Lie Detection Systems: Part 1

1. Introduction

Techniques for lie detection have existed for decades through the use of interviews, interrogations, and other means based on little scientific merit.  Today, modern lie detection techniques rely on measurable physiological responses, which serve as indicators of deception.  This White Paper explores the concept of veracity by summarizing the following physiological techniques used for either overt or covert lie detection scenarios:

  • Polygraph
  • Voice Stress Analyzers
  • Facial Thermography
  • Functional Brain Imaging

2. Polygraph

Considered one of the best know and widely utilized lie detection techniques in the U.S. and other countries like Israel, Japan, and Canada, the polygraph provides U.S. law enforcement and intelligence agencies with a tool that combines interrogation with physiological measurements obtained during the polygraph examine.  By recording a person’s respiration, heart rate, blood pressure, and electrical conductance at the surface of the skin, the polygraph examination applies to overt scenarios where a subject is asked a series of yes/no questions while wired sensors relay data about the person’s physiological attributes.  In addition to these traditional measurements of involuntary and somatic activity, other physiological events can be recorded non-invasively, including cardiac output, total peripheral resistance, skin temperature, and vascular perfusion in coetaneous tissue beds.

Trained polygraph practitioners emphasize that the polygraph instrument itself measures levels of deception indirectly, by measuring physiological responses that are believed to be stronger during acts of deception than at other times.  Collectively, these patterns of physiological responses to relevant questions asked by an investigator are recorded on an analog or digital chart and require human interpretation from the polygraph examiner.  Aside from interpretation of the polygraph chart, which can also be somewhat automated by computer algorithms, other factors that assist or inhibit the polygraph instrument’s ability to accurately perform a lie detection function include the potential for influence from drugs or alcohol, examiner’s expectations about the examinee’s truthfulness, and adverse physiological responses that have no direct correlation to the examinee’s intent to deceive.

Aside from its use as a diagnostic tool to test for deception where truth or deception decisions are made based on charts that are analyzed and scored, the polygraph has other practical applications such as:

  • Eliciting admissions from people who believe or are influenced to believe that the polygraph machine will accurately detect their attempts at deception.
  • Testing the level of cooperation with an investigative effort through suspicion or detection of countermeasures used by the examinee during polygraph testing.

Overall, polygraph examines are considered to be an effective tool for lie detection when combined with information from other sources used to judge truthfulness or deception (i.e., pretest interviews, comparison question testing, observation of examinee’s demeanor, etc.).

2.1 Commercially Available Polygraphs

Digital quality, commercially available polygraph systems consist of either a complete hardware/software integrated system on a laptop PC or a stand-alone data acquisition system that can be connected to an existing computer.  Some systems include scoring algorithm software and/or peripheral hardware used for motion sensing or to measure additional physiological parameters.  The following list describes a few of the commercially available polygraph products:

  • LX4000 – manufactured by Lafayette Instrument.  Records, stores, and analyzes physiological characteristics derived from respiration, galvanic skin response, and blood volume/pulse rate.
  • 4-6 Channel S/Box Package – from Axciton Systems, Inc. provides a customized polygraph system designed to accommodate 4, 5, or 6, channel physiological parameters.
  • Computerized Polygraph System (CPS) – manufactured by Stoelting Polygraphs.  Claims to be the only computerized polygraph system containing a scoring methodology based on verified criminal data from a major government law enforcement agency.

2.2 Conclusions

As the subject of hundreds of controlled, scientific studies regarding polygraph effectiveness, a final and concrete determination of the polygraph’s accurateness still hinges on research information contained in classified national security documents as well as proprietary information about computer scoring algorithms or other trade secrets that equipment vendors will not divulge.  The American Polygraph Institute cites a 70% accuracy rating among polygraph skeptics with a 90% accuracy rate among proponents.  A 2003 report conducted by the National Academy of Sciences, which examined 57 previous polygraph studies to quantify the accuracy of polygraph testing within the scope of personnel security screening concluded, “The inherent ambiguity of the physiological measures used in the polygraph suggests that further investments in improving polygraph technique and interpretation will bring only modest improvements in accuracy.”  This study also pointed out that polygraph countermeasures deployed by major security threats could seriously undercut the value of polygraph security screening.  Nonetheless, this same study concluded that the polygraph technique is the best tool currently available to detect deception and assess credibility.

3. Voice Stress Analyzers

Touted as a lower cost, less invasive lie detection method, commercially available voice stress analyzers (VSAs) have been in use since the early 1970’s through efforts between private industry and the U.S. Army.  Based on the presumption that liars experience more stress than truth-tellers, a VSA works by measuring microtremors associated with laryngeal muscles used during voiced excitation.  The microtremor are defined as inaudible vibrations that speed up uncontrollably in the human voice during an act of deception.  The level of microtremor maintains an inverse relationship to a person’s stress level where more stress denotes less tremor.  Slow microtremors occur at rates between     3-5 Hz while more rapid tremors can occur at 6-12 Hz.  Microtremors can be affected by numerous variables, including age, stress, drugs, alcohol, medical illness, brain disorders, and multiple sclerosis.  Major issues surrounding VSA validity and accuracy remain focused on how stress impacts the laryngeal muscles during normal speech production and whether VSA speech processing algorithms can effectively extract and quantify the existence of microtremor information.  Proponents of VSAs point out several benefits of using their equipment in lieu of more traditional polygraph techniques, namely:

  • Applicability to covert scenarios.
  • Less training time required to learn and operate.
  • No academic prerequisites for training.
  • 30-50% less time to administer the testing regiment.
  • Voice recordings can be processed as well as live speech.
  • Lower cost of ownership.

3.1 Commercially Available VSAs

Commercially available VSAs use some form of speech signal processing to extract excitation information related to microtremors.  The following VSAs provide a sample of these commercially available products:

  • Psychological Stress Evaluator (PSE) – patented in the 1970’s by Allan D. Bell and marketed through Dektor Counterintelligence and Security, Inc.
  • Truster – developed by an Israeli company named Makh-Shevet.
  • Computerized Voice Stress Analyzer (CVSA) – developed in the late 1980s by the National Institute for Truth Verification (NITV), which claims their system is in use by more than 500 law enforcement agencies.
  • Lantern – developed by Diogenes Group, Inc.
  • Vericator (formerly known as Truster Pro) – manufactured by Trustech Ltd. Integritek Systems, Inc.
  • VSA Mark 1000 – manufactured by CCS International, Inc. and marketed as a covert electronic lie detection system.

3.2 Conclusions

One technical report conducted in 1999 by ACS Defense, Inc. and the U.S. Air Force Research Laboratory in Rome, N.Y., concedes that information from previous studies of speech under stress combined with their own Air Force evaluations and experiments using commercial VSAs suggests that a speaker’s voice characteristics change when the speaker is under stress.  However, as stated previously in other studies, a variety of factors in addition to stress can reflect changes in the human speech production process, including the presence or absence of microtremors.  In its final conclusions, the Air Force study determined that the level and degree to which changes in muscle control associated with speech production impart more or less fluctuation in the speech signal cannot be conclusively determined.  In other words, focusing on the absence or presence of microtremors alone does not conclusively define the accuracy of VSAs.  Furthermore, the study recommends that several speech features may be needed to accurately capture the subtle differences in how speaker’s convey their stress in various speech scenarios.

Another study conducted in 2000 by the Department of Defense Polygraph Institute (DoDPI) and the U.S. Army Walter Reed Hospital, also concluded that the relationship between microtremors and a speaker’s deception might not be experimentally sound and that the use of microtremor analysis to detect deception is nothing better than chance.  In addition, a 2002 study, conducted by the DoDPI research division staff to investigate the NITV’s CVSA, provided no evidence to support the CVSA for its ability to identify stress-related changes in voice.  Lastly, a 2002 VSA literature review conducted by the National Research Council revealed that VSA accuracy rates from commercially available systems remain at or below chance probability levels.  Still, despite the doubt from many researchers and published reports citing a lack of scientific evidence to support industry claims, the commercially available CVSA system, which retails for about $10,000, claims an accuracy rate of 98%.

Leave a Reply

Your email address will not be published. Required fields are marked *