Introduction

Introduction to the History & Problems of Motor Control

August 28, 2009

I. History of motor control in psychology

A. "Modern" psychology has ignored movement for the most part...

1. Thought that movement is simple - animals can do it!

2. Real human abilities are perceptual and cognitive

3. Modern cognitive psychology arose from 'information processing' and so was aligned more with computer science.

a. e.g., computers can be programmed to play chess

B. ...but it should not.

1. Real movement is complicated

a. e.g., building operational robots has been extremely challenging.

(1) Imagine building a chess-playing computer that actually picked up and moved the pieces.

2. Further, none of the movements we make as humans are fundamentally much different from what other animals do.

a. Activities that have been regard as distinctively human...

(1) communication - speaking, writing, pointing

(2) tool manufacture and use

b. are based on skills that are evident in other species

(1) chimpanzees have rich vocal repertoire of communication and they use tools, e.g., sticks to get ants.

3. Also the relation between perception and action was a dominant theme at the birth of experimental psychology in 19th century.

a. Wilhelm Wundt

(1) established first laboratory of experimental psychology in 1879 at University of Leipzig

(2) 1881 founded first journal for psychology

(3) Wrote first textbook, Principles of Physiological Psychology

(a) discussed motor control in his text under 'Volitional Processes'

b. William James

(1) established psychology at Harvard, 1876

(2) conceived of psychology as an experimental science based on physiology of the brain and not just philosophy

(3) wrote 2 volume Principles of Psychology in 1890

(a) devoted 3 chapters - "The production of movement", "Instinct", and "Will".

c. In general the view was of multilevel sensorimotor control in which volitional action was built on foundation of reflex. Experimental data used in support of ideas included reaction time as today.

4. Two problems were recognized

a. What is the role of perception in the production of action?

In other words, what is the origin of basic action?

(1) James proposed that perception and imagery was origin of action:

(a) "We may ... lay it down for certain that every representation of a movement awakens in some degree the actual movement which is its object; and awakens it in a maximum degree whenever it is not kept from so doing by an antagonistic representation present simultaneously to the mind. Movement is the natural immediate effect of feeling, irrespective of what the quality of the feeling may be. It is so in reflex action, it is so in emotional expression, it is so in voluntary life. (Principles of Psychology, 1890, p 526, 527)

(2) but precisely how "feeling" resulted in movement was never clear

b. What is the role of action in perception?

(1) Commonly accepted since Berkeley that perception was determined by action.

(a) e.g., what you see depends on where you look.

(2) But there was a problem of how space itself was perceived and how spatial relations between agents and objects in the world could seem constant when perception varied by action so much.

(a) e.g., with each eye movement how does your brain “know” that it was your eye and not the world that moved?

(b) Space constancy was solved by the idea of corollary discharge

i) a neural representation of a movement against which the movement's sensory consequences are matched. Idea traced to Helmholtz in Physiological Optics, but gained real currency in 1950s (von Holst & Mittelstaedt, Sperry, Teuber).

ii) There were 2 competing notions about how this happened.

a) On the one hand, sensory information could be merged with innervation feelings to produce spatial perception.

b) On the other hand, kinesthetic feedback for each movement could be used.

c) Thus, in 19th century the 'inflow' vs. 'outflow' debate had started which continues today!

C. At the beginning of the 20th century there developed a schism between the investigation of perception and of action with the birth of behaviorism and of gestalt psychology

1. Classical experimental psychology died because it proved unable to answer practical questions

a. e.g. what are the best educational practices?

2. One shortcoming of 19th century psychology was its claim for special status. Wundt argued that all other sciences were concerned with 'mediate experience', the outer physical world. The ultimate topic of psychology was 'immediate experience'. Thus psychology isolated itself from the rest of science.

3. Behaviorism and gestalt psychology corrected this.

a. Behaviorism was based on biology.

(1) Watson, Skinner

b. Gestalt psychology was rooted in physics.

(1) Wertheimer, Köhler, Koffka

c. This close relation with the growing fields of biology and physics made it possible to effectively wipe out experimental psychology as an intellectual field in 1 decade, 1910-1920.

d. Both behaviorism and gestalt psychology would claim to offer an account of all the problems that 19th century experimental psychology addressed - including movement. But in practice they would fall short and lose sight of problems and issues dealt with in 19th century.

e. The classic questions were 1) How is sensory information used to control action? and 2) How is perception influenced by movement?

(1) Behaviorism said little if anything about #2, and disposed of #1 by committing itself to explaining which stimuli lead to which responses and not how the relation happened.

(2) Gestalt psychology did not directly address action. Movements were thought to occur to result in better organization of the psychophysical field. It does not explain how movements are selected and executed.

(3) Both behaviorism and gestalt psychology had simplistic notion that movements were automatically executed once the sensory stimulus or perceptual state was dictated. Thus, from 1920 to 1950 research on cognitive processes and research on control of behavior took place in different labs (different countries) under different theories.

D. The 1960s was a period of reunification of experimental psychology with the development of the information processing perspective which arose from a number of influences: 1) human factors research in WWII, 2) Shannon's information theory of communication, 3) cybernetics, 4) computer technology

1. Many of the problems of the 19th century were rediscovered

2. This brand of psychology was not a school per se, being less based on a theory than on a metaphor or set of metaphors that related brain to the most advanced computational machinery and concepts available at the time.

3. Early on, action was part of the package in the sense of controlling processes. One of the most influential publication series in the field is entitled Attention and Performance.

4. But there was a legacy of the original split between perception and action. What has become cognitive psychology became more concerned with perception and cognitive processes and until recently virtually ignored movement.

a. A founding father of the approach, Ulric Neisser, described the central tenet of cognitive psychology as "attempting to trace the fate of the input", a pursuit that stops short of control of overt action. Thus psychology has typically stopped at memory and thought.

b. BUT even just to do the experiments required overt movements

5. Why did cognitive psychology ignore movement?

a. The nature of the computational metaphor.

(1) Most computers had input and processing; but there was no action.

(2) Also, computational machinery has until recently been too limited to be very effective in controlling processes in real time.

b. Cognitive psychology picked up the work on the problems left by behaviorism and Gestalt psychology which were mainly perceptual and cognitive. Also as interest moved toward higher cognitive abilities like mental representations, categories & concepts, and language, primary action was forgotten.

E. Current status

1. With development of connnectionism and neural networks in mid 1980s, cognitive psychology is returning to its biological roots.

a. “Cognitive Neuroscience”

2. Development of active perception approach in robotics

a. Traditional AI approach to problems of perception yielded ill-posed problems

(1) e.g., specifying the 3-d reconstruction from a stereo image - in a single snap-shot frame - has more variables than there are equations.

b. However, if the cameras moved, then the resulting mathematical problems had explicit solutions.

c. New generation of robots that link perception and action

F. Important to learn that psychology (as all of science ) has not progressed cumulatively. The work of one line of thought can dry up and be replaced by views and findings of different approaches.

II. Overview of central issues of motor control

A. Two fundamental aspect of motor control to remember

1. How are successful movements generated?

2. How is stability maintained?

B. Consider physiological & psychological levels of explanation

1. e.g., Parkinson's disease

C. Four major problems of motor control

1. The Degrees-of-freedom problem

a. Define motor equivalence - ability to achieve goal, perform task using variety of movements, e.g., writing with either hand, foot or mouth

b. Specification of the parameters of movement is computationally complex because there are so many degrees of freedom

(1) e.g., arm has 7 degrees of freedom (shoulder = 3, elbow = 2, wrist = 2)

c. Solution of efficiency

(1) avoid extreme joint angles

(2) minimize jerk (derivative of acceleration)

(3) minimize changes in muscle torque or muscle stiffness

d. Synergies

(1) dependencies between components of motor system

(a) e.g., blink during sneezing

(b) rhythmic movements with 2 hands have similar frequency

i) but skilled jazz drummers can hold 2 different beats with the two hands!

e. Relying on mechanics

(1) biomechanical properties simplify degrees of freedom problem

f. Path planning, inverse kinematics, and inverse dynamics

(1) Consider different levels

(a) path planning - which route for limb to move to achieve objective

(b) inverse kinematics - convert desired path into desired joint angles

(2) People generally make straight line movements, even curved trajectories can be shown to be comprised of straight segments

(a) 3 possible reasons

i) Path planning is done with respect to the Cartesian coordinates in which the hand moves. This assumes that path planning is under geometric constraints.

ii) Minimizing jerk. This assumes kinematic constraints.

iii) Minimizing torque changes at joints. This assumes dynamic constraints.

(b) General rule of biological systems is that every possible trick will be used, so each of these factors probably contributes. Thus, high level control is constrained by and influenced by lower level control factors.

2. The serial-order problem

a. How do we produce sequences of movements

b. Speech errors

c. Action slips

d. Coarticulation

3. The perceptual-motor integration problem

a. Feedback and feedforward control

(1) ballistic versus controlled corrective movements

(2) use of error to guide movement is negative feedback, servomechanism

(a) control systems approach to motor control

i) reference signal

ii) plant

iii) comparator

iv) gain

v) closed loop

vi) open loop

b. Spatial coordinates

(1) body-centered

(2) environment-centered

c. Movement enhances perception

(1) Perception is necessary for effective movement...

(2) but movement is necessary for effective perception.

d. Suppression effects

(1) When appropriate, movement impairs perception

(2) Necessary to distinguish sensory changes that are due to movement of the world from changes due to movement of the sensory receptor relative to the world.

D. The skill-acquisition problem

1. Extent of innate skills?

2. How are skills learned?

3. Once learned, what is form of memory?

4. Example of vestibulo-ocular reflex

5. More complex examples - playing the piano, hitting a ball, performing surgery, etc.