Copyright © 2007-2017 Russ Dewey
The information processing approach to motor behavior began after World War II. Norbert Weiner's influential book Cybernetics proposed a general theory of control systems and guided movement with clear relevance to motor control.
In principle, the flight of a guided missile and the reaching movement of an arm toward a target are similar. Both involve the same sort of information processing.
A guided missile and a reaching arm both use negative feedback. To an engineer, negative feedback does not mean criticism. A negative feedback process in a system is one that reduces deviation from a goal state.
What is negative feedback and how does it relate to motor movement?
A guided missile control system feeds back information about the direction of its target to the guidance system. The information is used to adjust the fins on the missile, turning it so it points toward the target.
As the missile travels, it reduces the deviation (distance) between itself and the goal. The word negative in negative feedback refers to this reduction in the distance between (1) the present state of the system, and (2) the goal state.
A guidance system must be a negative feedback or deviation-reducing system. That means it pursues goals. Guidance systems are interesting to psychologists and scientists working on robotics because almost all intelligent motor behavior is directed at goals.
In motor movements, often there are two distinct components:
1. A ballistic component that launches the system in the right general direction (like launching a missile)
2. A zeroing in component with negative feedback processes (like guiding a missile to its target)
What are two components of many guided movements? How do they show up in eye movements?
Jerky eye movements called saccades are the most common variety of eye movement. They have both ballistic and zeroing-in components.
First the eye movement is launched toward a target (the ballistic component) then a network of muscles perform tiny adjustments (the zeroing in component). The final adjustment allows the most important part of the visual image to land on the fovea centralis, the most sensitive receptor surface in the eye.
As a result of the two components, a diagram of a saccade often looks like a cane. There is a long straight jump followed by a little hook.
All movements require some form of aiming to achieve accuracy. The faster a person executes a movement, the less time is available for aiming.
In 1954 Paul Fitts, a leading researcher in the field, outlined equations describing the speed-accuracy trade-off. The basic idea is that the faster you perform a motor movement, the more mistakes you make. The trade-off was summarized in an equation called Fitts' Law.
What relationship did Fitts's Law express?
Paul Fitts died unexpectedly in 1965. The death of one researcher should not be enough to damage a whole field of research, but the death of Paul Fitts seemed to symbolize what happened to motor research over the next few years.
Federal funding for research into motor behavior was generous in the post-war era of the 1950s. Then it disappeared in the 1960s.
Conditions became so unfavorable for researchers that an academic funeral for the discipline of motor research was held at Tulane University. "Renowned motor behavior psychologists gathered to read the last rites and bid each other farewell, as each moved on to other related topics in psychology." (Schmidt, 1982)
What caused a new interest in motor research?
But the field of motor research was not dead. The resurgence of cognitive psychology in the late 1960s and 1970s stimulated new research on motor behavior.
Now motor behavior was recognized as a form of cognition or information processing. Motor tasks (such as teaching a robot to navigate a crowded room) involved all sorts of cognitive processes.
A robot must combine perceptual processes such as pattern recognition, problem solving, and motor behavior. Motor behavior became recognized as one of a suite of skills required of any intelligent system. It fit right in with perceptual research and language research as a form of information processing suitable for study by cognitive scientists.
Scientists who study motor skills speak of routines. A routine is a skill, a procedure, a recipe, a series of instructions for performing a task. A sub-routine is a recipe within a recipe. For example:
Skipping past the first 4513 steps in the program...
Step #4514: Add one cup of flour
Sub-routine 4514a: Get measuring cup
Sub-routine 4514a-1: Locate cup
Sub-routine 4514a-2: Pick up cup
Sub-routine 4514b: Add flour
Sub-sub-routine 4514b-1: Locate flour container
Sub-sub-routine 4514b-2: Pick up flour container [followed by instructions 4514b-2-a,b,c to navigate to it, grasp it, lift it]
Sub-sub-routine 4514b-3: Take lid off flour container [followed by instructions 4514b-3-a,b,c to twist or pry the lid, remove it, set it somewhere convenient...]
Sub-sub-routine 4514c-1: Dip scoop into flour
Wait! Where did the scoop come from? We need to add in another sub-routine before 4514c to locate and pick up a scoop!
[And so forth.]
If you were designing a robot to bake a cake, you would have to think in terms of complete job descriptions. You could not leave anything out.
If you failed to specify even one tiny but necessary sub-routine, your robot would look helpless and foolish. It might try to remove flour from a container without a scoop or something equally mindless.
The point is: even a relatively simple task (for humans!) consists of multiple steps. Each step has sub-routines embedded within it, to make that step possible.
Each sub-routine is a recipe within a recipe. The numbers of sub-routines embedded inside other sub-routines seems endless. Yet humans self-program this type of thing within our nervous systems, routinely, to guide activity.
Perceptual skills also must be woven in with motor skills. The robot cake-maker must "locate flour container" at one step. That might involve sophisticated pattern recognition skills, or considerable trial and error, as various containers are inspected or opened to see if they contain flour.
If you need a scoop, how will your robot recognize it? Even if it has access to a drawer full of kitchen tools, it must be able to pick out a scoop from all the other tools. [And so forth, on and on.) If any one of these sub-routines was improperly specified, the robot would fail in its task.
What is a routine? A sub-routine?
To completely specify the operations in even a simple motor task is very challenging! Looking back to the early days of AI in the 1960s, it is hard to believe serious scholars predicted household robots would be helping with housework by the 1980s.
As the complexity of even simple motor activity became obvious, ambitions of researchers have been scaled down. Rodney Brooks of MIT and Randall Beer of Case Western University became famous for building robot insects. They could scuttle about, pursue or avoid light, respond to sound, and other simple things. The first appeared in 1987.
Many robot insects are available by now, so beginning students in robotics can assemble them for practice. We have robotic cockroaches today, although we may have to wait many years for robot housecleaners beyond the floor vacuum cleaners we have today.
How have ambitions for robots been lowered?
Actually, to be realistic, we do not have anything remotely as complex as a cockroach available in robotics today. A cockroach is quite a piece of work.
It navigates through complex environment, climbs walls with aplomb, finds food, and reproduces itself. It is far more complex system than our most complex industrial robots.
There has been good progress on bipedalism: creating robots that walk on two legs. Even that sort of task requires complicated sub-routines; for example, each time a robotic foot is set down, it must sense the stability or tilt of the ground, then make appropriate adjustments.
If you recall from Chapter 4, a gestalt was defined as a figure or form. The examples in chapter 4 involved visual forms, but motor gestalts also exist.
A gestalt or schema is a pattern based on previous experience. It has an existence "other than its parts." In other words, it exists independently of any particular expression of it.
Neisser (1967) suggests a simple experiment for demonstrating motor schemata:
Ask someone to trace a letter of the alphabet on your back with his finger. You will have little difficulty in identifying the letter he marks out, although it is quite unlikely that such a pattern ever appeared on your back before.
This indifference to locus, and even to modality, is a remarkable phenomenon. In many ways, it seems closely akin to the transferability of motor skills. Having learned to make letters with a pencil in your hand, you can also make them, perhaps a little awkwardly, with one held in your teeth or your toes or even the crook of your elbow. (p.53)
What experiment, suggested by Neisser, shows transferability of schemata?
Motor activity can be guided by something abstract–a schema. It does not depend on any particular set of muscles.
It can be transferred in the production phase (making letters with your toes) or the comprehension phase (deciphering a letter traced on your back). The schema is an abstract pattern that can be expressed or detected in many ways.
Many different muscle groups might be used to execute a particular movement, but this is not what defines the movement, unless you are interested in muscle physiology. The meaning of the movement is the overall form, the pattern, the gestalt.
The "same movement" might be achieved in several ways, by different sets of muscles, once you learn how to make it. This is what Neisser refers to as the transferability of motor skills.
How can transferability of schemata aid the handicapped?
Transferability of schemata is important in overcoming handicaps. People who are born without arms use their feet in strikingly hand-like ways, driving and holding silverware with their feet, naturally and gracefully.
Video records of such individuals create a definite impression that one is watching a hand in action rather than a foot. Brain areas responsible for manual dexterity, normally used by the hands, are successfully re-assigned to the feet in such individuals, resulting in fluid and dexterous movements.
Just as visual images can be imagined, so can motor productions. For example, vivid movements occur in dreams. They are accompanied by signals sent toward muscles but blocked by the mechanisms that inhibit movement during REM sleep).
We know that neural impulses generated by imagination during dreams are capable of powering realistic body movements. In the sleep disorder known as REM sleep disorder, people act out their dreams, sometimes endangering themselves or others.
What is evidence that dream movements are partly transmitted to muscles?
Motor imagination can be stimulated by observation. Hebb (1960) noted that we naturally put ourselves in another person's place when watching body activity. Hebb reported that when he was in a back brace after an injury and had to stand up very carefully, he winced when he saw somebody stand up quickly.
Similarly, many people wince when they see somebody do a difficult gymnastic move such as the splits. It does not hurt the person who does the splits, but the same movement would hurt the observer.
Such motor empathy probably helps humans learn by observation. When we closely observe an activity, part of our brain is acting it out in imagination.
Hebb's speculation was supported by the discover in the 1990s of mirror neurons: neurons that fire patterns in the motor cortex of observing animals, similar to neural activity in the motor cortex of animals actually producing a movement.
What did Hebb point out? How might motor empathy help humans learn?
Motor empathy occurs, but motor imagination is even more realistic. Carpenter (1985) described an experiment that showed subjects could not remember whether they had actually traced a design or just imagined tracing it. They did not suffer similar confusions between acting out a movement and watching somebody else act it out.
Guillot, Di Rienzo, MacIntyre, Moran, and Collet (2012) reviewed the considerable literature on motor imagery. They drew several conclusions:
The last point supports athletes using imagery to prepare for sporting events. They may actually experience learning and improved performance, just by imagining the appropriate movements.
Why might visualization help athletes?
The authors also note long-standing observations of electromyogram (EMG) activity in muscles, sometimes showing traces of activity during motor imagery. This does not always occur. Several studies have reported no EMG activity during imagined movements.
Because the EMG activation sometimes occurs in peripheral muscles during MI, this supports the assumption that imaginary movement is generally inhibited at the brain level. Sometimes a bit of activity can "leak through" to the body's muscles.
Consistent with the observations of REM sleep disorder, specific forms of brain damage can eliminate this inhibition. "Patients with specific brain damage fail to inhibit the motor action associated with its mental representation, and thus fully 'execute the imagined action.'
In a famous analysis of motor behavior, Hailman (1969) pointed out that even the simple pecking behavior by birds can be quite complex and creative. The peck is stereotyped in form, so one might assume it is the same every time it occurs. But a close analysis shows the peck is slightly different each time it is executed.
A bird adjusts its aim in subtle ways and can peck the same spot from a variety of starting positions. Hailman's work reminds us that there is just as much creative variation in motor activity as there is in perception or language.
What did Hailman point out about pecking?
There is no such thing as a behavior that is repeated at different times, unless you define behavior by its effect or consequence in the environment (as behaviorists define operants). The functional definition of behavior was necessary for B.F. Skinner because it is outcomes that are repeated when a rat hits the bar in a Skinner Box, not exact movements.
If you define behavior by the exact sequence of muscles activated, each repetition of a behavior is different. Bartlett (1932) expressed this well when he wrote:
Suppose I am making a stroke in a quick game, such as tennis or cricket. How I make the stroke depends on the relating of certain new experiences, most of them visual, to other immediately preceding visual experiences and to my posture, or balance of postures, at the moment...
When I make the stroke I do not, as a matter of fact, produce something absolutely new, and I never merely repeat something old. The stroke is literally manufactured out of the living visual and postural "schemata" of the moment and their interrelations.
I may say, I may think that I reproduce exactly a series of textbook movements, but demonstrably I do not... (pp. 201-202)
What point did Bartlett make about simple movements in 1932?
Handwriting is an example of a complex motor production that naturally varies each time it is produced. This predictable variation was an important clue for handwriting expert Charles Hamilton Jr. when he debunked the fake "Hitler Diaries" sold to the German magazine Stern in 1983.
He knew the diaries were not actually written by Hitler because the handwriting was too consistent. "If Hitler had written these spontaneously, each letter would be a little different," he said, "whereas a forger would copy the way Hitler formed his 'a' and would be afraid to vary it." (McFadden 1983)
Why was Hamilton skeptical of the Hitler diaries?
Human speech is a marvel of motor coordination, as explained by Fry (1979):
In the space of 30 milliseconds it may be necessary for the brain to switch off a noise generator at the front of the mouth, switch on the larynx vibration, move the soft palate, change the configuration of the tongue, modify the lip shape, and initiate a certain pattern of frequency change by the vocal folds. These actions are not synchronized but must be carried out with various specified leads and lags [i.e. careful timing]. (Fry, 1979)
To summarize, motor activity is much like other cognitive activity. It is (1) based on schemata, (2) subject to great variation, and (3) displays great creativity.
Bartlett, F. C. (1932). Remembering: A Study in Experimental and Social Psychology. Cambridge, UK: University Press.
Carpenter, E. (1985, September). Not all in your head. Psychology Today pp.8-9.
Fry, D. (1979, February). How can you say that? Human Nature, p.38.
Guillot, A., Di Rienzo, F., MacIntyre, T., Moran, A., & Collet, C. (2012) Imagining is Not Doing but Involves Specific Motor Commands: A Review of Experimental Data Related to Motor Inhibition. Frontiers of Human Neuroscience, 6, 247.
Hailman, J.P. (1969) How an instinct is learned. Scientific American, 221, 98-106.
Hebb, D. O. (1960). The American Revolution. American Psychologist, 15, 735-745.
McFadden, R. D. (1983, April 25). Skepticism growing over Hitler diaries. New York Times, p.1.
Neisser, U. (1967). Cognitive Psychology. Englewood Cliffs, NJ: Prentice Hall.
Schmidt, R. A. (1982) Motor Control and Learning. Champaign, Illinois: U Illinois Press.
Write to Dr. Dewey at firstname.lastname@example.org.
Copyright © 2007-2017 Russ Dewey