Counterintelligence and Neuroscience: The Merits and Limitations of the Polygraph, fMRI, and EEG
By: Sarah Mesich
Author and lecturer Terrence McKenna speculated that language was developed in pursuit of deception – in other words, that our most powerful tool of communication was principally aimed at lying. The introduction of the lie fundamentally alters human interaction. The primordial deception in the garden, when man discovered his nakedness and decided to hide himself from God, is an early account of that distinctly human tendency to deflect the gaze of another. The Genesis story also tells us of the inevitable consequences of introducing the lie – the severing of relationship and access.
Counterintelligence, as an effort to guard against and root out subversive elements targeting one’s nation, is principally concerned with methods of lie detection and best practices of employment. This paper analyzes the well-worn counterintelligence tool – the polygraph – as well as two emerging technologies: functional magnetic resonance imaging (or fMRI) and electroencephalography (EEG). Accordingly, the following analysis is broken down into four sections. Sections I and II focus on the polygraph. The first section gives an overview of the test, its methods, and delicacies of administration. The second section enumerates the polygraph’s shortcomings and how it might best be utilized for CI purposes. Sections III and IV focus on more recently emerging technologies. The third section provides general descriptions of these tests and the final section describes how these tools can be used to supplement the polygraph and bolster CI security measures. While these technologies do have limitations, used in conjunction they are valuable tools for the preservation of national security.
I. The Polygraph
The polygraph is the product of a long succession of developers. After William Moulton Marston’s development of the systolic blood pressure test, other tests were created to measure breathing patterns, electrodermal activity, and pulse. In 1921, John Larson was the first to combine multiple physiological measurements into a single test (Krapohl and Shaw 17). Lie detection tests involve presenting the examinee with a series of stimuli in the form of questions. Multiple physiological responses are then recorded and correlated to the subject’s response to each question. Overlaying measurements increases the validity of the test and allows examiners to make more reliable predictions. Modern polygraphs measure blood pressure, changes in breathing, and sweating.
Test Question Construction
The trick of effective polygraph testing lies largely in mastering Test Question Construction (TQC). Cognitive responses to the novelty, intensity, and salience of the stimuli can evoke physiological reactions. Polygraph examiners must control for novelty and intensity in the construction of their test questions. The pretest interview also aids in controlling for novelty (Krapohl and Shaw 61). Questions must be asked in exactly the same manner as in the pretest interview. Otherwise the examinee may register surprise responses that will interfere with the detection of deception. Additionally, the examiner must maintain the same volume and demeanor for each question to control for intensity (62). If salience can successfully be isolated as the source of physiological responses, these responses can then be used to signal personal significance and can be linked to deceptiveness.
Not all salient responses indicate deceptiveness, however. An additional difficulty in crafting good questions is avoiding emotionally charged language. Questions involving the examinee’s name, social taboos, or graphic descriptions (i.e. slaughter, massacre, etc.) can all generate responses of salience that are not linked to deceptiveness. In sum, there are many ways in which a polygraph can produce false positives if the examiner does not carefully control the questions.
There are two mainstream techniques in polygraph administration and analysis. The Peak of Tension (POT) technique and the Concealed Information Test (CIT).
“Peak of Tension” was coined by Leonarde Keeler in the 1930’s. It was later developed into two subcategories: Type A and Type B. Type A tests are used in cases in which the details of the crime are known, whereas Type B is used in cases when the details are still unknown. The general idea of POT tests is that, since the examinee knows the order of questions from the pretest, they will anticipate the anxiety-inducing question(s) and stress symptoms will build until the question(s) is asked and will decrease – as a sign of relief – afterwards. Thus, the polygraph data would display a “peak” of tension and indicate at what point the examinee was most uncomfortable (Krapohl and Shaw 167-168).
The Concealed Information Test, previously called the Guilty Knowledge Test (GKT), was pioneered by David Lykken in the late 1950’s. The CIT rests on the assumption that a guilty examinee will react more strongly to the “key” (or the critical detail) than to the buffers or padding (the random details). An innocent examinee, on the other hand, should react randomly to all details (Krapohl and Shaw 173). The CIT is much more thoroughly researched and is generally the preferred technique in polygraph administration today. The National Research Council conducted a 2003 report which found the CIT to have an 88% accuracy rating (174).
Despite historical improvements in technique and equipment, polygraphs still suffer from various inaccuracies and limitations. The following section discusses potential failings of the polygraph and best practices in their employment for counterintelligence purposes.
II. The Polygraph in Counterintelligence
In their article In Search of the Magic Lasso: The Truth About the Polygraph, Stephen Fienberg and Paul Stern argue that “[t]he validity of polygraph testing depends in part on the purpose for which it is used” (251). The article reviews the report of the National Research Council after the Wen Ho Lee debacle. In the report, the NRC recommended that the Department of Energy, as well as other intelligence agencies, reduce their reliance on screening tools such as the polygraph in favor of other more effective security measures.
The Problem of Validity
I described how the polygraph isolates for salience. The salience factor must be carefully controlled in order to be justifiably linked to deceptiveness. This is not always successfully carried out, however. The main problem with the polygraph is that its question-form stimuli often leave room for ambiguity and thus, are subject to everyone’s personal range of comfortability with dishonesty.
For example, when confronted with the question “Have you had any contact with a foreign national in the last 6 months?” one examinee may rapidly perspire and experience an immediate increase in heart rate as they explain that yes, they called their mother in Canada just that morning. Another examinee may answer no, remembering the few emails back and forth with a Chinese university about a call for papers, but not be particularly disturbed by it because they never met in person and the exchange was quite brief. The second examinee may not demonstrate any abnormal fluctuations in heart rate, perspiration, blood pressure, or breathing. This example serves to reemphasize the fact that polygraphs indicate salience (i.e. personal significance or value) and not necessarily concealment of information that investigators find relevant.
The National Research Council Report
In the NRC report, the committee suggested that the use of the polygraph in employment screenings may not be an effective counterintelligence tool. The reason for this is that the more “event-specific” the polygraph is, the more concrete and less ambiguous the questions can be (Fienberg and Stern 251). Regular screenings, however, are more likely to be routine questions, less specific and therefore yield less valid results.
The NRC report included calculations of 10,000 hypothetical examinees who were administered a polygraph as a part of procedural screenings. Ten of the examinees were designated as spies and 9,990 were innocents. Assuming beyond-the-best-case administration, polygraph accuracy was calculated in two scenarios. In scenario 1 – the “suspicious” mode – the test was strict enough to catch 8 of the 10 spies. However, this scenario also falsely implicated 1,598 innocent examinees (Fienberg and Stern 253). Thus, using the polygraph in an aggressive and mole-focused manner would still require CI professionals to investigate and eliminate nearly 1,600 innocent suspects before honing in on the true suspects. In scenario 2 – the “friendly” mode – the test is designed to protect innocents. In this scenario, only 39 innocents are falsely under suspicion. However, scenario 2 implicates only 2 of the 10 spies, producing fewer red herrings but also fewer leads overall. “The committee concluded that for practical security screening applications, polygraph testing is not accurate enough to rely on for detecting deception” (254).
The Double Bind of Deterrence and Skepticism
This is not to say that polygraphs should be thrown out due to their lack of scientific validity. The issue is much more nuanced than that. One of the most valuable aspects of the polygraph is their deterrent force. The perception of a test that can identify a lie or force the truth from unwilling subjects is a powerful tool in itself. Consider American spies like David Henry Barnett, who remained in a low-access position as a training officer due to his fear of taking a polygraph exam. Ana Montes is another example: choosing to stay with DIA instead of applying to CIA for fear of information coming out in extensive polygraph checks. Both cases demonstrate the deterrent capacity of the polygraph.
Preserving this deterrent capacity, however, presents counterintelligence professionals with a double bind. The effectiveness of the polygraph resides in the belief that deception will be revealed, but overconfidence in the test by CI investigators can lead to security oversights. Fienberg and Stern remark, “It may be harmless if television fails to discriminate between science and science fiction, but it is dangerous when government does not know the difference” (254). Effective counterintelligence policy may therefore continue to include polygraphs as a part of the screening application process while also seeking additional and more innovative ways of screening applicants and employees.
As previously stated, the efficacy of the polygraph was increased enormously when Larson began to combine measurements of multiple physiological responses. By identifying the convergence of multiple inputs, CI professionals can be more confident in their assessments. The following section looks to two emerging technologies as effective instruments for supplementing the limitations of the polygraph.
III. New Methods of Lie Detection
In two words, the fMRI and EEG are both “brain scans.” Each technology observes a different process to identify brain activity. Through this process of “brain mapping,” investigators are able to associate different regions of the brain with different cognitive states. These technologies are not yet as popular as the polygraph, but both fMRI and EEG have potential to provide massive breakthroughs in counterintelligence investigations.
Functional Magnetic Resonance Imaging (fMRI)
Since neurons require oxygen to function, the fMRI principally analyzes the flow of blood to brain regions. Areas of the brain with more relative blood flow indicate increased oxygen uptake and thus, increased activity (Shen 679-680). When oxygen binds with hemoglobin proteins in the blood, it produces oxyhemoglobin. When neurons activate, they use up the oxygen and release deoxyhemoglobin back into the blood stream. Because deoxyhemoglobin is magnetic, functional magnetic resonance imaging (fMRI) can map areas of relatively high deoxyhemoglobin. These regions indicate neural spaces that have recently used oxygen and therefore, demonstrate increased activity (687-688).
Like the polygraph, fMRI lie detection methods usually rely on the Concealed Information Test (CIT) model. Research and methods for fMRI administration and analysis are far from perfect. However, corroborating polygraph results with fMRI results may be an effective way of ruling out false positives and narrowing in on actual suspects.
Instead of measuring blood flow, electroencephalography (EEG) measures electrical pulses in the brain. The exam involves the placement of electrodes on the examinee’s scalp to record electrical activity. EEG exams also utilize the CIT model (Shen 684). Electrical activity is measured as “event-related potentials” (ERPs). An ERP is the electrophysiological response to specific stimulus, or “event.” One of the more notable ERPs in electroencephalography is the P300. J. Peter Rosenberg explains, “The P300 is a special ERP component that results whenever a meaningful piece of information is rarely presented among a random series of more frequently presented, nonmeaningful stimuli often of the same category as the meaningful stimulus” (64). The P300 will be discussed in more detail in the following section.
IV. The fMRI & EEG in Counterintelligence
These emerging technologies have two significant capacities for counterintelligence purposes: lie detection and memory detection. The fMRI can be used similar to the polygraph in a classic lie detection test. Both fMRI and EEG technology can be used for memory recognition tests.
Using a CIT model, the fMRI can measure neural activity in areas of the brain associated with lying. These fMRI methods use questions as stimuli and record relative blood flow as the examinee responds. This method encounters similar limitations to the polygraph. Namely, these areas of increased activity are only associated with lying but may not directly implicate dishonesty. Researchers refer to this as the “reverse inference fallacy.” It is possible that any number of cognitive processes (random thoughts, distractions, triggered memories) could activate regions of the brain that would interfere with the assessment of deception. However, if administered in conjunction with a standard polygraph, the fMRI’s lie detection methods can provide valuable corroborating results.
Another powerful capacity of the fMRI is in memory detection. This test typically involves visual stimuli (i.e. presenting the examinee with a picture of a face they may have seen before). Studies have been able to reliably identify when examinees are in subjective memory states – meaning that they believe they have seen the object or person before (Shen 683). In a study about facial recognition, “researchers could tell with great accuracy whether a subject remembered seeing a particular face” (683). This tool could be used to confirm whether a suspect has memory of known foreign operatives or even locations or files they shouldn’t have access to. Sensitive information used in exams such as this must be carefully selected to not unnecessarily spill secrets. If done masterfully, fMRI memory detection tests can be an invaluable instrument of CI investigation.
EEG memory recognition works in much the same way. “The logic is that the brain will react differently to a stimulus (such as a photo of a particular aspect of a crime scene) if that person recognizes the stimulus” (Shen 684). If a suspect denies knowledge in an interview setting but is later found to have memory of the item in question, they can be identified as a probable suspect. For example, if the technology had existed at the time, the British Foreign Office could have administered an EEG memory detection test to Elyesa “Cicero” Bazna. The Chief of Security could have showed him just the memo heading of the leaked diplomatic transcript and compared the strength of the P300 event-related potential to his reactions when looking at the morning newspaper, a randomly selected page of Crime and Punishment, and his own portrait. The relative strength of the ERP would indicate that Bazna had seen the memo before and implicate him as the spy in the British embassy. The charade he presented as an illiterate and simple man would shatter if investigators could prove he’d laid eyes on something he should have never seen.
In conclusion, while the polygraph may suffer from a lack of scientific validity, it nonetheless provides a valuable deterrent force to potential counterintelligence threats. The polygraph should continue to be used in CI screenings but should be supplemented with additional investigative technologies such as fMRI and EEG. The fMRI can be used as verification of polygraph findings, as well as a method of memory detection. EEG is also a valuable method of memory detection and should be used to aid counterintelligence professionals. Threats to information security will always exist, but with the assistance of emerging technology, investigators can leverage the subconscious unease individuals experience when they attempt to deflect the gaze of another.