control engineering = Control Technology ). The importance of control for movement learning was recognized through collaboration between movement scientists and engineers around 1940 (see also cybernetics ), in Germany in behavioral physiology, in the USA in the training of pilots for the Second World War.

It was already known to the behaviorists that without confirmation of the correctness (reinforcement, reward ) of a movement, no learning can take place. When it came to movement sequences, she believed that when a movement chain was set up, each chain link was reinforced and thus controlled when it was acquired and the entire sequence could then run automatically - without individual checks. The result is a kind of exercise program that is initially learned in sections, but then runs unconsciously.

If a movement is only viewed from the outside (as in behaviorism ), it prepares - despite the existing knowledge of physiological feedback mechanisms, e.g. B. the gamma loop of muscle control (see: motor skills ) - the understanding of constant monitoring of movement sequences difficulties. A control did not seem useful, since corrections of a movement were only considered possible after its completion with effectiveness for the next sequence, because the time for a correction of the sequence takes too long (it was assumed about 200 ms) than it would be during a sequence of movements can become effective. However, this did not apply to slow movements. There, for example, the executor could make corrections himself.

Therefore, there was a distinction between "open loop" running fast movements, which can only be assessed (controlled) after their execution and corrected for the next sequence (for example the throwing of a ball) and "closed loop" running slow movements, in which during the sequence Corrections are possible (e.g. drawing a circle).

For the slow sequence of a “closed loop” movement, the engineering-style representation using flow and structure diagrams, which are supposed to represent the information processing during the sequence, was increasingly used. Although these structure diagrams are an invitation to think about the processes in the organism, because they contain elements inside the "organism box" that would have to be filled with physiological knowledge, the behaviorists found it difficult to revise their concept of the black box . Exercise programs offered a solution to this problem.

Exercise programs

At that time, however, programs (e.g. computer programs) were considered to be rigid (idea: machine: a coin (= the desire to move) is inserted at the top. Then the exact same sequence of movements always comes out). This corresponded to the idea of ​​the "unconscious" running sequence of movements and could also be observed in this way. But that was undesirable because movement sequences have to be flexible, i. H. be able to adapt to different conditions (in the organism, e.g. tiredness, or in the environment, e.g. rough ground).

The advent of branching programs in computer science seemed to solve this problem of rigidity in motion programs in a simple manner.

The analogy with the branched computer programs, however, led to the problem of how the organism can learn such a branched program, i. H. programmed: how does he know where the branch points have to be placed and according to which criteria one or the other path has to be taken. Control of such a program must also be learned.

Another problem arose that people had to learn and keep a great many different exercise programs in the course of their lives. These must be stored in a long-term memory. A certain solution to this problem could be achieved by the so-called schema theory . This means that when a new movement sequence is learned, a coarser scheme of the overall sequence is also formed, so that the learned movement does not have to be learned from scratch under similar conditions, but only has to be adapted and saved as such. Example: We learn to write with the right hand. We can then not only do it correctly on the blackboard, but also with our left hand or foot in a fairly legible manner, although we have to activate completely different muscles than those with which we learned to write. Richard A. Schmidt has dealt intensively with this schema theory.

The psychologists G. Miller, E. Galanter K. Pribram made an important contribution to understanding the control of human behavior with their TOTE (Test-Operate-Test-Exit) model.

An approach to explain and understand the physiological fundamentals of movement control is offered by Erich v. Holst with the principle of reactivity .

Information processing theory

The control theory was further developed by psychologists in the 1960s and 1970s into the theory of information processing in the human organism. This helped to solve some problems from control theory.

In the theory of information processing, as in control theory, human action is represented as a structure diagram, but the individual elements are filled with the corresponding physiological and neurophysiological content. These provide information about the way in which the information from the environment that affects the organism is processed before the result affects the environment. The psychologists mainly limit themselves to perception and memory .

Model for the theory of information processing

For the movement scientists, however, the execution of the action / movement with the preparation and control of the muscle activity is particularly important. The Canadian Ronald G. Marteniuk implemented this theoretical approach for movement learning in 1976 based on preliminary work by HTA Whiting. This was made known in Germany by Heidrun Schewe.

The model

The result is a structure with five main elements:

1. On the sense organs are signals ( information ) from the environment into the nervous system was added. This is a purely biological process.
2. In the next step - the beginning of cognitive processing - the information that is important for the specific movement is filtered out and grouped in such a way that the situation in which the organism is located is identified and its significance for the current situation or for a planned action is recognized. At the end of this step, which you as a range of perception ( perception , perception mechanism ) refers, is the classification of the overall situation.
3. If the situation can be clearly assigned to a certain already known class of situations, and there is only one alternative course of action, this can be called up and carried out. If a clear classification is not possible or if there are several alternatives (e.g. in a sports game or different routes for a car trip), an adequate or successful solution strategy must be selected or sought. This has for a decision ( decision ) for (mentally) played through and be checked for its consequences through. This takes time. If this is not available or if the person performing the task is impatient, wrong decisions can be made that lead to the failure of the movement. If a solution strategy is found, the consequences of which can be accepted, this intention to move is passed on to the "execution area".
4. In the execution area ( effector mechanism ) the sequence patterns of the movement are put together. If the intended movement has already been carried out several times - this is usually the case because we collect movement experience from childhood on, and suitable innervation patterns are available for all our muscles, so that completely new movements hardly occur in adulthood - the movement patterns required for this only have to be created deployed and activated. If the intended movement is new or relatively new, suitable patterns (partial patterns) must be collected and coordinated with one another.
5. The last section, the “muscle system”, is responsible for converting the cognitive results into the mechanical movement sequence. This takes place under constant control.

In each processing step, all previous processing steps can be used. They are constantly compared with the surrounding situation and the intended goal.

The conclusion of the overall process is the determination of whether the intended goal has been achieved or not. Accordingly, the process is saved as positive or negative ( knowledge of result ) = evaluation. This model also allows a detailed analysis of the overall process for both the teacher and the learner ( knowledge of performance ). The importance of these feedback mechanisms is discussed in detail by A. Gentile, also by H. Schewe.

The advantages of this theory for those performing and teaching are the ability to follow and analyze the development and execution of a movement sequence in its individual steps . This is important for planning, analyzing, assessing and correcting, especially when learning to move. By including additional psychological knowledge (e.g. attention to perception), differences in performance in execution can be explained and influenced.

The problem of control theory, how new programs and their control can be learned, is solved by resorting to the movement experiences of the individual.

The theory of information processing represents the knowledge that is valid today about the occurrence of the behavior of living systems, including humans. This basis has led to further questions. The first and simple question about this model is in which structures of the organism these processing steps take place. The nervous system as the information system is of course responsible for this, but the locations had to be sought for the individual sub-processes and, above all, the transition structures had to be examined. Much of that happened today. Because of the current investigation methods, many of the processes involved know where they take place.

Another question is about the time required to carry out the individual processing steps. This leads to 2 other problems. On the one hand, the question of how much time do you need to prepare a movement - how does that work with spontaneous movements? The second, particularly important, is: under these conditions, can a movement be corrected during its course? That is the question of the control of the movement (open loop - closed loop). In the graphic model, feedback loops are entered for each of the processing steps - this means that internal control of the individual steps is assumed. This problem is currently still a main subject of movement research.

As a kind of intermediate result, an old idea about the execution of motion sequences can be described. This means that the control of motion sequences has a hierarchical structure, i.e. it takes place on different levels. The reflexes on the short spinal cord arches are the lowest level and the conscious control of a process is the top level. This can also be represented well with the ideas of the reafferent principle and is generally accepted today.

This representation of the hierarchy of movement control is also linked to the idea of ​​the awareness of the execution of movement sequences. This question of whether and under what conditions the execution is accessible to the consciousness, and whether it can be learned, must be the subject of philosophical consideration and research. Results on this are now more of a speculative nature.

Neural Group Selection

While the theses of the theories described so far were derived from experimental investigations of behavior, the theory of Neural Group Selection (TNGS) (see Evolution ) draws its conclusions directly from the studies of the brain and its activity.

New questions and doubts - e.g. B. how the classes of actions / movements, according to which one decides on a certain execution, could have arisen, since a corresponding pre-structuring must exist for this; that variants of movements - to adapt to new environmental conditions - were only conceivable within a range of natural noise; and that on the neural level (see nerve cell ) it could not be explained why large amounts of neural connections seem to be broken during movements - made new considerations necessary.

Learning to move

According to the theory of the Neural Group Selection (TNGS) (see Evolution ), it follows that all processes (mental and motor) - every thought process, every action and every movement - are not based on programs learned in the brain in a representative form , but according to the evolutionary one Selection principle planned, developed, coordinated and executed. So every process is completely rebuilt and put together. Existing neural connections are used and changed at the same time. This also applies to movement learning.

The principle is to make a suitable selection from a very large (several billion elements) population (in the organism: populations of synapses on neurons (see nerve cells , neuron networks and groups of neuron networks, to groupings, repertoires and maps )). These respective groupings within a repertoire are similar but not the same ( degeneracy ), so that the same performance can be achieved even with different choices. Through constant recursive exchange of simultaneously active groups of neurons in the reciprocally connected regions that are distributed over the entire cortex (see cerebral cortex ) ( reentry ), the sensory and motor events are spatially and temporally coordinated.

The selection made is different for each execution and it is usually not an exact, but a sufficiently good solution, which is optimized by the current constant checking and changing of the synapses and can be almost optimal at the highest skill level.

With this method, completely new processes can also be carried out successfully. These need more time for their planning and preparation (selection and coordination through reentry ). They are also less precise. But repetition ( practice ) improves choice, and coordination can be more economical and faster.

If a new sequence of movements is to be learned, it is therefore sensible to ask the learner to vary the execution from the outset so that more “suitable” connections are available for the selection of the respective optimal solution.

Problems solved

This theory solves the problem of memory as a representative store of all learned movement sequences, since every movement is recomposed and optimized from the most suitable links.

The observation is also confirmed and explained that at the beginning of the learning of movements the consciousness is involved to a greater degree than at higher levels of ability. The sequence of movements is automated in the course of frequent reuse (exercise), as they say. There will then be less attention d. H. also requires energy for the process such. B. for everyday movements. However, that does not mean that it runs without control.

This theory is also suitable for explaining many of the contradicting research results from behavioral learning research (e.g. massed or distributed practice).

Introduction aid: muscles are activated at the last stage of execution. A muscle consists of a very large number of motor units (see motor skills ). These motor units are innervated by motor neurons in the spinal cord. However, only a limited number of motor units is required for each movement. You should consider whether the same motor units are always activated for a muscle contraction in which a predefined, defined force is to be applied (program presentation) or whether by selection (see evolution ) more suitable, but different motor units are not also activated required precise output (performance) comes.

Critical Rationalism

→ Main article : Critical rationalism

Karl Popper's Critical Rationalism chooses a philosophical, metaphysical and epistemological approach to the problem of learning to move, which, however, critically includes scientific knowledge: evolutionary epistemology . Critical rationalism emphasizes the essential connection between a movement of the organism and the organism's knowledge of movement sequences and thus the relevance of epistemology in this question. He rejects all behavioristic and biologistic positions on movement learning as reductionistic and scientistic . He assumes that movements are actions and thus attempt to solve problems, i.e. aim at the realization of a purpose, even if purposes as such are metaphysical.

Critical rationalism emphasizes a fundamental dichotomy in the learning process. The first part of learning to move about is therefore learning an adequate sequence of movements itself. For only a few sequences of movements are adequate with regard to a given purpose, and an organism will go through many attempts and failures until the movement leads to the desired purpose, or the purpose itself as is abandoned inaccessible. He ultimately chooses these attempts at movement with the help of chance, but not purely by chance, as he will already have background knowledge of many facts and thus can rule out many attempts at movement as being unsuitable from the outset (e.g. that to get to the fruit on the tree , jumping the legs is more promising than wiggling the ears). In lower living beings, these purposes, as well as the knowledge of movement processes, are more or less firmly built into the genetic material and can only be changed through mutation and natural selection. Higher living beings, on the other hand, have higher control organs (e.g. brain) through which they can consciously set these purposes and also change them in the course of life, up to and including humans, who can even choose their purposes in a more conscious and far-sighted manner . Accordingly, higher living beings can consciously recognize failures in a sequence of movements as a falsification of the assumption that it achieves the purpose. You can then consciously change the sequence of movements and try again.

If a higher organism has at some point learned an adequate sequence of movements in this way, the second stage of the learning process follows, namely the constant repetition of the sequence of movements. Through this repetition, it is practiced and thereby leaves the consciousness, is imprinted in the subconscious and can then be called up and executed by the organism unconsciously and “automatically” so to speak. A sequence of movements in the subconscious can, however, return to the level of consciousness at any time if, under other circumstances, it should no longer prove to be adequate for achieving the purpose. And it can be consciously brought back into consciousness in people even without a concrete failure and examined in the strictest critical manner, also with foresight with regard to circumstances that have not yet occurred. Popper sees in this the essence of human rationality and ultimately of science. In addition, humans can check movement sequences not only for adequacy for an immediate purpose, but at the same time also for several other, possibly higher purposes, such as morality. He can thus consciously refrain from movements that are adequate with regard to an immediate purpose (e.g. moving forward towards a well in order to drink, although another person is in the way), but which run counter to a higher moral purpose (that you shouldn't bump into other people).

The essential, first part of the learning process is thus dominated asymmetrically by the negative element of failure, i.e. H. of inadequacy for one or more purposes. Confirmation, on the other hand, does not play a role at all, not even in the second part of the learning process, at least insofar as confirmation should be more than the mere (temporary) absence of the failure. This is shown by the fact that even an inadequate movement process that is not or no longer confirmed as appropriate can be shifted into the subconscious through repetition ( quirks ). According to Popper, the main mistake of most approaches is to overlook the dichotomy of the learning process and to understand repetition as the real and essential element of movement learning, and thus amount to a mistaken inductivist conditioning theory .

literature

• Shelly A. Boudreau, Dario Farina, Deborah Falla: The role of motor learning and neuroplasticity in designing rehabilitation approaches for muscucoskeletal pain disorders . In: Manual Therapy , 2010, doi: 10.1016 / J.math.2010.05.008
• Zoubin Ghahramani, Daniel M. Wolpert: Modular Decomposition in visuomotor Learning. In: Nature. 386, March 1997, pp. 392-395.
• Richard A. Schmidt, Timothy D. Lee: Motor Control and Learning. 4th edition. Champaign IL 2005.
• Anne Shumway-Cook, Marjorie H. Woollacot: Motor Control. 3. Edition. Philadelphia 2007, p. 22.
• Richard A. Schmidt: A Schema Theory of Discrete Motor Skill Learning. In: Psychological Review. 82: 225-260 (1975).
• Ronald G. Marteniuk: Information Processing in Motor Skills . Holt Rinehart and Winston, New York 1976.
• Eric R. Kandel, James H. Schwartz, Thomas M. Jessel: Principles of Neural Science. 4th edition. McGraw-Hill Companies, New York 2000, ISBN 0-8385-7701-6 .
• Heidrun H. Schewe: The human movement . Thieme Verlag, Stuttgart 1988, ISBN 3-13-706901-7 .
• Heidrun H. Schewe: The human-performance-model - an explanatory approach for movement learning . In: Physiotherapy 31 (1979) pp. 541-554
• Heidrun H. Schewe: Ways to understand movement and movement learning . In: B. Lipp, W. Schlegel, K. Nielsen, M. Streubelt: Caught in your own body - solutions / neurorehabilitation. Neckar-Verlag, Villingen-Schwenningen 2000.
• Erich von Holst, Horst Mittelstaedt: The Reafferenzprinzip. In: Natural Sciences. 37 (1950) pp. 464-476.
• Peter H. Lindsay, Donald A. Norman: Human Information Processing, an Introduction to Psychology . Academic Press, New York / London 1977.
• Thorsten Kahnt, Marcus Grueschow, Oliver Speck, John Dylan Haynes: Perceptual Learning and Decision Making in Human Medial Frontal Cortex. In: Neuron. 70, 2011, pp. 549-559.
• Christian Rosenmund, Carsten Stevens: Definition of the Readily Releasable Pool of Vesicles at Hippocampal Synapses. In: Neuron. 16, 1996, pp. 1187-1207.
• Burrus Frederic Skinner: What is Behaviorism . Rowohlt Verlag GmbH. Reinbek near Hamburg. 1978.
• Thorsten Kahnt, Marcus Grueschow, Oliver Speck, John Dylan Haynes: Perceptual Learning and Decision Making in Human Medial Frontal Cortex . In: Neuron. 70, 2011, pp. 549-559.
• David W. Franklin, Theodore E. Milner: Adaptive Control of Stiffness to Stabilize Hand Position with Large Loads . in: Experimental Brain Research 152 (2003) pp. 211-220.
• Joseph Y. Nashed, Isaac L. Kurtzer and Stephen Scott: Context-dependent inhibition of unloaded muscles during long-latency epoch , in: Journal of Neurophysiology 113 (2015) pp. 192-202.
• David W. Franklin Daniel M. Wolpert: Specificity of Reflex Adaptation for Task-Relevant Variability in: The Journal of Neuroscience (Behavioral / Systems / Cognitive) 24 (2008) pp. 14165-14175
• Maurice A. Smith, Ali Ghazizadeh, Reza Shadmer: Interacting Adaptive Processes with different Timescales Underlie Short-Term Motor Learning . Plos Biology. Volume 4, June 2006, Issue 6, e179
• Dieter Kruber: Sport . In the series: Erhard Schmidt, editor; Standard work by the teacher Ferdinand Kamp Verlag Bochum 1973.
• Friedrich Fetz: General methodology of physical exercises . 4th edition Limpert Verlag Frankfurt 1971.
• Daniel M. Wolpert, Jörn Diedrichsen, J. Randall Flanagan: Principles of Sensorimotor Learning. In: Neuroscience. 12, 2011, pp. 739-751.
• Michael J. Zigmond, Floyd E. Bloom, Story C. Landis, James L. Roberts, Larry R. Squire: Fundamental Neuroscience . Academic Press. San Diego / London 1999.

Individual evidence

1. ^ Daniel M. Wolpert, Jörn Diedrichsen, J. Randall Flanagan: Principles of Sensorimotor Learning. In: Neuroscience. 12, 2011, pp. 747-748.
2. ↑ Michael J. Zigmond, Floyd E. Bloom, Story C. Landis, James L. Roberts, Larry R. Squire: Fundamental Neuroscience . Academic Press. San Diego / London 1999. Chapter 9. Neurotransmitter Receptors pp. 235-268, and Chapter 55 Learning and Memory. Basic mechanisms pp. 1411-1453
3. ↑ Michael J. Zigmond, Floyd E. Bloom, Story C. Landis, James L. Roberts, Larry R. Squire: Fundamental Neuroscience . Academic Press. San Diego / London 1999. Chapter 10 Intracellular Signaling pp. 269-315
4. ↑ Shelly A. Boudreau, Dario Farina, Deborah Falla: The role of motor learning and neuroplasticity in designing rehabilitation approaches for muscucoskeletal pain disorders . In: Manual Therapy , 2010, doi: 10.1016 / J.math.2010.05.008 , pp. 2/3
5. ↑ Erich von Holst, Horst Mittelstaedt: The Reafferenzprinzip. In: Natural Sciences. 37 (1950) pp. 464-476.
6. ^ Peter H. Lindsay, Donald A. Norman ' Human Information Processing, an Introduction to Psychology. Academic Press, New York / London 1977.
7. ↑ Ronald G. Marteniuk. Information Processing in Motor Skills. Holt Rinehart and Winston New York 1976.
8. ↑ Heidrun H. Schewe: The movement of people. Thieme Verlag, Stuttgart 1988, ISBN 3-13-706901-7 .
9. ↑ see for example: Zoubin Ghahramani, Daniel M. Wolpert: Modular Decomposition in visuomotor Learning. In: Nature. 386, March 1997, pp. 392-395; and: Thorsten Kahnt, Marcus Grueschow, Oliver Speck, John Dylan Haynes: Perceptual Learning and Decision Making in Human Medial Frontal Cortex. In: Neuron. 70, 2011, pp. 549-559.
10. ↑ see for example: Christian Rosenmund, Carsten Stevens: Definition of the Readily Releasable Pool of Vesicles at Hippocampal Synapses. In: Neuron. 16, 1996, pp. 1187-1207.
11. ↑ for example: Daniel M. Wolpert, Jörn Diedrichsen, J. Randall Flanagan; Principles of sensorimotor learning. In: Nature Reviews 2011. pp. 730-751.
12. ↑ Eric R. Kandel, James H. Schwartz, Thomas M. Jessel: Principles of Neural Science. 4th edition. McGraw-Hill Companies, New York 2000,
13. ↑ for example: or David W. Franklin, Theodore E. Milner; Adaptive Control of Stiffness to Stabilize Hand Position with Large Loads in: Experimental Brain Research 152 (2003) pp. 211-220; or: Daniel M. Wolpert, Jörn Diedrichsen and Randall Flanagan; Principles of Sensorimotor Learning in: Nature Reviews (Neuroscience) 12 (2011) pp. 739-751
14. ↑ for example: Joseph Y. Nashed, Isaac L. Kurtzer and Stephen Scott; Context-dependent inhibition of unloaded muscles durin long-latency epoch. In: Journal of Neurophysiology 113 (2015) pp. 192-202; or: David W. Franklin and Daniel M. Wolpert; Specificity of Reflex Adaptation for Task-Relevant Variability in: The Journal of Neuroscience (Behavioral / Systems / Cognitive) 24 (2008) pp. 14165-14175
15. ↑ Maurice A. Smith, Ali Ghazizadeh, Reza Shadmer; Interacting Adaptive Processes with different Timescales Underlie Short-Term Motor Learning. Plos Biology. Volume 4, June 2006, Issue 6, e179
16. ↑ Dieter Kruber. Sport in the series: Erhard Schmidt editor; Standard work by the teacher Ferdinand Kamp Verlag Bochum 1973. P. 66 See also: Frieder Fetz. General methodology of physical exercise . 4th edition Limpert Verlag Frankfurt 1971 p. 91 ff.
17. ↑ Natural gymnastics , edited by Hans Groll, 3 volumes. Vienna / Munich. 1930-1942
18. ^ John Broadus Watson: Behaviorism . German translation. Cologne 1968.
19. ↑ for example: Burrus Frederic Skinner; What is behaviorism . Rowohlt Verlag GmbH. Reinbek near Hamburg. 1978. pp. 9-10
20. ^ "After: Robert Mills Gagné, The conditions of human learning. Schroedel Verlag, 1969, p. 103.
21. ↑ C. Ferster, B. Skinner: Schedules of Reinforcement. New York 1957.
22. ↑ Erich v. Holst: On behavioral physiology in animals and humans. Collected papers Volume 1 and 2. Piper paperback, Munich 1970.
23. ↑ see also: Richard A. Schmidt, Timothy D. Lee: Motor Control and Learning. A behavioral emphasis. 4th edition. Human Kinetics Publishers, Champaign IL 2005, p. 164.
24. ^ Richard A. Schmidt: A Schema Theory of Discrete Motor Skill Learning. In: Psychological Review. 82: 225-260 (1975).
25. ↑ GA Miller, E. Galanter, KH Pribram: Plans and Structures of Behavior . London 1960. German: strategies of action. Plans and structures of behavior . Stuttgart 1974.
26. ↑ Erich. v. Holst, Horst Mittelstaedt: The Reafferenzprinzip. In: Natural Sciences. 37 (1950) pp. 464-476.
27. ^ Peter H. Lindsay, Donald A. Norman: Human Information Processing, an Introduction to Psychology. Academic Press, New York / London 1977.
28. ↑ Ronald G. Marteniuk: Information Processing in Motor Skills . Holt Rinehart and Winston, New York 1976.
29. ↑ HTA Whiting (Ed.): Readings in Human Performance . London 1975.
30. ↑ AM Gentile: A Working Model of Skill Acquisition with Application to Teaching. In: Quest. 17. 1972, pp. 42-45.
31. ↑ Heidrun Schewe: Ways to understand movement and movement learning. In: B. Lipp, W. Schlegel, K. Nielsen, M. Streubelt: Caught in your own body - solutions / neurorehabilitation. Neckar-Verlag, Villingen-Schwenningen 2000.
32. ↑ for example: F. Crevecoeur, I. Kurtzer, SH Scott: Fast corrective responses are evoked by perturbations approaching the natural variability of posture and movement tasks. In: Journal of Neurophysiology. 107 (2012), pp. 2821-2832. or J. Andrew Pruszinski, Isaac Kurtzer, Stephen H. Scott: The long latency reflex is composed of at least two functionally independent processes. In: Journal of Neurophysiology. 106 (2011) or Joseph Y. Nashed, Isaac L. Kurtzer and Stephen Scott: Context-dependent inhibition of unloaded muscles during long-latency epoch, in: Journal of Neurophysiology 113 (2015) pp. 192-202.
33. ↑ Gerald M. Edelmann: Neural Darwinism. The Theory of Neural Group Selection. Basic Books, New York 1987.
34. ↑ Gerald M. Edelmann, Giulio Tononi: Brain and mind, how consciousness arises from matter. CH Beck Verlag, Munich 2002, p. 68.
35. ↑ Karl R. Popper, John C. Eccles: The I and its brain. 1977.
36. ^ Karl R. Popper: Objective knowledge. 1973.
37. ↑ Karl R. Popper: Knowledge and the body-soul problem. 1994.
38. ^ Karl R. Popper, Konrad Lorenz, Franz Kreuzer: The future is open. The Altenberg conversation. 1985