B. Frontal Area

Stimulus Type: Target _____ Irrelevant......... Probe_____

Average brain responses recorded in response to three types of stimuli: (1) Targets are phrases that are made relevant to the subject by instructing him/her to press a particular button when they appear. Note the large P300/MERMER (memory and encoding related multifaceted electroencephalographic response), which indicates that these stimuli are relevant to the subject. (2) Irrelevants are phrases that are not relevant. The P300/MERMER is lacking. (3) Probes are FBI-relevant phrases that the subject gives no overt indication of recognizing. The large P300/MERMER indicates that these FBI-relevant stimuli are relevant for this subject. A subject unfamiliar with FBI training would not recognize these Probe stimuli as significant, and would not display a P300/MERMER in response to them. Thus, it can be determined that this subject is an FBI agent (defined in this study as an "Information Present" determination). (Scale: -10 to +10 micro volts x 2250 msec: 250 pre-stimulus to 2000 post-stimulus.)

Brain Fingerprinting System: Detection of FBI Agent Knowledge

"Information Absent" Brain Response

A. Parietal Area

B. Frontal Area

Stimulus Type: Target _____ Irrelevant......... Probe_____

This subject shows a P300/MERMER only in response to the Targets. In this subject's brain responses to the Probe stimuli, the P300/MERMER is lacking, indicating that these FBI-relevant stimuli are not relevant for this subject. Thus, it can be determined that this subject is not an FBI agent (defined in this study as an "Information Absent" determination). (Scale: -10 to +10 micro volts x 2250 msec: 250 pre-stimulus to 2000 post-stimulus.)

Summary

The detection of concealed information stored in the brain of suspects, witnesses, intelligence sources, and others is of central concern to all phases of law enforcement and intelligence operations. Brain Fingerprinting technology presents a new paradigm in the psychophysiological detection of concealed information. This new system detects information directly, on the basis of the electrophysiological manifestations of information-processing brain activity, measured non-invasively from the scalp. Brain Fingerprinting technology depends only on brain information processing, it does not depend on the emotional response of the subject.

Brain Fingerprinting technology utilizes multifaceted electroencephalographic response analysis (MERA) to detect information stored in the human brain. A memory and encoding related multifaceted electroencephalographic response (MERMER) is elicited when an individual recognizes and processes an incoming stimulus that is significant or noteworthy. When an irrelevant stimulus is seen, the P300/MERMER is absent. This pattern occurs within less than a second after the stimulus presentation, and can be readily detected using EEG amplifiers and a computerized signal-detection algorithm.

Brain Fingerprinting technology incorporates the following procedure: A sequence of words, phrases, or pictures is presented on a video monitor under computer control. Each stimulus appears for a fraction of a second. Three types of stimuli are presented: "targets," "irrelevants," and "probes." The targets are made relevant and noteworthy to all subjects: the subject is given a list of the target stimuli and instructed to press a particular button in response to targets and another button in response to all other stimuli. Since the targets are noteworthy for the subject, they elicit a P300/MERMER. Most of the non-target stimuli are irrelevant, having no relation to the situation under investigation. These irrelevants do not elicit a P300/MERMER. Some of the non-target stimuli are relevant to the situation under investigation. These relevant stimuli are referred to as probes. For a subject who as knowledge of the situation under investigation, the probes are noteworthy due to the subject's knowledge of that situation, and therefore probes elicit a P300/MERMER when the subject is "knowledgeable". Probes are indistinguishable from the irrelevants for a subject who is not knowledgeable, and thus probes do not elicit a P300/MERMER if the subject has no knowledge of the situation under investigation.

The entire Brain Fingerprinting system is under computer control, including presentation of the stimuli, recording of electrical brain activity, a mathematical data analysis algorithm that compares the responses to the three types of stimuli and produces a determination of "information absent" or "information present," and a statistical confidence level for this determination.

Dr. Lawrence A. Farwell, Director and Chief Scientist of Brain Fingerprinting Laboratories, Inc. in collaboration with then SSA Drew C. Richardson, Ph.D., FSRTC, Laboratory Division, FBI applied the Brain Fingerprinting system in the detection of FBI new agent trainees using FBI-relevant probes. The system correctly classified 100% of the 17 FBI new agents tested, as well as 4 control subjects who were lacking any knowledge regarding the FBI.

The Brain Fingerprinting system has potential application in a wide range of law enforcement and intelligence operations, from detecting whether a suspect has knowledge that would identify him as the perpetrator of a crime to detecting whether an individual has knowledge that would indicate that he had undergone training by a foreign intelligence or terrorist organization.

I. Introduction

This report describes new and potentially revolutionary technology for the detection of concealed information that revolves around the non-invasive recording of electrical brain activity. The electrical brain activity pattern recorded and of interest is a specific multifaceted electroencephalographic response (MER) that occurs within less than one second after an examinee is visually presented (via a computer screen) with words, short phrases, acronyms, or pictures that are recognized and cognitively processed by that subject. This phenomenon, coupled with its absence following the presentation of the same information to a subject for whom the material is unknown or irrelevant, is the basis for discriminating between a subject with “information present” and “information absent”. This would potentially allow for the determination of a whole host of issues of interest to the law enforcement and intelligence communities, e.g., (1) does a suspect have knowledge connecting him to specific investigated criminal activity, (2) does an intelligence source have knowledge of the internal workings of a hostile intelligence agency that would indicate that he was an intelligence officer of that agency rather than who he claimed to be, (3) has an informant, a debriefed spy, or a suspected member of a criminal organization accurately described the entirety of his actions and knowledge, (4) did a convicted serial killer who claims to have killed 40 to 50 individuals, other than the one(s) he was convicted of, actually commit these acts, or are these claims merely the bravado of a condemned prisoner.

The potential benefit of this program extends to a broad range of law enforcement applications, including organized crime, violent crime, white-collar crime, drug-related crime, foreign counterintelligence, non-traditional targets, and other categories of casework as well. This new technology promises to be of tremendous benefit both at the national level and for state and local law enforcement agencies.

This document describes a revolutionary technology that is capable of detecting concealed information stored in the brain through the electrophysiological manifestations of information-processing brain activity. Previous research conducted by Dr. Lawrence Farwell and his colleagues has shown this technique to be highly accurate in distinguishing between individuals with “information present” and “information absent” in mock crimes and actual minor crimes (see 1-9). Research conducted by Dr. Farwell in collaboration with SSA Drew C. Richardson, Ph.D., FSRTC, Laboratory Division, FBI indicates that the system is also effective in distinguishing between members of a particular organization (in this case, the FBI) and others who are not knowledgeable regarding that organization.

When a crime is committed, traces of the event are left at the scene of the crime and elsewhere. The task of the investigators is to reconstruct what has happened and who has been involved, based on the collection of such evidence. In addition to the physical and circumstantial evidence that can be obtained, there is one place where an extensive record of the crime is stored — in the brain of the perpetrator. If this record could be tapped, criminal investigation and counterintelligence could be revolutionized.

Until recently, the only method of attempting to discern what information regarding a crime or other situation of interest was stored in the brain of a suspect or witness has been (1) to interrogate the subject, and (2) to attempt to determine whether or not the subject is lying.

Conventional control question (CQT) polygraphy has been used as an aid in the attempt to detect deception in such reports. The fundamental theory of conventional polygraphy is that a deceptive individual will be more concerned with and experience more emotional arousal in response to relevant questions than control questions, and this emotional arousal will be accompanied by corresponding physiological arousal which can be measured. Traditional interrogative polygraph ("lie detection") methods rely upon using questioning formats in conjunction with the recording of physiological parameters that reflect autonomic nervous system (ANS) activity (e.g. blood pressure, heart rate, sweating, etc.). This information is peripheral to the cognitive aspects of deception or of concealing guilty information.

Multifaceted electroencephalographic response analysis (MERA) technology presents the possibility of focusing on the origins (at the level of subject recognition of guilty knowledge) of concealed information rather than the peripheral physiological manifestations of that knowledge. In addition to being a more direct physiological approach (central nervous system vs. peripheral) to the question at hand, the Brain Fingerprinting system may well overcome certain difficulties inherent with standard polygraphy: (1) Innocent as well as guilty individuals may respond emotionally and physiologically to crime-relevant questions, which may result in an innocent subject falsely being found deceptive; (2) guilty individuals may fail to respond in the expected way either emotionally or physiologically; (3) certain mental and physical countermeasures can be practiced successfully with standard technology; and (4) a conventional polygraph exam is highly stressful for the examinee, and involves deception by the polygraphers.

In a conventional polygraph test, emotion-driven physiological responses to relevant questions (regarding the situation under investigation) are compared to responses to control questions, which are invasive, personal questions not relevant to the issue at hand that are designed to be emotionally and physiologically disturbing to the subject. A greater response to the relevant questions leads to a deceptive ("guilty") determination; a greater response to the control questions leads to a non-deceptive ("innocent") determination. In an attempt to avoid a false positive result (non-deceptive subject falsely found deceptive), the examiner must ask penetrating questions in the pre-test interview to find personal material sufficiently disturbing and stress-producing to produce effective control questions. To elicit a stress response to the control questions during the test, the examiner typically deceives the subject, leading him to believe that a large response to control questions will make him appear guilty (deceptive), rather than innocent (non-deceptive). This deception by the examiner is necessary, or at least highly instrumental, to produce the response. Thus, in conventional polygraphy, innocent subjects — even if they are correctly determined to be innocent and truthful — are deceived and subjected to a highly invasive and stressful situation both during the pre-test interview and during the test.

This latter shortcoming is generally justified by the correct end result of finding an innocent subject non-deceptive to the relevant questions, but could be avoided altogether with Brain Fingerprinting technology, which depends entirely on information processing brain activity (i.e., recognition and processing of significant information) rather than an artful and disturbing manipulation designed to produce emotional and physiological responses to control question material. In fact, the pre-test interview for a Brain Fingerprinting technology exam is a very clinical, emotionally neutral experience for all subjects, whether or not they have specific information relevant to the situation under investigation. The in-test portion of the Brain Fingerprinting technology exam does not involve the asking of any questions, only the non-invasive recording of brain electrical activity as a subject views verbal or pictorial information on a computer screen.

Dr. Farwell, the Director and Chief Scientist of Brain Fingerprinting Laboratories, Inc., Fairfield, Iowa, and his colleagues have been conducting research on this new technique for several years. This research has been reviewed in a technical report prepared by Dr. Farwell for the United States Government (3), in US Patent #4,941,477 and three other US Patents; U.S. Patent #5,363,858: U.S. Patent #5,406,956 U.S. Patent #5,467,777 and in refereed scientific journal reports (see 1-9). Results have shown this technique to be capable of producing an "information absent" or "information present" determination, with a strong statistical confidence, in approximately 90% of the cases studied. All of the determinations were accurate: there were no false positives and no false negatives. In the other 10% of cases the mathematical algorithm determined that there was insufficient information to make either determination, and no determination was made, (i.e., "indeterminate" result). The cases studied included actual minor crimes and other real-life events as well as mock crimes.

Another study conducted by Dr. Farwell in collaboration with SSA Drew C. Richardson, Ph.D., FSRTC, FBI Laboratory, has also shown Brain Fingerprinting testing to be capable of detecting whether or not an individual has participated in FBI new agent training at the Academy. New FBI agents in training at the FBI Academy at Quantico were correctly identified as such, and individuals unfamiliar with the FBI were also correctly classified. The application of this technique in foreign counterintelligence is obvious: if this technology can be utilized to detect an FBI agent, it can also be used to detect agents of other organizations, including both intelligence organizations, international criminal organizations, and terrorist groups.

The detection of information stored in the brain is central to the investigation of all types of crimes; organized crime, violent crime, white-collar crime, drug-related crime, industrial espionage, non-traditional targets, as well as foreign counterintelligence operations. The far-reaching implications of possessing technology to accomplish this end are obvious. With the potential availability of such technology, it is felt imperative that this methodology be further tested as soon as possible, that the research and development necessary to make this a practical technique for field use be undertaken immediately, and that the technique be implemented as soon as possible by law enforcement agencies, if and when found valid and feasible. The research reported here is the first step in making this new technique and technology available not only to the FBI and other national agencies, but also potentially to state and local law enforcement agencies.

II. Overview of Brain Fingerprinting Technology:

Instrumental Requirements and Multifaceted Electroencephalographic Response Analysis

The equipment required for Brain Fingerprinting technology consists of a personal computer (e.g., 486 - 66 MHz Gateway 2000), a data acquisition board (e.g., Scientific Solutions Lab Master AD), a graphics card for driving two monitors from one PC (e.g., Colorgraphics Super Dual VGA), a four-channel EEG amplifier system (e.g., Neuroscience), and the software developed by the Brain Fingerprinting Laboratories for data acquisition and analysis. The electrodes used to measure electrical brain activity are held in place by a special headband designed and constructed by Brain Fingerprinting Laboratories for this purpose. (This new method for attaching electrodes is more convenient and comfortable for the subject as well as quicker and easier for the operator than previously available methods.) The software necessary to present the stimuli, collect the electroencephalographic data, and analyze the data have been developed by Brain Fingerprinting Laboratories.

In experiments using the Brain Fingerprinting system, brain responses to certain types of stimuli are analyzed to detect a specific multifaceted electroencephalographic response (MER) known as the memory and encoding related multifaceted electroencephalographic response (MERMER). The P300/MERMER contains several facets, both in the time and frequency domains, that can be detected with sophisticated signal analysis procedures. It is elicited by stimuli that are noteworthy to a subject.

The Brain Fingerprinting system presents visual stimuli consisting of short phrases, acronyms, or pictures on a video screen under computer control. Three categories of stimuli are presented: "probes," "targets," and "irrelevants."

Probes are stimuli relevant to the crime under investigation. Irrelevants are, as the name implies, irrelevant. For each probe stimulus, there are approximately four irrelevant stimuli. The stimuli are structured such that the probes and irrelevants are indistinguishable for a subject lacking specific information relevant to the situation under investigation. That is, if a given probe is an article of clothing relevant to the crime, four articles of clothing irrelevant to the crime are also presented; if a particular probe stimulus is a name, there are four irrelevant stimuli that are also names, and so on.

In addition to the probes and the irrelevants, a third type of stimuli, designated as targets, is presented. About one-sixth of the stimuli are targets, one for each probe. The subject is given a list of the targets, and is required to press a particular button whenever a target is presented. (For all other stimuli, the subject is instructed to press another button.) Each target is the same type of item as one of the probes and the several corresponding irrelevants. The targets, since they are recognized and require a particular response, are noteworthy for all subjects. The irrelevants are not noteworthy for any subjects. The probes are noteworthy only to the subjects who possess the knowledge necessary to recognize them — that is, the knowledge specific to the situation under investigation.

Research has shown that brain information processing of noteworthy stimuli results in a characteristic brain electrical response known as a memory and encoding related multifaceted electroencephalographic response (MERMER). One of the most easily measured aspects of this response (and the only one measured in early research) is an electrically positive component, maximal at the midline parietal area of the head, with a peak latency of approximately 300 to 800 msec. It is referred to variously as P300, P3, P3b, or late positive component (LPC). Another more recently discovered aspect of the MERMER is an electrically negative component, maximal at the midline frontal area, with an onset latency of approximately 800 to 1200 msec. These components can be readily recognized through signal averaging procedures (1-9). Recent research suggests that a third aspect of the P300/MERMER is a pattern of changes in the frequency domain characterized by a phasic shift in the frequency power spectrum that can be detected using single-trial analysis techniques (8).

A subject lacking specific information relevant to the situation under investigation recognizes only two types of stimuli: relevant, noteworthy, rare targets and irrelevant, frequent stimuli (consisting in fact of true irrelevants, plus probes — which he does not distinguish as being different from the irrelevants). The targets elicit a P300/MERMER, and the irrelevants and (unrecognized) probes do not. A subject with specific information relevant to the situation under investigation, however, recognizes a second noteworthy type of stimuli, namely the probes, which are relevant to a crime or other situation in which he has participated. Thus, for a guilty subject, the probes, too, elicit a P300/MERMER.

What this experimental design accomplishes, essentially, is to create a two-stimulus series for an individual without the specific information under investigation, and a three-stimulus series (with the same stimuli) for an individual who possesses knowledge of the specific information under investigation. The targets provide a template for a response to stimuli known to be particularly noteworthy — P300/MERMER-producing stimuli. The irrelevants provide a template for a response to stimuli that are irrelevant — non-P300/MERMER-producing stimuli.

The determination of information present or information absent consists of comparing the probe responses to the target responses, which contain a P300/MERMER, and to the irrelevant responses, which do not. If the probe responses are similar to the target responses, one can conclude that the subject recognizes the probes — which only someone knowledgeable about the crime would do — and therefore is "knowledgeable". If the brain responses to the probes are like those to the irrelevants — i.e., lacking a P300/MERMER — then the subject can be determined to be lacking in knowledge of the situation under investigation. (Note that what is detected is not actually guilt or innocence, but knowledge or lack of knowledge regarding the situation under investigation.) In order for the test to be an effective indicator of knowledge or lack thereof, stimuli must be structured such that only a knowledgeable person would recognize the probe stimuli. The statistical technique of bootstrapping is employed to compare the brain responses to the different types of stimuli, to make a determination of "information absent" or "information present," and to provide a statistical confidence for this determination.

Table 1 summarizes the three types of stimuli presented and the predicted brain responses to each type of stimulus.

Table 1

 BRAIN FINGERPRINTING TECHNOLOGY

TYPES OF STIMULI AND PREDICTED BRAIN RESPONSES

III. Results of the FBI Agent Study

The results of the study of FBI agents were as predicted. In every case, FBI agents showed large P300/MERMERs in response both to the targets and to the FBI-relevant probes. The subjects who were not knowledgeable regarding the FBI showed large P300/MERMERs only to the targets.

Bootstrapping analysis was conducted to make a determination of whether the subject was knowledgeable regarding the FBI (an "information present" determination), or was lacking in such knowledge. The Brain Fingerprinting system yielded a correct determination in every case, both for the FBI agents and for the control group, along with a statistical confidence for the determination. The determinations are summarized in Table 2. The determinations and the statistical confidence for each are listed in Table 3.

As can be seen in Tables 2 and 3, 100% of the determinations were correct. There were no false positives, no false negatives, and no indeterminates.

Table 2

 BRAIN FINGERPRINTING TECHNOLOGY:

SUMMARY OF DETERMINATIONS

Validity (excluding inconclusives) 100%

Validity (including inconclusives) 100%

Table 3

 BRAIN FINGERPRINTING TECHNOLOGY:

DETERMINATIONS AND STATISTICAL CONFIDENCE

A. FBI Agents ("Information Present")

B. Non-FBI Agents ("Information Absent")

V. Abstracts of Recent Research

The following pages present abstracts of recent presentations and publications of research on earlier versions of Brain Fingerprinting technology. At the time of these publications, the MERMER had not yet been discovered, and discriminations between different brain responses were made on the basis of the P300.

Professional Profiles: Event-related Potential
Detection of Concealed Occupational Information in FBI Agents

Lawrence A. Farwell

Abstract of an address presented at the Annual Meeting of the

American Polygraph Association, Orlando, FL,
August 1992

The EEG-based Brain Fingerprinting system (Farwell, 1991; see also Farwell, 1992; Farwell & Donchin, 1986, 1991) presents a new paradigm in the psychophysiological detection of concealed information. In previous research, the system was used to detect possession of knowledge regarding real and mock crimes. Three types of stimuli were used: crime relevant "probes," "irrelevant" stimuli, and "targets," a subset of non-crime-relevant stimuli in response to which a particular button press was required. Characteristic brain responses were elicited by the rare, relevant stimuli: targets for all subjects, and both targets and probes for subjects with specific information relevant to the situation under investigation. Comparison of the event-related potential responses using a mathematical bootstrapping algorithm produced an information present or information absent determination and a statistical confidence for the same. The present study introduces two innovations: 1) what is detected is not ‘guilty’ knowledge regarding a specific incident, but rather a more generalized class of information that would be possessed only by individuals with particular occupational knowledge or professional expertise, in this case, FBI agents; and 2) the targets are a subset of the relevant stimuli. Targets, as before, have been identified to the subject, and a unique task is required in response to them. The pot