Saturday, April 25, 2009

Lie To Me


Background

Relevant Parts of the Brain

Anterior Cingulate Cortex
:
  • Connected with the prefrontal and parietal cortex
  • Regulates autonomic function including blood pressure and heart rate
  • Also involved in rational cognitive functions such as reward-anticipation, decision making, empathy and emotion

Dorsal Lateral Prefrontal Cortex:
  • Responsible for motor planning, organization and regulation
  • Associated with Working-memory; the storage and manipulation of information


The Parietal Cortex:
  • Determination of spatial sense and navigation


The Polygraph

  • Inadmissible in court
  • Federal Polygraph Handbook:
  • Question Form - Be clear and concise; Avoid legal terms when possible; Be constructed so that they may be answered yes or no; Not be worded in the form of an accusation or contain an inference that presupposes knowledge or guilt; In specific issue tests when testing for multiple items or amounts of money, use the phrase, "any of', e.g., "Did you steal any of that money?"; In specific issue tests, only address one issue in each question; In specific issue tests, only address one incident in each series.
  • Question Type - Primary Relevant. This question tests the possible direct involvement of the examinee. In PDD screening questioning formats, all relevant questions are considered primary relevant questions;
  • Secondary Relevant. This question tests the examinee's possible involvement in the
    offense under investigation. A secondary relevant question should be constructed to address a secondary issue such as help, plan, or participate; test for secondary involvement in, such as seeing, hearing, or knowing; or focus on the nature or location of evidence and/or physical acts that support the primary offense. There are three types of secondary relevant questions:
  • 1. Evidence-Connecting. An evidence-connecting question is designed to
    determine if the examinee was involved with any of the evidence of the crime or is aware of the nature or location of various items of evidence.

    Do you know where any of that money is now?
  • 2. Guilty-Knowledge. A guilty-knowledge relevant question is used to determine
    if the examinee has any knowledge of who committed the incident under investigation.
Do you know who stole any of that money?
  • 3. Secondary-Involvement. A secondary-involvement relevant question tests for
    secondary involvement such as seeing or hearing or focuses on physical acts that support the primary offense.
Did you participate in the theft of any of that money?
  • Comparison Question. Physiological responses of comparison questions are
    compared to physiological responses of relevant questions. The comparison question is designed to produce a greater physiological response for the non-deceptive person. The probable lie and directed lie questions are the two types of comparison questions used within the Federal Government.
  • 1. Probable-Lie Comparison Question. This question is designed to be a probable lie
    for the examinee. The PLC question should be similar in nature but unrelated by time, place or category to the specific issue. However, in screening examinations the PLC can be related to the issue(s) as long as the screening comparison question establishes a dichotomy between the relevant and comparison issues. A comparison question should be broad in scope and time so that it captures as many of the examinee's past life experiences as possible. The physiological responses to the PLC are compared to the responses of the designated relevant questions. The exclusionary and screening comparison questions are the two types of PLC questions used within the federal government.
Theft issue: Before 1997, did you ever steal anything?
  • 2. Exclusionary Comparison Question (ECQ). A probable-lie question
    should be similar in nature but unrelated to the issue being tested. The question should be
    separated from the relevant issue by time, place or category. The comparison question should use the same action verb or similar in nature action verb as that of the relevant issue. A comparison question should be broad in scope and time so that it captures as many of the examinee's past life experiences as possible.
Before this year, did you ever steal anything from someone who trusted you?
  • Directed-Lie Comparison. The DLC question is a specialized comparison
    question. A properly constructed DLC question involves a minor transgression which should have some personal significance to the examinee. Upon acknowledging having committed such a transgression, the examinee is directed to lie when asked that question on the test. The question is separated from the relevant issue by category.
Did you ever commit a minor traffic violation?
  • Sacrifice Relevant Question. When used, this is the first question that refers to the
    relevant issue, and it prepares the examinee for the introduction of the relevant questions.
    Sacrifice relevant questions are not scored during the test data analysis phase of a PDD
    examination.
Regarding the theft of that car, do you intend to answer each question truthfully?
  • Symptomatic Question. This question is designed to test for an outside issue that
    could be more significant for an examinee than the relevant and comparison issues. Symptomatic question responses are evaluated, though not numerically scored, during the test data analysis phase of a PDD examination.
Is there something else you are afraid I will ask you a question about?
  • Stimulus Questions: designed to measure capacity for response. These can be relevant or irrelevant:
Does 10+ 9= 19?
Were you on that bridge that night?

  • What do the question formats mean?
  • The yes/no format means the linguistic construction is always in the hand of the examiner, not the examinee.
  • The built in shifts from relevant to irrelevant to comparison and control questions indicate an acknowledgment of the difficulty of maintaining accuracy and control in the testing environment.
  • Problems with the Polygraph:
  • For instance, using the Directed-Lie Comparison question as an example of how the subject lies poses several real problems since Polygraph measures physiological response and not neurological response. The form of a directed lie question dictates that the subject is instructed to lie. The point of polygraph is to discern a lie based on a spike in stress levels resulting from the concealment of a lie. The Comparison question removes the element of concealment and therefore removes exactly what the examiner wishes to observe: the effects of concealment.
  • The problems with Polygraph tests are well known to most people today, they measure stress without having a way of differentiating between the stress caused by crafting a lie and the stress caused by being subjected to a polygraph.
  • Langleben: "Its specificity is limited because it relies on the correlates of peripheral nervous system activity, while deception in a cognition event with top-down control by the central nervous system"
  • This is the shift his research is trying to make: from physiology to cognition.
The Next Generation of Lie Detection

The Experiment
Subjects: 23 right-handed male undergraduate students (mean age: 19.36)
Task: A modification of the Guilty Knowledge Task (GKT)

  • Instruction Protocol:
  • Participants were given an envelope with two playing cards, a 5 of clubs and a 7 of spades, along with $20.
  • The investigator instructed the participants to deny possession of one of the cards and acknowledge possession of the other during the imaging phase.
  • Participants were told they could keep the money only if they successfully concealed the identity of the lie card during the scan session.
  • A third party then led them to the scanner and instructed them to answer each question as truthfully as possible.
  • A sequence of photographed playing cards was presented
  • The series of cards included five stimulus classes:
  • 1. Lie (5 of clubs or 7 of spades)
  • 2. Truth (5 of clubs or 7 of spades)
  • 3. Recurrent distracter (2 of hearts)
  • 4. Variant distracter (remaining cards, 2- 10 all suits)
  • 5. Null (back of card)
  • Cards were presented for 2 seconds followed by a variable (0 -16 s) during which the null card was shown.
  • 24 Lie, Truth and recurrent distracter cards and 168 variant distracter cards were shown.
  • Participants were instructed to press a left button (yes) to confirm that a card was in their possession or the right button (no) to deny it.
  • In total, 432 stimuli were presented with a session length of 14.4 minutes.
The experiment still operates with yes/no responses, like the polygraph.
  • Results:
  • No difference between those who reported using a strategy and those who reported using no strategy.
  • "The key point is that you need to exercise a system that is in charge of regulating and controlling your behavior when you lie more than when you just say the truth," Langleben said. "Three areas of the brain generally become more active during deception: the anterior cingulated cortex, the dorsal lateral prefrontal cortex and the parietal cortex."
"Complications arise when we examine Langleben group’s attempts at replication when using this same paradigm. In their 2005 papers (Davatzikos et al., 2005; Langleben et al., 2005), where the same data appear to be reported twice (analysed in two different ways), we find that:

1. The initial ACC and inferior parietal lobe findings are not replicated and the frontal activations described are of insufficient statistical significance to be reported by
most neuroimaging groups (Z values of less than 3);

2.The parietal lobes (bilaterally) now exhibit greater activation during truthful
responding than deceptive responding (c.f. Langleben et al., 2002); and

3.The behavioural and functional anatomical findings are reported differently across the 2005 papers: while truthful RTs appear only qualitatively longer than lie responses in Langleben et al. (2005), they are reported as significantly longer in Davazitkos et al. (2005); while very many regions exhibit greater activation during truthful responding in Langleben et al. (2005), a different subset of foci are maximally informative in Davazitkos et al. (2005; admittedly, following application of a different, and novel, analytic method)."

Conclusion
  • Admits that brain activity previously associated with deception was driven by the task design
  • Lie and Truth operate with a similar frontoparietal network, but lie still appears to be a more working-memory associated task
Work Cited

Langleben, et. al. Telling Truth From Lie in Individual Subjects With Fast Event-Related fMRI. University of Pennsylvania, 2005.

Federal Psycho-physiological Detection Of Deception Examiner Handbook. COUNTERINTELLIGENCE FIELD ACTIVITY TECHNICAL MANUAL
October 2,2006

Temple-Raston.
Neuroscientist Uses Brain Scan to See Lies Form. http://www.npr.org/templates/story/story.php?storyId=15744871. NPR, 2009.

Spence, Sean.
Playing Devil's Advocate: The Case against fMRI Lie Detection. Legal and Criminal Psychology (2008) 13, 11-25. The British Psychological Society, 2008.

No comments:

Post a Comment