ERP and fMRI Research in the Detection of Deception

1
ERP and fMRI Research in the Detection of
Deception
Jennifer M. C. Vendemia Michael Schillaci -
Robert Buzan - Eric Green - Scott Meek
2
Organization

• Theoretical Research with ERPs
• fMRI in 5 seconds
• Current Deception Research with fMRI
• Techniques combining multiple methods

3
Amplifier Buzz
Scalp, Skull
High Impendence
Random Brain Activity
Muscle
4
Aspects of deception in the ERP
P3a An early attention related component with an
anterior distribution and positive deflection.
Occurs when one switches tasks such as from
telling the truth to telling a lie.
P3b A late component that is related to
decision making, workload, inhibition, and
attention, and context updating.
N400 A component that occurs when what weve
heard, said, or seen does not match the contents
of our semantic (and possibly) episodic memory.
Anterior distribution, negative deflection
0 250 500 750ms
5
ERPs The Signal-Noise Approach

• Signal Brain wave associated with deception
• Several potential waveforms are studied
• The most consistent is the P300 evoked in an
  oddball paradigm with concealed information
• Noise Every other signal generated by the human
  brain

6
The Signal-Noise Approach has Historically Proven
Inadequate for BOTH Polygraph and ERPs

• Reliability Both measures deliver consistent
  results across repeated tests
• Validity But NEITHER measure has been
  experimentally validated
• They measure something, but not necessarily
  deception

The Signal-Noise Approach has no History with fMRI

• Reliability Few consistent findings
• Validity No validity with reliability

7
Functional MRI

• This technique allows us to watch the human brain
  in action.

8
Functional MRI

• An MRI scanner can detect the magnetic change as
  blood flow increases in certain parts of the
  brain.
• We can use this to determine which parts of the
  brain are most active.

9
Functional MRI

• Example ask a person to move their eyes

M1 movement
S1 sensation
10
Activated Parts of the Brain
11
Current Research in fMRI Regions of Activation
sited in Bhaat et al (in press)
Area 9, 10
VLPFC
Area 32
Area 8
Sensory Motor Strip
Area 21
Area 17
Caudate
Cerebellum
Additional regions Hippocampal gyrus, left
inferior parietal
12
Individual Trials fMRI studies
13
Variability in fMRI Approaches

• Within Subject Noise
• Subject movement
• Respiratory, cardiac artifacts
• Scanner instability
• Attentional modulation
• Inconsistent cognitive strategy
• Learning effects
• Drugs and medications
• Anxiety
• Countermeasures
• Between Subject Noise
• Consistent differences in factors related to
  within subject noise
• Anatomic variability
• Cytoarchitectonic variability
• Variability in venous drainage patterns
• Differences in hemoglobin concentrations

• Between Paradigm Noise
• Inconsistent definition of the type of deception
• DIfferences in the rate, number, and type of
  stimuli presented
• Differences in the type of memory to which the
  participants deceive
• Differences in reward/punishment scenario

14
(No Transcript)
15
Incidental Measurement Differences
16
Example paradigm differences

• Spence, Farrow, Herford, Wilkinson, Zheng, and
  Woodruff (2001) directed lies to episodic memory
• Langleben, Schroeder, Maldjian, Gur, McDonald,
  Ragland, OBrien, and Childress (2002) directed
  lies in a digit recall type task with cards
• Lee, Liu, Tan, Chan, Mahankali, Feng, Hou, Fox,
  and Gao (2002) Feigned memory impairment to
  digit span and autobiographical memories
• Ganis, Kosslyn, Stose, Thompson, and
  Yurgelun-Todd (2003) Planned lies vs.
  spontaneous lies to long latency episodic
  information
• Kozel, Padgett, and George (2004) Planned lies
  to recently short latency episodic information
• Faro, Mohamed, Gordon, Platek, Williams, and
  Ahmad (2004) Planned lies to short latency
  episodic information

17
Deception is a complicated socio-dynamic
cognitive process

• It is possible that deception is not the result
  of a unique structure or system within the human
  brain
• Rather it is the result of several sub-processes
  that are also recruited during other
  socio-dynamic cognitive processes (like lecturing)

18
Conclusions from MRI Studies

• Motivation Kozel, Langleben, Phan
• Orbitofrontal activation only present in Kozel
• Autobiographical Memory Ganis, Lee, Spence
• Temporal activation present only in LEE
• Weighing of multiple information sourcesall
  studies
• Prefrontal cortex Lee, Ganis, Kozel, Faro
• Resource allocation, attention switching,
  response conflict all studies
• Lee, Langleben, Ganis, Kozel, Faro, Spence, Phan
• Regions of confusion
• Cuneus, cerebellum

19
Conclusions

• These technologies are not ready for practical
  application
• The issues that limit the utility of ERP and fMRI
  have nothing to do with the equipment
• The major problems all revolve around the
  supporting science
• The science is currently in its infancy, and has
  thus far had a troubled development