KAAP 428 Exam 2

what is motor performance?

what we measure when a person performs a skill.

-can be an outcome or process measure

outcome measures

indicate the outcome of performing a skill:

-accuracy

-time/speed

-magnitude

Why measure error?

accuracy is a major component of human skill from everyday tasks to sports performance. The way in which we understand the accuracy of performance is simply to measure the extent to which performance differed from a criterion

how do we measure error?

the criteria can be a specific target in space, such as an archery target or a time, such as matching a rhythm

Dimensionality of Error Scores

-1 dimension (x-axis)

-2 dimensions (x and y-axis)

-3 dimensions (x, y, z-axis)

Information Obtained from Error Scores

a dichotomy (hit/miss)

-there can be the same number of hits, but performance can differ if one person is more consistent than the other

absolute error (AE)

the absolute difference in relevant units between the criterion and the performance outcome

-abs(xi-T)

-error along a single dimension

-we take multiple trials to gain a representative measure of performance - the average

-AE = sum of(abs(xi-T))/n

radial error (RE)

absolute error for two dimensional tasks

-takes into account both x and y dimensions

RE = SQRT(Error x^2 + Error y^2)

(uses pythagorean theorem)

-to find radial error for multiple target locations or trials, take the sum of all RE's and divide by n (number of trials)

Bias in Performance Outcomes

absolute measures of accuracy may hold insufficient information. For example, they fail to describe tendencies to over or undershoot the target.

constant error (CE)

represents the magnitude of error in a specific direction (no longer absolute)

CE = sum of (xi -T)/n

variability in performance outcomes

a measure of consistency in performance. the typical measure is the standard deviation (SD)

SD = SQRT( (sum of (x-m)^2) / n-1 )

variable error (VE)

an index of how much variability there is in the accuracy of performance

who would be more skilled: lower AE but higher VE or higher AE but lower VE

lower absolute and constant error means that they are closer to the target on average, but higher variable error means they are inconsistent. The individual with lower variable error is more skilled because they have more consistency, despite being further from the target on average (lower AE and CE)

Continuous Skills

root mean square error (RMSE)

-error between a participant's displacement curve and a criterion (ideal) curve

-computes one error score for the duration of the task

performance outcome measures of accuracy

-dichotomy (hit/miss)

-absolute error (AE)

-radial error (RE)

-constant error (CE)

-variable error (VE)

-RMSE

reaction time (RT)

the interval between the onset of a signal or stimulus to the initiation of a response

simple reaction time

there is one stimulus and one reaction once that stimulus is perceived. the quickest response time for types of reaction times

-ex: hearing a starting gun go off and beginning to run your race

choice reaction time

A measure of the speed of mental processing that takes place when someone must choose between several responses, depending on which stimulus is presented

-ex: there could be stimuli of 3 different colors with a specific response for each color. when a specific color is identified, the correct response needs to be produced

-ex: looking at a stop light and determining what to do based on the color.

discrimination reaction time

there are multiple stimuli and you have to react to the correct stimuli using the best response

-ex: determining the pitcher is throwing a fastball as opposed to a curve ball and hitting responsively) (3:1 stimuli to response)

-ex: 3 different color stimuli are possible. a reaction should only be produced when one specific color is shown

-slowest of all reaction time tests

-used in information processing studies

what are typical reaction time values for simple RT

light = 240 ms

siren = 220 ms

electrical shock = 210 ms

all together = 180 ms

Typical RT values for choice RT (based on number of choices)

1 = 200 ms

2 = 350 ms

3 = 400 ms

4 = 450 ms

5 = 500 ms

6 = 550 ms

7 = 600 ms

8 = 600 ms

9 = 650 ms

10 = 650 ms

reaction time and sprinting

in track and field, the governing body (IAAF) has a rule that if the athlete moves within 0.10 seconds after the gun has fired, the athlete has false started

-this is because a human cannot react that quickly and if they did, it means they were not moving based off of the sound but instead predicted the sound and moved before they perceived it

movement time (MT)

the interval of time from the initiation of the response to the completion of the movement

response time

the sum of RT and MT. from the onset of a "go" stimulus to the completion of the movement

EMG in RT measures

EMG, which shows the electrical activity of the muscle, has been used to separate RT into central and peripheral components

-research shows that following the presentation of a stimulus, for a portion of the RT, there is no electrical activity. This 40-80 ms period is known as the pre-motor RT and represents central processes (decision making, perceptual processes, etc)

-there is a motor RT period where there is electrical activity before movement

-the pre-motor RT + motor RT is the reaction time

Electromyography (EMG)

-measures the electrical activity in muscle

-electrodes are attached to the skin superficial to the muscle belly

-electrodes detect electrical activity which result in muscular contraction. Electrical signals are amplified and recorded

-data described temporal patterning, and amplitude

anticipatory signal in gastrocnemius during motor reaction time

-gastrocnemius muscle fires before biceps in a pulling task

-provides postural support for movement of the arm

-predicts postural disturbance and generates stabilizing response

measures of magnitude

indicate the "size" of an outcome, and have a particular relevance in sports settings

-distance

-height

-weight

Performance process measures

tell how a result was achieved; what underlies performance

Examples:

-kinematics

-kinetics

-EMG

-brain acitivty/imaging (EEG, PET, fMRI, MEG)

kinematics

measures which describe motion, without regard to the cause of that motion

-modern systems (Optotrak) use infra-red technology to relay the saptio-temporal positions of markers

-can provide raw data in the form of x, y, and z coordinates. From this, we can calculate:

-displacement (linear/angular)

-velocity (linear/angular)

-acceleration (linear/angular)

displacement

change in spatial position

velocity

rate of change in position (displacement)

acceleration

rate of change in velocity

kinetics

measurements of the forces which cause motion (mostly applying Newton's laws of motion)

Examples:

-force: push or pull on an object; product of an object's mass and acceleration (e.g. ground reaction force)

-torque: angular force directed around an axis of rotation; product of force and perpendicular distance to the axis

-momentum: product of an object's mass and velocity

equipment includes force platforms, strain gauges, etc

before you start measuring peformance...

Is it an objective measure?

-a measure is objective if it can be employed consistently by different people. It is also objective if the measurement scale is appropriate.

Is it a valid measure?

-this refers to whether a test measures what it is supposed to measure. Does your measure have construct validity? Magnitude measures almost always do. However, is accuracy, consistency, bias a construct of your skill?

Is it a reliable measure?

-is the measurement repeatable? Deviations in the way a test is performed can result in markedly different results. As a result, change may be incorrectly attributed to difference in performance.

improving your tests' objectivity, reliability, and validity

1. consider the purpose of the skill

2. keep the test environment consistent

3. document your methodology

4. standardize measures

5. don't test yourself!

Scientific Method

Step 1: develop a testable research question

Step 2: formulate hypotheses

Step 3: operationally define independent and dependent variables

-independent variable- the "manipulated" variables/factors

-dependent variable- the measured variable (presumably influenced by the independent variable(s)

Step 4: design study to test research

-subjects, trials, controlling extraneous variables, etc

Step 5: observe behavior and gather data

Step 6: analyze and interpret results of the study

-descriptive and inferential statistics

the cognitive approach to psychology

examines processes of thinking, language, remembering, learning (no longer ignore cognitive processes, but infer from them research to understand what goes on)

information processing

the process of thinking is analogous to how a computer processes information

-stimulus identification, response selection, response programing

-sense data (meaningless) --> interpreter --> meaningful information --> decision center --> meaningful action

Assumes that information in the environment must be:

-accepted into a storage system (memory)

-processed (coded)

-input/signals (environmental information) --> processing (human brain) --> output/motor action (observable behavior)

Helmholtz' Epistemic Mediation

3-D world is seen by the eye and is 2-D on the retina where the brain then makes sense of it. This requires epistemic mediation.

-the processes that facilitate the alignment and integration of diverse knowledge systems to create a shared understanding and collaborative action, like in complex contexts

Turing Machine

Alan Turing - British mathematician

-in 1936, developed the concept of a simple "theoretical" machine, consisting of:

-an infinitely long tape divided into identical squares

-a scanner to read the tape

the machine could:

-move the tape to the right

-move the tape to the left

-print a slash

-erase a slash

-proved mathematically that through this "binary code" one could devise and execute an indefinite number of programs and that machines operating on this principle could be constructed (computers)

computer programs

John von Neumann - mathematician

-in 1940s made an idea of a program to instruct the Turing machine

-idea of a stored program within the computer's internal memory. A computer can prepare and execute its own programs.

neuronal model

they showed that:

-the operation of a nerve cell and its connections could be modeled in terms of "logical elements"

-the logical scheme is that of a neuron being activated and then causing another neuron to fire

cybernetics

Norbert Wiener - engineer

-1948, published Cybernetics which integrated developments in the understanding of the human nervous system, the electronic computer as well as the operation of other machines

-introduced the idea that machines are purposeful. Those that exhibit feedback can be considered to be "striving towards goals"

information theroy

in 1940s, Shannon developed the idea that information can be thought of as simply a single decision between 2 alternatives

-basic unit of information = bit (binary digit)

-electrical circuits could embody fundamental operations of thought

Neuropsychological syndromes

after WWII, mental pathologies due to brain injury were heavily studied

-labs globally found considerable convergence across cultural and linguistic boundaries

-ex: aphasia assumed similar forms despite wide differences across languages

common thread of cognitive process findings

-one can consider cognitive processes apart from any physical embodiment or structure

-the focus is instead on the efficacy of any transmission of information via any mechanism

-human brain is a thinking machine which may be understood through the concepts of information processing

information

a quantity that reduces uncertainty

-role as a constraint specifies the pattern to be learned, memorized, performed

information processing model overview

-information is thought to proceed through a set of identifiable states before movement occurs

-analyze various mental (and unobservable) mental operations that precede action

-what is the role of the control system in sensing, attending to, transforming, retaining, and transmitting information?

chronometric approach for inferring cognitive processes

-examines the temporal aspects of information processing, concentrating on the duration of processes

-extensively uses RT studies. By carefully controlling the test environment, we can infer that changes in RT are a result of a cognitive process

assumptions of the information processing model

1. numerous processing stages that occur between stimulus and response

2. sequence of processing stages is initiated by stimulus presentation

3. each stage operates only on information available to it

4. each stage transforms in some way the information supplied to it, an event which takes time

5. upon completion of processing performed at one stage, the transformed information is made available to the next stage of processing

Donders' RT studies

compared to a simple RT, choice RT and discrimination RT require additional processing which account for greater time

-discrimination RT: stimulus discriminaiton

-choice RT: stimulus discrimination --> response selection

parallel vs serial processing

Parallel- happens at the same time

-electronics assembly, engine assembly, body assembly all happening simultaneously

series- one part comes before another

-final assembly --> test drives

parallel vs serial processing response time

RT studies are through to indicate that human IP occurs partly in parallel and partly serial

stimulus identification stage

-represents the detection of a stimulus, and its recognition as part of a pattern

stimulus detection:

light in the optic array --> transformed into code of neurological impulses --> more processing to contact memory

-increasing stimulus clarity and intensity shortens the stimulus identification stage

Pattern Recognition

-most pattern detection is learned, and often differentiates experts and novices

static- chess

dynamic- patterns of play in soccer

-automaticity in motor performance develops due to sensory info being detected quickly and accurately

response selection stage

the perceiver takes the "knowing" of the environment achieved in the stimulus detection stage and selects an appropriate response

Number of Stimulus-Response (S-R) Alternatives

in choice RT, increased time with increased number of choices represents increased processing of information at the response-selection stage.

-curvilinear relationship (Hick's Law)

Hick's Law

choice RT is linearly related to the log of the number of stimulus options (response selection stage)

-choice RT = k log2 (N+1)

N= number of S-R alternatives

k is a constant (usually simple RT)

-in information theory, log2 represents a binary unit (bit) of information. it is a yes/no alternative

- 1 bit decision = 2 alternatives (yes/no)

- 2 bit decision = 4 alternatives

- 3 bit decision = 8 alternatives

-increases in RT are a curvilinear function of the number of alternatives

-increases in RT are a linear function of the number of bits of information

-increases in RT are a linear function of the log of the number of alternatives

-correctly predicts RT increases with increasing choice alternatives, and predicts size of increase

2 caveats:

1. excess familiarity/practice and the law doesn't hold

2. unequal probability of possible choices

stimulus-response compatibility

S-R compatibility concerns the extent to which the stimulus and its associated response are connected in a "natural" manner.

The Simon Effect:

- reaction time is fast when task is press left key when you hear a sound in left ear and right key when the sound is in the right ear

- when you have to press the right key for a left sound and the left key for a right sound, RT decreases significantly

-incompatibility appears to cause interference in selecting a response

-response selection phase of RT is susceptible to compatibility effects when the stimuli and respose have a learned association

-interference occurs if response is not consistent with that learned association

The Stroop Effect

-when given a list of colors that are a different color than the word says, it is difficult to only state the color of the word and not what the word actually says

-we learn to process the color and meaning in parallel. This task would be easier for young children who have learned their colors but not learned to read

-there is response competition, causing delay. There may also be delay in stimulus identification

response programming stage

the third stage of information processing; during this stage, the motor system is organized to produce the desired movement

-The Henry-Rogers experiment

- manipulated complexity of the movement, while keeping the stimulus the same

- complex movements required not only longer MT, but longer RT, suggesting longer movement programming/planning

- the more complex the movement, the longer period of time to initiate the movement (identification/selection the same across conditions)

characteristics of a good theory

-general

-testable predictions

-elegant

-statistical

a theory of motor control...

describes unobservable structures or processes and relates them to each other and to observable events

-makes predictions for future events, guides practitioners

-ID performance problems, develop interventions, predict effectiveness of interventions, develop systematic practice methods, create new intervention strategies, evaluate new interventions

closed-loop control systems

commands are modifiable via feedback

-feedback --> movement control center --> movement commands --> movement effectors --> feedback

-mechanical example: thermostat in the house

-human motor skill example: driving a car

open-loop control systems

commands are pre-set, there is no feedback during the action

-movement control center --> movement commands --> movement effectors

-mechanical example: videocassette recorder

-human motor skill example: throwing a dart at the dartboard

motor program theory

-a motor program is a pre-structured set of motor commands selected by the CNS and sent to the peripheral nervous system

-it is a memory-based representation that stores information needed for action

-the concept of a motor program is couched in information processing and the computer analogy

early motor program theory

-all control is cortical (top-down)

-sensory inputs not necessary or used

-control directed at muscle level

-separate programs for each specific movement

-"a set of muscle commands that are structured before a movement sequence begins that allows the entire sequence to be carried out uninfluenced by peripheral feedback"

rapid movements

sensory feedback processing is too slow to correct a movement

-open-loop control

accurate movements without sensory feedback

normal monkeys learned a no-vision pointing task, then underwent surgery to cut the sensory nerves from the right arm. They were still able to perform the task

inability to stop a movement

-when a movement is initiated less than 80 ms before "target"

EMG patterns with and without unexpected blocking of the movement

EMG patterns are the same if the movement is unblocked or if it is unexpectedly blocked

motor equivalence

the ability to produce the same movement outcome with a variety of different muscle groups or effectors

the ghost in the machine

concept that the body is separate from the mind and that there is a consciousness apart from a physical entity

-transfers the "problem" of control to the "little human in the head"

-understanding --> eliminating the homunculus

degrees of freedom

the number of independent elements to be controlled

degrees of freedom problem

how can a complex system act to constrain so many degrees of freedom into a functional unit?

defining the DOF of the human arm

7 degrees of freedom

-3 at the shoulder

-1 at the elbow

-1 at radioulnar joint

-2 at the wrist

defining DOFs

DOF complexity also in terms of: # of planned movements

-more planned movements --> longer RT

BUT

RT of sprinters equivalent to key press about 200 ms

-components and number of movements involved in starting to sprint is far greater than a key press

-how do skilled athletes decrease RT for complex movement tasks?

equations of constraint

-consider 2 components, A & B, represented in a 2D space

-need 2 coordinates for each element (4 DOF)

-same 2 components now connected by a line: (xB-xA)^2 + (yB-yA)^2 = L^2 (equation of contraint)

-if coordinates of A are chosen, then only one coordinate of B can vary freely

-REDUCTION IN DOF FROM 4 TO 3

# of DOF = N * D - C

N= # of elements

D= # of dimensions

C= #of equations of constraint

EX: 2 independent components A & B: DOF = (2*2)-0 = 4

2 components with a line: DOF = (2*2)-1 = 3

how to simplify the control problem

link the parts with equations of constraint

in an aircraft: there are 2 ailerons, 2 elevations, 1 rudder

1. when left aileron goes up one position, rudder goes right by one position. position of left aileron = rudder position

2. when right aileron goes up one position, rudder goes left one position. Position of right aileron = -rudder position

2 ailerons + 1 rudder now = aileron-rudder subsystem (3 elements --> 1 element)

2 elevators can also be linked:

3. right elevator position = left elevator position (Le = Re)

now link aileron-rudder to elevator

4. aileron-rudder system varies in constant proportion to elevator position (AR = k(E))

DOF = ND - C = (5*1)-4 = 1

Original 5 DOF system now only 1 DOF

example: internal model - "anticipatory" postural responses

-muscle synergies are linked "remotely"

example: skilled behavior - in a professional sharpshooter, any movement at the wrist is matched by an equal and opposite movement at the shoulder

learning a motor skill?

setting up the appropriate equations of constraint:

-reduces DOFs

-simplifies control

-enables better precision, accuracy, etc

unit of control- not individual muscles/limbs, but groups of muscles/limbs linked functionally (a coordinative structure)

laws of control - constraints on action

-Fitt's Law - the speed accuracy tradeoff

-Ecological theory of perception-action

-dynamical systems theory

Fitt's Law: Speed-accuracy trade-off

MT - a + b log2(2D/W)

-a and b are constants dependent on body parts, tools used, etc.

D = distance of movement to object

W = width of target

more easily used:

The index of difficulty = log2(2D/W)

ecological theory of perception-action

direct perception

-minds are not like computers

-the environment contains information for action in the optic array without need for cognitive processing

-the environment provides affordances for action - information in action-relevant terms

optic array

instantaneous pattern of light reaching space from all directions (moving = optic flow field)

Direct perception in visual control of action

Tc = distance/velocity

-avoidance of collision, intercepting objects

-diving gannets (birds)

-retinal image increases as object approaches

dynamical systems theory

-behavior arises from interaction of multiple subsystems

-behavior emerges through self-organization - not prescribed

-patterns are the unit of analysis

executive control: brain tells muscles what to do

self-organized: sees what has to be done and responds accordingly

-abrupt changes in behavior represent a reassembly of components into new patterns that are better suited to accomplish the task

-ex: a horse uses a certain amount of oxygen when walking at maximum walking speed. They use the same amount of oxygen when moving faster when trotting. This is because the new pattern (galloping) is better at accomplishing the task (moving fast)

dynamical systems phase transitions - bimanual coordination

together = in-phase

alternating = anti-phase

fingers want to alternate even if you start moving them togethet

constraints

act to limit the observed behaviors

-organism

-environment

-task

understanding sensorimotor learning

learning is a relatively permanent change in the ability to perform a skill

-learning cannot be directly measured - it is inferred from behavior and performance

performance:

-observable behavior

-temporary

-not necessarily a result of practice

-influenced by performance variables

Characteristics of the learning process

1. performance of skills shows an improvement over time

-early learning = large errors

-late learning = little to no errors

-many learning studies examine reduction of motor error, which allows us to be able to infer that learning has occurred

2. performance becomes more consistent (less variable)

3. there is a greater persistence in the improvements made

-improvements are made and stick around

-relearning is quicker than learning

4. performance of skill becomes more adaptable (generalizes)

-rehabilitation takes advantage of the learning process. Doing one skill to build strength that will allow the person to perform a different, similar skill

how do we quantify learning?

performance curves are a method of assessing learning by recording levels of performance across practice

-dependent variable for learning vs time

-dependent variables may be absolute error, variable error, RT, etc

performance curves

learning a new skill is typically characterized by one of four performance curves

-linear, positively accelerated, negatively accelerated (ceiling effect), S-shaped (ceiling effect)

-negatively accelerated curve is most common in motor skill learning

-representative of the classic power law of skill learning

performance curves in kinematics

performance curves can also show changes between and within trials

-improvement in performance can be assessed by how close the movement pattern matches the criterion

-consistency can be assessed by the extent to which the movement patterns vary

assessment of learning

inferring learning from practice can cause problems:

1. practice data shows no evidence of permanent change in behavior

2. performance during practice is susceptible to over and under estimating learning

3. performance in practice may temporarily plateau

assessment of learning by retention

typical design of studies that aim to assess learning by retention:

-pre-test: measures ability to perform the task before the treatment

-practice/acquisition

-post-test: measures ability to perform the task after a period of practice

-retention test: 1-7 days after post-test; measures ability to perform the task after a no-practice retention interval

-comparing pre and post test allows us to examine the effect of practice on performance

-comparing pre-test and retention allows us to examine the effect of practice on learning

-comparing post-test and retention allows us to examine decay of performance

retention paradigm

-tracking task conducted with practice over 2 days

-retention tested at 2 time points

--immediate (same day)

--delayed (1 day later)

-50% = people that progressively received less visual feedback throughout practice

-no difference in acquisition between the two conditions, but the 50% group had better retention

-different forms of skill practice can impact retention

assessing learning by transfer

describes how previous practice on a task influences the learning of a new skill

transfer can be:

1. negative: where previous practice of one skill hinders learning of a new skill

2. positive: where previous practice in one skill assists learning of a new skill

significance of transfer

1. it can define the appropriate sequencing of skills to be learned

-curricula tend to be organized in a simple-to-complex order

-early fundamentals need to be in place before moving on

-skill classifications can be a useful tool to guide transfer

-rehabilitation clinicians need to treat impairments in a functional order

2. it can assess the effectiveness of practice conditions

-does practicing a certain skill result in better performance

Can learning in an aiming task transfer to different directions?

1. pre-test conducted at 5 different target distances

2. subjects practiced moving the cursor to only one distance (54 degrees)

3. post-test shows that learning from the 54 degree target carried over to all the other tested distances