Psychophysiological Methods

1
Psychophysiological Methods
2
Electrodermal Measurement

• Galvanic skin response as indicative of the
  sympathetic branch of the autonomic nervous
  system
• Sweat glands provide a shunt between skin and
  deeper tissues
• Measures indicative of arousal, stress-strain,
  and emotion
• Autonomic habituation provides a physiological
  measure of information processing capacity needed
  to complete a task
• Used as measure of workload, mental strain, and
  emotional strain.

3
Procedure

• Sampling across various locations of the body
• Typically 3-4 cm distance between electrodes
  sampling dc current using a bioampifier
• Sampling at 20hz sufficient to calculate Skin
  Conductive Response (SCR)
• Amplitude, rise time and recovery time are
  measured
• May be used to determine tonic Electrodermal
  Activity (EDA) to measure readiness for action

4
Three-Arousal Method of Measurement for the use
of Psychophysiology in Ergonomics
5
(No Transcript)
6
Typical Epidermal Response
7
Advantages/Disadvantages

• Easy to measure and interpret the physiological
  signal
• Pure measure of the sympathetic branch of the ANS
• Sensitivity to workload and emotional strain

• Somewhat difficult to record
• Prone to artifacts in non-laboratory settings
• Indiscriminately sensitive to any ANS activity
• Several months of lab training to be able to use
  plus training for use in an ambulatory setting

8
Reliability/Validity

• Short term reliability (within days) is fairly
  good (.80 to .90)
• Longer term reliability is more limited (.60)
• Tonic EDA more reliable than SCL (test-retest
  correlations of .76 and .61 at one year)

• Validity at or above .90 for EDA-Emotional
  strength in LAB setting
• No similar data for applied settings
• Validity based more on strength of emotion
  related to strain than physical relationship
• Heart rate and BP are yield better validity than
  EDA (.68 to .86)

9
Electromyography

• Studies muscle function using electrical analysis
  of signals emanating at muscle contraction
• Motor activity
• Anterior horn of the spinal cord, transmitted via
  alpha motor neurons to muscle
• Each muscle fiber consists of multiple chains of
  contractile sarcomeres (actin-myosin-filaments)
• These filaments create muscle contraction
• Motor unit chemically activates the muscle fibers
  connected as myoneural junction is depolarized
  (amplitude of about 100mV with a 2-14msec
  duration
• Muscle action potential causes sarcomeres to
  contract
• Electrodes in tissue or skin can measure these
  action potentials (electrolytic response)

10
EMG (continued)

• Muscle force defined by motor units activated
• EMG forms a quasi-randomly shaped spikes of
  amplitude and duration but no identifiable
  sequence
• Correlation between number and intensity of
  generation of amplitude spikes and muscle
  contraction force
• Remember, doesnt measure force, joint position
  but rather voltage associated with local muscle
  recruitment

11
Use of EMG in the Workplace

• Used for workplace and tool design
• Muscle load (static and dynamic)
• Local muscle fatigue due to overload
• Muscle timing and coordination
• Motor-unit recruitment
• EMGs complemented by use of measures of external
  load, body posture, joint measurement
• EMGs can be performed with needle electrodes
  inserted into muscle or surface electrodes

12
Procedure

• Placement of electrodes or needles in muscles,
  signal passed to preamplifier, processed with
  band-pass filters for frequencies related to
  muscular activity
• Select muscles related to action (may depend on
  how specific you want to be measuring potential
  between muscle and ground
• Amplify, filter and store results
• Signal Processing
• Scaling

13
Surface EMG Signal Processing
14
(No Transcript)
15
Correlation between feedback control and
execution speed during learning an assembly task
16
EMG Advantages/Disadvantages

• Continuous and quantitative measured data
• High temporal resolution with marginal
  interference with task execution
• Allows detection of muscle fatigue at early
  stages providing objective measure
• Multi-channel EMG can identify muscular
  bottle-necks

• Surface EMG limited to muscles directly beneath
  area accessed by skin electrode
• Only feasible for single muscles in individuals
  not too obese
• Requires careful calibration, instrumentation,
  data manipulation, and interpretation
• Setup is fairly time consuming
• Interpretation requires data analysis and data
  integration
• Calibration lacks reliability
• Requires individual calibration, poor reliability
• Needle method more specific but invasive and
  quite painful

17
Heart Rate/Variability

• Various Measures
• Electrocardiogram (ECG)
• Duration between heartbeats (HR)
• Mean heart period or Interbeat Interval (IBI)
• Heart Rate Variability (HRV)
• Normal Rhythm
• Cardiac Sinusoidal Mode
• Modulated by innervations from the sympathetic
  and parasympathetic branches of the ANS
• Heart Rate
• Controlled by nuclei in the brain stem and guided
  by the hypothalamus and prefrontal cortical
  structures
• Two control modes
• Parasympathetic (Vagal) and Sypathetic output
  (Pores (1995))
• Mediation of bororeflex activity

18
Use of ECG

• HRV related to changes in autonomic control
• Vagal Gating
• NSR is vagally determined
• HRV and other cardiovascular variables modulated
  by baroreflex gain.

• General cardiac response found in mental-effort
  studies characterized by increased HR and BP and
  decreased HRV and BP variability at all
  frequencies
• Compatible with fight-flight reaction (lab
  studies, short-lasting tasks, challenging mental
  operations in working memory.
• Mid freq. band most sensitive to variation in
  mental effort due to decreased vagal activation
  and increased sympathetic activation.

19
Procedure

• Three or 7 lead methods
• Sampling and R-peak detection
• Artifact detection and Correction
• Spectral procedures
• HR, IBI or Normalized Values?
• Logarithmic Transformation

20
Disadvantages

• HR and HRV used as indicators of mental effort
• Higher invested effort, higher HR and lower HRV
• Complex relationship between HR with baroreflex
  BP control and autonomous nervous activity

• Most stable results only really found in lab
  settings
• Restrictions in sensitivity for artifacts in
  obtained IBI series and sensitivity for changes
  in respiration
• Artifact correction time consuming
• Newer technologies are mitigating some of these
  limitations

21
Reliability and Validity

• R and V of HR and HRV in short duration mental
  loading lab results usually high
• Doesnt hold for practical settings
• Diagnostic validity and reliability debated

• Sensitivity of the measure not very high
• Difficult to distinguish levels of task load and
  related invested effort
• Requires multiple data collection sessions on
  single subject
• Validity affected by fight-flight mechanism
  versus compensatory mechanisms

22
Ambulatory EEG Methods

• Sleepiness linked with accidents (Comair 5191)
• Sleep loss, long time awake, work at circadian
  trough of physiological activation and alertness,
  monotony
• Effects of drugs, alcohol, sedative, hypnotics,
  antihistamines all can have an effect.
• Concept of sleepiness (various components)
• Subjective
• Behavioral
• Physiological

23
Methods

• EEG sum of electrical brain activity
• Recorded at scalp or needle electrodes
• When alertness falls, frequency of EEG falls and
  amplitude increases as more neurons synchronized
  to fire by the thalamus (rational behind EEG
  indicator of sleepiness)
• Progression from Alpha (8-12Hz) to Theta (4-8Hz)
  to Delta (0-2Hz)

24
Procedure/Example

• See text for in depth discussion of procedure for
  setup and use of EEG and EOG.
• Example 12.1 shows EEG/EOG pattern in severe
  sleepiness performing a task and demonstrates
  beta activity, increased alpha activity, eye
  closure, slow eye rolling movements, and dozing
  off, reappearing beta activity, return of eye
  blinks

25
EEG/EOG Recording
26
Application/Training

• Application takes 15-30 minutes minimum for
  electrode placement
• Learning to setup takes 10-20 hours for reliable
  recording
• Scorring takes several months to learn

• Requires repeated quality checks

27
Reliability/Validity

• No formal reliability established for ambulatory
  EEG methods
• Hard to define due to the definitions of the
  measures and the changing nature of the measure

• Several studies have defined validity between
  subjective sleepiness and sleepiness-impaired
  performance
• Purposeful interaction with the environment not
  possible when EEG dominated by alpha/theta and
  slow eye movements

28
Event Related Potentials (ERPs)

• ERP Transient series of voltage oscillations
• Recorded from the scalp
• Response to direct stimuli and responses
• Often defined in terms of polarity and minimum
  latency with respect to discrete
  stimulus/response
• Found to reflect perceptual, cognitive, motor
  processes
• Useful to decompose processing requirements of
  complex tasks (Fabiani et al., 2000)

29
Procedure

• Experimental paradigm design
• Subject preparation
• Preparation of ERP data for analysis
• Component definition and pattern recognition
• Data analysis

30
Advantages/Disadvantages

• We possess understanding concerning functional
  significance of different ERP components
• Brain regions from which component generated are
  known
• ERPs can be obtained in absence of operator
  action/performance

• Motion artifacts
• Require discrete stimulus or response
• Substantial training required for recording,
  analysis, interpretation

31
Sample ERP Outputs
32
Training/Application Times

• Depends on whether you want to learn the basics
  (2 months) or become knowledgeable about the
  basis of ERP signals (advanced degree)

• Application times from 15min for a few electrodes
  to 45min for large electrode array

33
Reliability/Validity

• Validity of ERP components to specific cognitive
  constructs convincingly demonstrated

• Reliability established through extensive
  replications
• Split-half reliability high for P300 amplitude
  (.92) and latency (.83)
• Test-retest reliability over several days for
  P300 amplitude .83 and P300 latency .63

34
EEG/MEG/fMRI

• Neural activity generates currents outside the
  skull which can be monitored by their electrical
  and magnetic fields
• Electroencephalogram (EEG)
• Magnetoencephalogram (MEG)
• Magnetic Resonance Imaging (MRI)
• Functional MRI (fMRI)
• Provide a basis for examining the neural
  substrate of specific cognitive processes

35
(No Transcript)
36
Research Areas amenable to MEG fMRI

• Movement-related brain activation
• Memory processes (encoding retrieval)
• Visual perception, attention, selection
• Auditory perception, attention selection
• Language production and processing
• Perception of music
• Learning and brain plasticity with respect to
  cognitive functions

• Uses
• Neurosurgery
• Localization of epileptic foci based on specific
  brain areas spiking
• Estimation of the impact of certain lesions on
  higher neural functioning
• MEG preferred for temporal resolution fMRI for
  spatial resolution and maximum information
  when/where the two methods can be combined.
• Limited to patients without ferromagnetic inserts

37
MEG/fMRI Mechanisms

• Requires use of superconducting quantum
  interference devices (SQUIDs)
• Exploit quantum mechanical Josephson-effect
• Modern MEG systems monitor signals from 150-300
  SQUIDs spread equally over the head surface
• Variants of MEG sensors known as Gradiometers

• Spin tilt of protons aligned with strong magnetic
  field is pertubated by a brief electromagnetic
  pulse
• Protons emit burst of RF energy as they return to
  their initial aligned state
• Strength of signal with particular RF signature
  allows determination of proton density
• Helps define tissue characteristics at that
  location

38
Procedure (MEG)

• Helmet like gantry placed over subjects head
• Coils fixed on head provide weak magnetic sources
  known as anatomical sites.

39
fMRI

• Subject reclined on movable gantry, shifted into
  the bore of a magnet.
• Structural scan
• Repeated functional scans

40
Experimental Setup

• During session, sequence of visual or auditory or
  other sensory stimuli is presented to the subject
  who has to process them according to a predefined
  task.
• Often the same or similar stimuli are presented
  with slightly different task requirements
• Differences between the BOLD responses in the
  different experimental conditions are evaluated
  to determine what brain regions are specifically
  activated by a particular task or sensory input

41
MEGData Analysis

• Event-locked epochs initially averaged separately
  for each subject, channel (sensor), task, or
  condition
• Epochs containing artifacts are rejected or
  correlated
• Grand average waveforms are scanned for
  components (peaks/troughs)(50-100msec)
• Time/amplitude measurements determined for each
  component, task, subject.

• Statistical analysis to ascertain significant
  differences which are localized to a region of
  the brain
• Various co registration techniques possible using
  nonlinear parameters, algorithms from chaos
  theory
• Analyzing continuous MEG data is based on chaos
  theory and beyond the scope of this class

42
fMRI Data Analysis

• Continuously recorded fMRI/BOLD data sorted
  according to different stimulus/task condition
• Any distortions are corrected by a processing
  algorithm.
• Those functional images showing significant
  differences in activation levels at the level of
  the individual voxels

• fMRI permits single subject statistical analysis
  due to better signal to noise ratios
• Comprehensive evaluation packages are available
• Statistical parametric mapping (SPM) package
• Also advanced applications that deconvolve
  temporally overlapping BOLD responses

43
Sample MEG Data
44
Sample fMRI Data
45
Training

• Predefined protocols performed by techs after two
  week training period
• For more advanced/sophisticated applications, at
  least one full time engineer or physicist should
  be available (6 month training).
• In clinical setting, support of a dedicated
  physician is required for data interpretation
• Neuroscientists should have a sound background in
  experimental design as well as neurophysiological
  education. Specific training per investigations
  being conducted
• Joint program between Emory and Georgia Tech
  provides degree and background in medical physics

46
Test Times

• Experimental run may require about 1.5 hours with
  15 minutes additional subject preparation

47
MEG Reliability/Validity

• MEG
• Artifacts can be minimized but not eliminated
• Raw signals, an experienced rater can recognize
  these distortions
• Potential artifacts depend on s/n
• Avoid misinterpretation of waveform
• Reasonable solutions possible but not foolproof.
• Reliability and validity rely on users experience

48
fMRI Reliability/Validity

• fMRI
• Continuous maintenance should result in
  acceptable raw images
• Artifacts due to discontinuities in magnetic
  susceptibility can lead to low structural
  distortions and signal loss
• Easily detected but not readily corrected
• Errors can be introduced to statistical analysis
• Even with formal fMRI handling correct, erroneous
  conclusions may be drown from data obtained using
  inappropriate experimental designs
• Reliability and validity of fMRI method also
  relies on user experience

49
Blood Pressure Measures

• Strain affects mood and behavior
• Performance and physiological effort have used
  strain analyses that are only psychological
• Progress in ambulatory measurement has allowed
  assessment of behavioral, emotional and
  activational interaction with workload under real
  work conditions.
• Low strain load has been used to investigate long
  term work (fatigue, boredom, vigilance) on health

50
Ambulatory BP

• Developed clinically to measure physical work
  effects
• Increased use for psychosocial work
  characteristics
• Use of portable recorders for non-invasive
  recording

51
Procedure

• Procure ABPM Device
• Select work analysis objective methods
• Develop prequestionnaire about normal activities
• Maintain diary
• Prepare BP monitor
• Fit monitor to subject
• Instruct subject
• Subject resumes daily routine
• Remove BP monitor, collect diaries
• Transfer data

52
Requirements

• Must have information about body position and
  motor activity at time of measurement
• Also should combine measurement with
  psychological data
• Must assess the nature of strain
• Include measures of perceived mental load,
  perceived control, mood, motivation

53
Advantages/Disadvantages

• Can simultaneously record workload and subjects
  strain experience behavior
• Work-strain-related effects on BP recovery can be
  investigated and assessed

• ABPM can have an artifact effect on daily
  activity (Hawthorn type effect)
• May have an effect on subjects sleep

54
Example of Recorded Data
55
Training/Application

• Investigator experienced in principles of BP
  measurement and interpretation of readings
• Main and artifact variables on BP

• Fitting takes 15-20 minutes
• Subject instruction about inflation/deflation of
  cuff
• Instruct subject on need for written diary
• Removal done by investigator with follow up

56
Eye Blink Measures

• Issues related to alertness/drowsiness
• Neurobiological products of interaction between
  endogenous circadian pacemaker and homeostatic
  need for sleep
• Objective biobehavioral signs often require
  intrusive physiological monitoring
• Growing use of these measures as humans subjected
  to 24 hour workdays
• Technological improvements are making measurement
  more affordable and less obtrusive
• PERCLOS

57
Procedure

• Use infrared, retinal reflectance monitor
• Uses CCD camera to record eye closure
  measurements in real time
• General use of two cameras situated at 90 degree
  angle
• 850-nm filtered bright eye camera and a 950-nm
  filtered dark-eye (dark pupil) image
• Calculate the changes in brightness of pupil
  based on average brightness

58
Sample Pupil Imaging
59
Advantages/Disadvantages

• Availability of on-line, near-real-time,
  automated slow eyelid PERCLOS system unobtrusive
  to user
• Ideal if used with preset thresholds versus
  self-report
• Can be used as an investigative and applied tool.

• May not work in all situations (requires
  restricted FOV)
• May create artifacts in completion of task
• Equipment may be too obtrusive in mobile
  real-world applied environment
• Not ideal in low humidity environments (not able
  to differentiate moistening of eyes and fatigue
  based closure
• Misuse in safety sensitive environment may
  generate risk.

60
Training/Application

• No training required other than to teach operator
  to interpret feedback indicating drowsiness

• Small, fairly easily applied and useful in many
  but not all environments

61
Reliability/Validity

• Loss of alertness, drowsiness, and hypovigilance
  must be theoretically linked to performance
  deficits

• Two levels of validation
• Biobehavioral parameters
• Specificity of biobehavioral measure used
• Both reliability and validity must be established
  across dynamic range of performance

62
Human Respiration

• Respiration linked to a variety of functional
  psychological dimensions
• Response requirements
• Appraisal patterns
• Mental effort investment
• Various dimensions of emotion
• Affect
• Mood

63
Respiratory Measures

• Assessment of how depth and frequency of
  breathing contributes to ventilation
• Expressed as tidal volume
• Frequency is respiration rate (BPM)
• Measurement of parameters associated with gas
  exchange
• Breathing cycle

64
(No Transcript)
65
Applications

• Can be used with verbal self report related to
  work
• Task demands
• System demands
• Operator workload
• Stressful/hazardous aspects of environment

• Must remain cautious of respiratory changes not
  related to work environment
• May be secondary or compensatory to respiratory
  volume changes

66
Measurement/Procedure

• Measure non-obtrusively motions of the rib cage
  and abdomen using an inductive respiratory
  plethysmography device
• Calibration techniques
• Often combined with other measures
  (accelerometry, ECG, oximetry, PetCO2

• Research Design
• Prepare Subject
• Physiological Monitoring
• Data Acquisition/Analysis

67
Equipment
68
Advantages/Disadvantages

• Valuable in applied studies of complex
  tasks/systems demands/effort investment
• Combines easily performance-based and subjective
  methods
• Demonstrates metabolic activity associated with
  task but may be affected by extraneous variables

• Respiration is intricate interplay between
  brainstem, metabolic, volitional influences
• Difficult to unravel
• May not be a convenient measure when interested
  in monitoring oxygen consumption
• Will the quantification of respiration answer the
  questions which researcher is interested in?

69
Training/Application

• Considerable investment in time, effort,
  resources to familiarize oneself with the
  underlying physiology, measurement, analysis
  required
• Basics in a few weeks, expert knowledge much more
  time

• Application of sensors, calibration, signal
  quality verification varies from 10-30min.

70
Reliability/Validity

• Problems with reliability related to posture
  changes, movement and respiration related
  movement
• Possible to filter some of these out but probably
  too complex for automatic filtering

• Measurement of PetCO2 may pose serious validity
  problems that need to be considered
• Breathing can vary widely and it may be difficult
  to correlate these changes to changes in work
  requirements