NEUROETHICS

They (Might) Know What You’re Thinking

Will fMRIs Be the New Polygraphs?

The notion of detecting deception or reading minds with a machine conjures Orwellian images of a world in which technology and government action render even privacy of thought impossible.  Ongoing research on functional magnetic resonance imaging (fMRI) suggests that such concerns might soon become reality.

Because mind reading and near-flawless detection of deception would avoid many of the limitations of conventional lie detectors, national security agencies and the courts stand to significantly benefit from this technology.  Investment in this research by the Pentagon and by private companies highlights interest in its development.

In the most common form of fMRI research, scientists focus their observations on oxygen in the brain, because the magnetic properties of hemoglobin change when it carries oxygen.  The fMRI scanner transmits information on the rate of oxygen-carrying blood flow in the brain to a computer, where software translates data into a representation of the brain. This is not a picture of the brain, but rather a map of blood flow in the brain.  With this information, researchers can correlate increases in blood flow during controlled activities with various brain functions.  Experiments built on this technology can determine with accuracy whether or not a subject under observation is making truthful statements.

Most ethical and policy concerns do not require immediate attention, given the present state of this technology.

Some cognitive researchers predict a startling increase in the accuracy of this technology, making the detection of deception nearly flawless. If that ever happens, concerns arise about forensic applications. For example, some worry that a machine of such accuracy will challenge the “province of juries.”  National security interests would almost certainly harness such technology for both interrogations and interviews.  And will this extend to regular penetration of civilian thought?  Fortunately, most ethical and policy concerns do not require immediate attention, given the present state of this technology, especially considering that national security and court use hinges on the ability to accurately scan individual subjects. To remove “background noise” from a scan, researchers establish a baseline and subtract this recording from experimentally obtained data. This works well for groups but poorly for individual analysis; distinguishing signal from background noise on an individual level proves difficult.[1]

Generating a representation of this neural activity presents additional problems. For example, current techniques measure changes in neural activity on the order of seconds, where most neural processes occur on the order of milliseconds. Additionally, when software translates data into images, smoothing and magnification to ensure a view of high-order brain functions reduces resolution.[2]

There are other limits to fMRI.  Few studies investigate variance between scans of individuals of differing age, sex, race, fitness level, cultural background, or account for the prevalence of mental disorder and other characteristics that might have a significant effect on the data gathered in experiments.  Researchers also face the challenge of testing fMRI for detecting deception in settings more realistic than the lab.  Lying in court generally entails both rehearsal and reasoning.  In this setting, suspects feel emotional pressure and stress as they try to deliver a convincing story. Lab studies don’t replicate that kind of pressure.

fMRI lie detection likely falls under the existing policy for polygraph use. State law and federal law differ with regard to lie detection through polygraph, and will probably differ with regard to fMRI lie detection.  Federal law seems more permissive of polygraph in courts, whereas 27 states and DC entirely prohibit its use.[3]   Until recently, a 1923 court precedent, called the Frye Standard, arising out of Frye v. United States, set the general acceptance of a technology or scientific development as the standard for admission in court.

Although national security interest in fMRI extends to interrogation, screening job candidates and employees seems the most practical application.

Today, Rule 702 of the Federal Rules of Evidence supplements this criterion, explaining, “if scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact at issue, a witness qualified as an expert by knowledge, skill, experience, training, or education may testify thereto in the form of an opinion or otherwise.”[4]  This rule allows judges to determine whether the science has been tested, subjected to peer review, the known or suspected rate of error, and whether or not the scientific community generally accepts the technology.[5]  Under the Frye Standard, courts prohibited polygraph use, however, after Rule 702, exclusions of polygraph tests no longer presented a foregone conclusion. If fMRI scans for lie detection reach 99 percent or better accuracy, standard use in courts seems possible.  So what happens then?

It is clear that regulation of fMRIs for lie detection parallels, in some ways, regulation of polygraphs.  For example, implementation in courts requires evaluation under Rule 702, including the general acceptance of efficacy, something which has yet to occur. Despite these similarities, fMRI brings up many new issues associated with both its potential accuracy and invasion of privacy. Jurors without knowledge of error rates and competing evidence will likely give undue credibility to an fMRI reading. The difference between perceived accuracy and actual accuracy jeopardizes the likelihood of an “impartial jury,” as guaranteed under the Sixth Amendment.  As neuroscience scholar and neuroethicist Judy Illes explains, “The sheer complexity of neuroscience research poses challenges for integration of knowledge and meaningful interpretation of data.”[6] The most challenging problem at this time, she argues, is that, “with dynamic images in hand, we may forget the limits of how the images were produced, including variability in research designs, statistical treatment of the data, and resolution.”[7]

There are other constitutional issues.  The courts will have to determine whether fMRI lie detection constitutes an unreasonable search or seizure under the Fourth Amendment, and high-accuracy lie detection likely carries implications on the Fifth Amendment clause on self-incrimination.

Specific issues arise with the use of fMRI in the context of national security.   For example, would this technology be applied in interrogations as an acceptable alternative to currently controversial techniques?  At least one former intelligence officer has stated that it has been used in the “war on terror.” However, applying fMRI in this context still seems to invite debate on questions of international law.  Although national security interest in fMRI extends to interrogation, screening job candidates and employees seems the most practical application.  National security agencies are among the largest users of polygraphs at present, despite their flaws, so a promising technology like fMRI surely opens new doors.

The Employee Polygraph Protection Act prohibits the administration of polygraph to civil employees, except in the case of ongoing investigations. However, this act does not apply to government employees, specifically exempting those working in national security with access to sensitive information.  So just as national security agencies legally use polygraphs, legal application of fMRI seems possible.

Even if researchers overcome the technical problems in brain-scanning lie detection, other practical issues make implementation of fMRI challenging.  For instance, the size of an fMRI machine and the necessary subject cooperation present serious problems—a person must lie prone inside a giant magnet and hold perfectly still for the device to function properly. Portable fMRI machines or long-range fMRI scanners might quell these issues, but such developments require significant scientific leaps. Gathering any significant valid data from a scan also requires an immobile subject.  Even moving one’s tongue during a scan compromises the validity of results.  At this point in time, the cost and inconvenience of using this technology makes its reliable use unrealistic, even if its accuracy is incredibly high, except in the most pressing of matters.

But the technology has to make significant strides before any of these issues become pressing. Improvements in fMRI brain scans stand to benefit medical treatments as well—imaging brain tumors, for example—and have applications outside of national security and the legal system.  The reservations and cautions expressed here are not reasons to inhibit the development of this technology, but are rather encouragement to continue.

Michael Peroski is an undergraduate majoring in biochemistry and philosophy at Allegheny College.

Notes

[1] Pearson, Helen. “Lure of Lie Detectors Spook Ethicists.” Nature 441 (2006):  918-919.

[2] Yousuf, Sameer.  “The Legal and Ethical Implications of fMRI Lie Detectors.”  Unpublished thesis (2007).

[3] Ford, Elizabeth B.  “Lie Detection: Historical, Neuropsychiatric, and Legal Dimensions.”  International Journal of Law and Psychiatry 29 (2006): 159-177.

[4] “Federal Rules of Evidence.” Cornell Law School.  18 Feb. 2008. http://www.law.cornell.edu/rules/fre/rules.htm

[5] Ford (2006); Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 US 579 (1993)

[6] Illes, Judy and Racine, Eric.  “Imaging or Imagining?  A Neuroethics Challenge Informed by Genetics.” The American Journal of Bioethics 5 (2005): 5-17.

[7] Illes and Racine (2005).

Comments on this article

You can skip to the end and leave a response. Pinging is currently not allowed.