Law & Psychiatry: Through a Glass Darkly: Functional Neuroimaging Evidence Enters the Courtroom

The use of a PET scan to support Braunstein's claim that he could not form an intent to commit the crimes with which he was charged is one of the more dramatic examples of the entry of functional brain scanning into the courts. But it was by no means the first time brain scans have played a prominent role at trial. Many Americans first became aware of structural brain scanning technology when a defense expert introduced computerized axial tomography (CAT) scans of John W. Hinckley's brain to support the contention that Hinckley—being tried for the attempted assassination of President Ronald Reagan—suffered from schizophrenia ( 6 ). The development of functional scanning, including PET, single-photon emission computed tomography (SPECT), and most recently, functional magnetic resonance imaging (fMRI), have multiplied the forensic uses of brain scanning technology.

Underlying all types of functional scanning evidence is the assumption that disruptions in the function of discrete parts of the brain can lead to alterations in particular aspects of cognition and behavior. The classic 19th century case of Phineas Gage, the railroad worker who suffered a dramatic change in his personality after an iron tamping bar was shot through the front of his brain, is often used to illustrate the phenomenon ( 7 ). Gage, a hard-working family man, became an impulsive, mildly antisocial ne'er-do-well in the wake of his accident, a change attributed to damage to his prefrontal cortex, often considered the brain's locus for reasoning, planning, impulse control, and moral judgment. Other brain areas linked to legally relevant functions include the amygdala (involved in control of emotions, including anger) and the anterior cingulate cortex (which plays a role in the generation of empathy) ( 8 ). Scanning techniques are based on the presumption that changes in blood flow—the parameter actually measured by the technology—correspond to changes in neural activity.

Functional neuroimaging evidence has been used in criminal cases in support of the insanity defense ( 9 ), claims of incompetence to stand trial ( 10 ), and pleas for mitigation in sentencing—a use pioneered by Dr. Buchsbaum ( 11 ). Civil cases in which functional imaging has been introduced typically involve claims of brain injury as a result of defendants' activities ( 12 ) but include more esoteric uses as well, such as to establish the effect of violent video games on children ( 13 ). A 2006 study found 133 reported opinions involving PET or SPECT evidence, in 85 of which the evidence was ruled admissible ( 12 ). Because only a fraction of legal cases yield written opinions, and given the greater ease and finer resolution of fMRI imaging compared with PET and SPECT, the number of cases involving functional neuroimaging by now is undoubtedly at least several times higher. But the courts' receptivity to functional imaging evidence has not gone unchallenged.

From a scientific perspective, there are a number of conceptual issues in relating imaging findings to particular cognitive functions (for example, to impulse control or planning), including the assumption that complex functions are localized to single areas of the brain ( 14 ). Not only do current models suggest that higher-level brain functions typically are dependent on interactions among a distributed web of neural centers, but the ability of other brain regions to take over from injured areas suggests limits to the conclusions that can be drawn from localized deficits. Moreover, to the extent that data link reduced activity in a particular brain region to a pattern of impaired cognition or behavior, the findings are almost always based on correlations between aggregate data and outcome measures, which can obscure the extent of individual differences. Thus a range of functional levels could be compatible with unimpaired cognition and behavior, including some that may be significantly different from the statistical mean ( 15 ).

How functional scanning results are presented can also be problematic. Levels of activation are determined by subtracting resting levels of activity from those achieved when the brain is being challenged to perform a particular task. Thus, when a bright patch of color fails to appear in a brain region of interest, the finding does not signify that the region was inactive, only that it was not significantly more active in response to the stimulus than in its absence ( 12 ). In theory, a person who is more efficient than the average subject at performing a particular task may show results that falsely suggest a failure to activate the relevant region. And the levels at which a significant difference is identified—perhaps not much above baseline—and the colors used to illustrate it are often selected arbitrarily. Thus dramatic contrasts on a scan can reflect relatively minor differences in actual levels of brain activity.

In the courtroom setting, however, perhaps the most important limitation on the current use of functional imaging evidence is the discrepancy between the inferences that can be drawn and the legal standards to be met. To be found not guilty by reason of insanity, for example, a defendant must demonstrate (in one common formulation) that he "lack[ed] substantial capacity either to appreciate the wrongfulness of his conduct or to conform his conduct to the requirements of the law." Although neuroscientists talk in general terms about brain regions associated with moral reasoning and behavioral control, at this point there are no data identifying a given degree of impairment in a particular locus of the brain that corresponds to substantial impairment in either of the capacities at issue. Hence, there is an understandable resistance among legal and neuroscience experts to accepting neuroimaging data as evidence of "aberrations in human thought, will, motivation or propensity for culpable behavior, or to show an incapacity to inhibit … behavior" ( 16 ). To behave otherwise, in one commentator's caustic turn of phrase, would be to exhibit symptoms of "brain overclaim syndrome," a condition that threatens to undermine the integrity of the process of determining responsibility for criminal acts ( 15 ).

Peter Braunstein's trial on charges of kidnapping, sexual abuse, burglary, robbery, and arson turned out to be a case in point. Deliberating only four hours before finding the defendant guilty on all charges except arson, the jury appeared to accept the conclusion of the prosecutor that "It really was an indefensible defense. All the actions, the record, his writings, his interviews were very clear. All the events leading up to it, the preparation for the crime, the actual execution of the crime, his getaway. It was a really conscious effort on his part to commit this crime" ( 17 ). Whatever the PET scans of Braunstein's frontal lobes showed, the jury was unwilling to accept that his deficits rendered him "completely unable to plan," which a finding of lack of intent would have required. The commonsense view that the crime was too intricate to have been carried out by someone incapable of forming a criminal intent won the day.

What is the likely future of neuroimaging evidence in court? As noted, functional scanning already appears to be playing a useful role in demonstrating impairments that result from physical injury, including exposure to toxins. In addition, in some cases where gross lesions are present and a person's behavior showed clear changes after the lesion developed ( 8 ), imaging data may be helpful in identifying the cause of the deviant actions, although structural imaging is more likely than functional imaging techniques to be used for this purpose. Moreover, to the extent that convincing data can be generated to demonstrate strong associations between specific deficits on scanning paradigms and relevant mental functions (for example, impulse control), such data could be used to support defendants' claims of absent or diminished responsibility. But without evidence drawn from defendants' own behavior and clinical state, functional scanning data are not likely to stand on their own for such purposes. It remains too early to judge whether neuroimaging evidence will become anything more than "a high-tech crystal ball" ( 18 ).

Dr. Appelbaum, who is editor of this column, is the Elizabeth K. Dollard Professor of Psychiatry, Medicine and Law, Department of Psychiatry, Columbia University. Send correspondence to him at New York State Psychiatric Institute, 1051 Riverside Dr., Unit 122, New York, NY 10032 (email: [email protected]).