NPR logo

Can MRIs Help Solve Crimes?

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Can MRIs Help Solve Crimes?


Can MRIs Help Solve Crimes?

Can MRIs Help Solve Crimes?

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

What if police could scan a suspect's brain to see if he was lying? Some companies claim the technology works, and it should be allowed as evidence at trial. Law professor Hank Greely explains the state of the technology and the ethical questions surrounding its use.


That's the flute, and that means we're switching to another subject and another guest. And as promised, neuroscience and the law.

At least two companies now claim that they've developed a brain scan that can tell you whether you are lying. You can't get away with it anymore, at least according to what they say.

Here's how it might work. You get arrested for a crime. No, nobody who listens to SCIENCE FRIDAY gets arrested for a crime, but somebody gets arrested for a crime. You say you are innocent. And in order to prove it, you are willing to undergo a brain scan. It's like a modern lie detector, but it's far more high-tech than those old things with all the wires and the, you know, the pen wiggling on the scrolling graph paper. I kind of like those things. But this is high-tech stuff. This is neuroscience.

One company offering the scan, its scan is called No Lie MRI. Honest, No Lie, that's what it's called. The company says its scans are 90 percent accurate. But is the technology really there yet? And even if the test was perfect or close to perfect, 90 percent accurate, could we always separate the liars and truth tellers? Do we - if we could separate them, do we have the right to use this on anyone? Does your brain have a right to privacy? It's been protected in that hard case for lo these millennia - now all of a sudden people are looking in there. Do you have that right to privacy? And should it be admitted in the courtroom?

So to talk about some of these issues and see if we can get all the answers, we have brought on an expert on neuroscience and the law, my guest, Professor Hank Greely, a professor of law at Stanford University. Thanks for talking with us today and volunteering to give us all the answers.

Professor HENRY GREELY (Stanford Law School): Well, my pleasure, although I'm not guaranteeing that I'm giving you all the answers.

(Soundbite of laughter)

RAEBURN: I'm trying to box you in a corner here before we get started.

Prof. GREELY: Yes, yes, no, C, and all of the above.

RAEBURN: Right. Spoken like a lawyer, what can I tell you? So anyway now, some of these companies that have talked about this are claiming these high accuracy rates. Do you believe them?

Prof. GREELY: No.

(Soundbite of laughter)

Prof. GREELY: No. Although in fairness there are now over 20 peer-reviewed published studies that find a statistically significant correlation between certain patterns of brain activation seen on fMRI and - when people, who are experimental subjects are saying something deceptive. So there is some scientific basis for thinking that there may be something to be found there. I don't think it's nearly proven enough to use in court or anyplace else where we rely on it. And I think the very high accuracy rates being portrayed by the companies are just wrong.

RAEBURN: Now, what is your background in this area? You're a lawyer. Did you -do you specialize in this particular neuroscience law intersection?

Prof. GREELY: Yes, in the last eight years, since 2002, most of our work has been on legal issues in neuroscience. I've been a co-director of the Law and Neuroscience Project funded by MacArthur Foundation. And I'm one of the directors of the Neuroethics Society, which is a scholarly organization devoted to looking at the ethical, legal and social implications of neuroscience on our society, which we think are going to be enormous.

RAEBURN: Now, I have to be at the AAAS session - the American Association for the Advancement of Science session in February. I guess it was in San Diego where you were the prosecutor in a mock trial...

Prof. GREELY: Okay.

RAEBURN: ...looking at neuroscience in the courtroom. They put on your - have you been a prosecutor?

Prof. GREELY: I am not. I was a lawyer. I'm a recovering lawyer. I'm, you know, 25 years into my recovery.

RAEBURN: You dont admit to being a lawyer any longer.

Prof. GREELY: I keep my license up but I'm not sure why.

(Soundbite of laughter)

RAEBURN: Okay. So, in any case, tell us a little bit about that case. That was a little bit different from lie detection as I recall.

Prof. GREELY: Yeah.

RAEBURN: What was the issue there? What were you trying to demonstrate as the prosecutor?

Prof. GREELY: Well, there are a couple different things going on, but that case was an example of where defense - criminal defendant was saying there's something unusual about my brain. That means you really shouldn't hold me responsible for my actions. So we created a mock case. It was the AAAS representatives who did most of the work on it, Deborah Runkle and Mark Frankel.

We created a mock case and then had a couple of lawyers, a real judge and a couple of real neurologists as expert witnesses. The theory was that this person had a problem in his frontal lobes that meant that he wouldn't had been able to form the intent to commit a murder, that he wouldn't had been able to premeditate, and that's what we argued in that particular case.

There are lots of different ways you can make these arguments about brain defects or brain problems leading to limited responsibility. You could try to argue that it's evidence for an insanity plea, it's evidence that the person is incompetent to stand trial, it's evidence that they couldn't form the right intent - ots of different arguments that can be used. It's probably come up most often, though only one so far with fMRI. Most of it has been structural MRI. In capital cases, where after the defendant is found guilty, there's a separate hearing to decide whether or not he should receive the death penalty or life in prison, brain scans have been used in those cases for a long time.

RAEBURN: So this is - there's a lot at stake with the accuracy of these things and the admissibility?

Prof. GREELY: Yeah.

RAEBURN: Lives are at stake if it's used in sensing procedures.

Prof. GREELY: Yes. And if it's used in guilt and innocence issues, how many years in prison is not a good thing for an innocent person to have to go through. So there's a - when you move this to the courtroom, the stakes are much, much higher than they are when you're rounding up the usual subjects, the undergraduate psych majors for an experiment in your lab.

RAEBURN: Now, as the prosecutor in that mock trial, you argue that the brain scan evidence should not be admitted that the person was guilty.

Prof. GREELY: Correct. Well, the brain scan evidence shouldn't be admitted, both because there wasn't enough of a scientific basis for it and because, even if there was, it was not legally relevant to that particular legal question. The law may care that whether you've got the intent to do something, and may care about whether you're insane and whether you're competent, but the law doesn't particularly care whether you had brain damage or not. Lots of peoples' brains are unusual. Everybody's brain, in some detail, is different from everybody else's. You've got to show something awfully strong and, I think, often through - much more frequently through behavioral evidence and through scanning evidence about a defect. But these cases are happening.

I'm part of a group that's looking for all the cases in California courts in the last three years that have involved neuroimaging evidence in criminal cases, and we've found about 25 to 30 of them so far.

RAEBURN: And what's - what has been - what are the outcome has been on those cases or how is the evidence been used?

Prof. GREELY: Well, we haven't full analyzed all of them. One of the things we've discovered is getting documents out of court clerks is a lot harder than most other things we've tried to do.

RAEBURN: I know a lot of reporters who would be sympathetic on that score, actually.

(Soundbite of laughter)

Prof. GREELY: Yeah. Yeah. My best research assistant is a law student who used to be a reporter and..

RAEBURN: Is that right?

Prof. GREELY: ...had been working very well on this.


Prof. GREELY: But we found that they've - they have had very mix success rates. They seemed most frequently to be used, or thought to be used, in the most serious cases, the murder cases, particularly those where capital punishment is an issue. And sometimes, they look like they may have been somewhat effective and other times not.

RAEBURN: Okay. And so it's a developing area of the law. Has anybody established any procedure yet for using these things?

Prof. GREELY: You know, not really. There isn't a great deal of agreement on what counts, what would be - what kind of a brain scan image would really indicate strongly something that would be of legal relevant. So, you know, we're still at very early days. I don't know how big a deal this will end up being in the law. I think the people who are involved in it from the legal side disagreed to some extent on how important to legal matter this will be.

I frankly think the more interesting and potentially important stuff will be the kind of lie detection or other kinds of mind reading that fMRI let us do. You know, one example that we don't think about very often - it's not very sexy but it's very important in the law - is pain. Whether somebody is having pain or not, and if so, how much, is a huge issue in lots of civil trials, usually not criminal cases. But everything from auto accidents, to workers compensation cases, to Social Security disability determinations. And there are neuroscientists who are beginning to think they can see pain in the brain. That could really revolutionize a lot of court cases.

We havent seen it introduced in court yet. At least I havent. But I know that at least one friend of mine has been was hired as an expert witness to testify about this, and the case had settled before it went to trial.

So between looking for pain, looking for deception, looking for evidence, whether or not a witness has remembered something by looking at brain scans, all these are going to, I think, raise major issues for the legal system.

And soon, in fact, just today, this morning, a hearing ended in Jackson, Tennessee, about whether an fMRI-based lie detection report should be admitted into evidence in the criminal case. We don't have an answer yet. The judge will make a decision - make a recommendation to the district court judge sometime in the next couple of weeks. But this is a very real issue. It's on us now.

RAEBURN: Let me take a pause to say I'm Paul Raeburn.

And this is SCIENCE FRIDAY from NPR.

So you've obviously got your tentacles out there. You're on top of these things, that this just happened this morning, and you're already aware of it.

Prof. GREELY: Well, it's been a pretty exciting time. There was this as far as we can tell, this is only the second time an argument has been made that fMRI-based lie detection should be allowed into court. The first time was last week, a case in Brooklyn - a civil case where the trial court judge decided not to allow the evidence in.

The difference between that case and this case is that in Brooklyn case, the judge did it without hearing any scientific evidence. He didn't rule on whether the evidence - whether the scientific method was reliable or not. He just said that in his view, it shouldn't be allowed to go to the credibility of the witness as that was the role of the jury.

In this federal case in Tennessee, though, there were expert witnesses, both trying to convince the judge that the reports should be admitted and trying to convince him that there wasn't enough science to back it up and it shouldnt be admitted. So this is a case to watch very closely.

RAEBURN: Let's take a question from Martin(ph) in Santa Clara. Are you with us, Martin?

MARTIN (Caller): Yes, I am. Thanks for the taking the call. A very good topic.

RAEBURN: Hi. Welcome to SCIENCE FRIDAY. Yeah, go ahead.

MARTIN: Well, real quick question, questions.


MARTIN: My sister certainly has her own mental issues. But for people that really believe that something has happened, my sister looks into her past and thinks things that happened and they really havent, I'm sure she could be convinced that she stole something or hit someone in a vehicle and it really didnt happen. But would her brain scans show that indeed she's guilty, is this where, you know, kind of a problem with this whole, you know, kind of technical analysis of someone's brain?

RAEBURN: Interesting question, Martin. Thanks for calling. Hank Greely, go ahead.

Prof. GREELY: It's a great question. And the answer I think has to be that we dont know yet. We suspect, I think, at least, I suspect, that people who are convinced that what they're saying is true, even if it isnt, would not fail an accurate lie detector. But nobody has really done that work yet.

There is a little bit of parallel work on the pain detection. People have hypnotized undergraduate students to feel pain when they shouldnt, when there's no real reason for them to feel pain. And their brains show that they're in pain. They've put people with amputees with phantom limb pain -whose hand, which they havent had for 20 years, still hurts - put them in the scanner and they show that they're having pain. That's a little bit of parallel evidence that what the brain perceives as real may show up in a scanner, even if we have other reasons to think it's not real. That might apply to these memories as well.

RAEBURN: And we just have a little bit of time left. So let me ask you the question all reporters like you ask, which is to predict the future. And that is in this case, are these things going to become more accurate? Will they become more widespread? Will they become used in the courtroom and in other aspects of our lives, do you believe?

Prof. GREELY: Well, I think, certainly our ability to read minds, and it's a strong term, but I think it's accurate one here, through fMRI and other neuroimaging, is going to get better and better, both from technical advances and through statistical advances. Whether it will ever be good enough to be used in the courtroom remains to be seen.

And I think the answer will probably vary from application to application, and context to context. It might be, for example, that pain is good enough to use, but deception is not; or looking for bias is good enough to use, but looking for memories may not turn out to be. It's going to require, though, a lot of research to see just how good these are, whether they work for everybody, whether they there are some people for whom these methods dont work at all.

So I think we've got a long way. Another three, five, 10 years before we have a very good sense of just how accurate any of this is.


Prof. GREELY: So if it is accurate, then we got to decide whether we want to use or not. And if so, how? Can we force somebody to go through a brain scan? Should we be able to get should the police be able to get a warrant to search your brain? Is there a Fifth Amendment issue here against self-incrimination? If this stuff works, all those questions are going to have to be answered.

RAEBURN: Okay. Well, thanks for taking the time to be with us. Can we kind of need you keep following this us for and report back?

Prof. GREELY: Oh, definitely.

RAEBURN: My guest has been Hank Greely, professor of law at Stanford University. Thanks for being on SCIENCE FRIDAY.

Prof. GREELY: My pleasure.

Copyright © 2010 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.