Brain scans can reveal how you think and feel, and even how you might behave. No wonder the CIA and big business are interested. By Ian Sample and David Adam
Thursday November 20, 2003 The Guardian
Earlier this year, a group of American students volunteered their brains for a cutting edge neuroscience project at Baylor College of Medicine in Houston. The research used a technique that could watch their brains at work as they made decisions. At first glance, this seems nothing extraordinary: brain-imaging tools have been used routinely for years to assess damage caused by stroke, to hunt for brain tumours, and even identify the grey matter associated with language, love and memories.
But this study was different. As each volunteer took their turn to slide into the coffin-like cylinder of the scanner, sticky fluids were squirted into their mouths. As unlikely as it sounds, the students were using multimillion pound medical equipment to take the Pepsi challenge. Read Montague, the neuroscientist behind the Baylor experiment, is not alone in pushing the boundaries of neuroscience beyond the clinical. In recent years, a growing number of researchers have used brain-imaging equipment to try to reveal our innermost thoughts and feelings in less conventional "social neuroscience" experiments. As well as brand loyalty and consumer choice, neuroscientists are probing violent tendencies, moral reasoning, feelings of love and trust, and notions of justice. Just this week, researchers claimed to have used a technique called functional magnetic resonance imaging (fMRI) to identify brain activity associated with racial prejudice.
While standard MRI machines like those still found in many hospitals take a snapshot of the brain, functional MRI is newer and more powerful because it takes lots of these snapshots one after the other, revealing how thoughts unfold over time. But the trend for using fMRI to probe social and behavioural issues is prompting some scientists to ask big questions about where this may all lead. Could it only be a matter of time before neuroscientists have techniques that can reveal secrets we would rather keep tucked under our skulls? According to some leading scientists, this isn't a paranoid over-reaction. "The CIA has been interested in fMRI for years as a means of doing lie-detection tests," says Bob Turner, an fMRI expert at University College London. After all, he says: "The brain can't lie."
As scientists unravel the links between how the brain looks and how it functions, some believe we will also be able to use images of the brain to see how people will behave. "There's no scientific distinction between prediction and understanding how the brain works," says Stephen Smith, associate director of the Centre for Functional Magnetic Resonance Imaging of the Brain at Oxford University.
The suggestion that brain scans could reveal not just our future health, but the intricacies of our personalities and how we might behave in a given situation, is unsettling enough to some scientists that they want legislation to stop brain-scan records falling into the wrong hands. "We're starting to get detailed information from these brain-scan experiments and soon people are going to be able to use it to predict an individual's behaviour," says Paul Glimcher at the Centre for Neuroscience at New York University. "That information has got to be proprietary to the individual."
The explosion in social neuroscience has been driven by the tumbling cost of scanning equipment. Brain scans used to be the preserve of medical and clinical experiments, because they relied on complex, expensive technology such as positron emission tomography (PET), which was only available in a handful of places. PET scanners, which rely on radioactive tracer materials, cost about £3m to buy and a single scan can cost as much as £2,000. In contrast, a new fMRI machine costs about £1.5m, and each scan works out at about £400.
The fMRI machines are essentially giant, powerful magnets that are used to detect the tiny magnetic fields carried by the hydrogen atoms in water (or blood). They allow a very detailed 3D map of blood flow to be built up, and in the head, blood flow means busy neurons. Studying which regions of the brain need the most blood tells scientists where the most thinking is going on. (The original MRI technique is identical to that used by chemists called nuclear magnetic resonance but the name was changed as it was thought nobody would want to be scanned by a machine with "nuclear" in its title).
As the scanners shifted from being an expensive piece of kit for specialist neuroscientists to a practical tool for anyone with the will to work it, the scientific questions the technique was used to investigate snowballed.
Three years ago, scientists at University College London used fMRI to investigate the essence of love. They recruited people who confessed to being hopelessly in love with their partners and showed them a series of photographs of people they knew, one of which was their partner. Although brain activity was different in each individual, the researchers found that in every case, four specific regions of the brain lit up each time they saw the one they loved. The researchers announced that they had discovered the brain's common denominator of romantic love.
A year later, Joshua Greene and colleagues at Princeton University in New Jersey studied how people solved moral dilemmas. In one test, volunteers were scanned while they were asked whether they would push a person in front of a speeding train if it meant saving the lives of five others. Unsurprisingly perhaps, the question caused a flurry of activity in parts of the brain linked with emotion, leading the researchers to conclude that such moral quandaries may not be solved purely by logical reasoning, but also by emotional reactions.
The technique has also been used to delve into the murky question of how we judge people. Last year, Ray Dolan's team at the Institute of Cognitive Neuroscience in London used fMRI to see how people judged the trustworthiness of strangers. Volunteers were shown a series of faces and asked to judge whether the person was trustworthy. The researchers found that a region of the brain called the amygdala and two other parts of the brain flickered more intensely when people were shown the faces of people they thought would not be trustworthy.
But according to some scientists, such studies are the tip of the iceberg. The better fMRI systems become, and the more adept scientists get at extracting information from them, the more they will be able to piece together the neural circuits that make us who we are.
One emerging field is that of "neuro-economics". At the Center for Neuro-economics at Claremont Graduate University in California, Paul Zak is using fMRI to study how people assign value to certain products and make choices about what they buy. "If I ask you why you made a certain decision, you might not really be sure," he says. "But what if I can look directly into your brain and see how you reached that decision? That's what we want to be able to do."
Slowly, he says, researchers are homing in on the neural circuits that are activated when we make decisions - our likes and dislikes or, for example, how much different people value cigarettes over other items. Know that, and you can start feeding the data into policies such as how you tax products, says Zak. "If you know how much people value something, you can work out at what point a price hike will stop people buying it," he says.
Zak says fMRI stands to make a big impact in what has been dubbed "neuro-marketing". As an example of how fMRI might be used, Zak proposes a company that wants to increase its sales of milk. One way it might is to gather a group of people who like milk and scan them as they drink a glass. Some of the regions of the brain that buzz with activity might be triggered by any drink, but others may be triggered only by milk. Find other stimuli that trigger these regions of the brain and it could help you work out what it is that makes milk enjoyable, says Zak. Suppose objects from your childhood made those regions of your brain flicker. It might be that milk was evoking a sense of nostalgia, reminding you of when you got milk at school.
"If it turned out that milk was pleasurable to drink because it evokes memories of your childhood, you could market it as 'good when you were a kid, great when you're an adult'," he says. It's just an idea, and we're not there yet, but Zak says this is not pie-in-the-sky stuff. "A couple of years ago there was a lot of hostility to this kind of research, but now people are realising there's potential in it. Of course there will be a lot of crappy studies, but done properly, it allows us to get answers to questions we could never get before."
At Glimcher's lab in New York, progress is being made into understanding how the brain allow us to make certain decisions. Using fMRI scans and another technique that measures the activity of single neurons, Glimcher has recreated in a computer the neural programs that monkeys use to make decisions in a simple financial game. "Their behaviour is quite erratic and very similar to that of humans, but the program predicts what they will do to about 95% accuracy. It's spooky," he says. Ultimately, says Glimcher, neuroscientists should be able to use techniques like this to work out what a person will do in a specific situation, such as what he or she might buy when they walk into a shop.
At least one company, the BrightHouse Institute for Thought Sciences, in Atlanta, has been set up to exploit brain scans to inform marketing strategies. Instead of using focus groups, it is trying to use scans to tell companies what people think of their products and commercials.
Not everyone is convinced of the approach though. Donald Kennedy, the Stanford University-based editor of the journal Science and one of America's most eminent scientists, says: "You could just ask people what they think."
While Glimcher concedes that using brain scans to predict behaviour is a long way off, the progress is such that we should think about the implications, he says. "It raises serious philosophical questions, because it reduces us to a machine, but there's also a huge moral issue." Who should be allowed access to our brain scans, if they can reveal so much about us, he asks. "Within 10 years, we will need legislation that protects brain-scan information in the same way genetic information is protected," he says.
If using brain scans to predict specific behaviour is not on the cards, using them to judge if we will suffer from mental disease later in life is. Studies have shown that fMRI scans can be used to reveal early signs of multiple sclerosis and even go some way to predicting who might be most susceptible to dementias such as Alzheimer's. "For severe mental illness and dementias it is a serious proposition," says Sean Spence, a psychiatry researcher at Sheffield University. "There are changes in their brain before they begin to lose their memory. It's quite conceivable people could use that."
Stanford's Kennedy says it is the potential to use scans to predict people's health that is a concern. "I'm worried about fMRI scans being preserved after they have been taken," he says. "There's a push to prevent genetic information being used by companies for adverse selection, and at least equal protection should be given to brain scan data."
Glimcher says legislation banning access to people's brain scans should be drawn up to keep the data private before it's too late. "It's only a matter of time before the insurance companies come calling," he says. "It is going to happen and it's a big issue. It has to be dealt with soon."
'I feel as if I've been entombed'
It's easier to get your brain scanned than you might realise. Recent growth in research using fMRI machines means that there aren't enough willing brains to go around. That's why neuroscientist Andrea Mechelli, who works with fMRI at University College London, has had his own brain screened more than 30 times in the past five years. And that's how I found myself in Mechelli's lab on Monday morning, stripped of belt and metal jewellery, wearing earplugs and gigantic headphones, and about to be slid headfirst into the bowels of a gigantic magnet.
"This is for emergencies," Mechelli tells me, thrusting a squeezy plastic ball connected to a long tube into my hands. "Squeeze it if you have a panic attack or anything like that and I'll come in and get you out. Though I hope you don't." I hope so too, I tell him, through the metal cage that he swung down to cover my face.
Mechelli and his colleagues at UCL's Institute of Neurology are delighted to give me my scan, even at short notice, but there is a catch. With scanner time a valuable resource, tightly allocated to research projects, I first have to agree to be a bona-fide volunteer on one of their projects.
The project in question doesn't sound too bad, though. Mechelli is interested in the parts of the brain that hold and process information related to words and language. Specifically, he is seeking to understand why people who suffer brain injuries such as a stroke often find themselves unable to pronounce words they still know the meaning of, or vice versa, able to read words out loud with no idea of what they mean. To neuroscientists and linguists these separate functions are also important in understanding dyslexia. "I want to see what happens in the brain when people see objects and name them," Mechelli explains.
He is also seeking to confirm previous findings that thought processes relating to animals and man-made objects such as tools take place in two distinct places. Bizarrely, stroke and accident victims often find they can recall the names of animals such as rabbits and rhinos but not, say, a rolling pin.
After a final check that I don't suffer from claustrophobia (I don't) and that I'm not fitted with a pacemaker (I'm not), Mechelli searches me for metal objects that would interfere with the magnetic field - the photographer is not even allowed to carry her metal equipment into the room containing the scanner - and I lie down. It's important that I move my head as little as possible during the scan, so Mechelli wedges it between two vertical metal struts using small cushions. After fitting the headphones he will use to communicate with me inside the machine, he brings down the metal mask to cover my face. It's a tight fit - my nose pokes through two of the bars - and it's fitted with a rear-view mirror that Mechelli angles until I can see the screen fitted inside the machine behind me. And then he pushes the button that slides me head-first into the scanner.
I discover later that this is an older version of these machines, which means there is very little room between where my head ends and the giant magnet begins. It also means that my head, shoulders, chest and some of my lower abdomen must be placed inside the machine. The test itself takes 30 minutes, divided into two equal slots. In all I see 1,200 images or words, presented in rapid pairs, both of which I must name out loud (or rather whisper quietly, as this triggers less head movement).
With my head fixed in its upward position, the lights off in the room outside and just the flickering screen running through a series of black and white images for company, I feel as if I am entombed in my own personal silent movie theatre. Except it's noisy inside the machine. Very noisy. After the tests finish, Mechelli tells me through the headphones to close my eyes and sleep if I wish while the machine takes a few minutes to record a high-resolution structural scan. I don't think he's joking, but as I close my eyes an unholy clatter like a combination of the old Grandstand teleprinter and an air horn blowing every two seconds begins. It lasts for 11 minutes.
I'm a little disorientated when I emerge. "Do you want to see your brain?" Mechelli asks, pointing to a familiar looking image on a computer screen. This is a good sign - a sign that he thinks my brain is (probably) healthy - because very occasionally, the structural scans of volunteers in research projects do reveal tumours. And as a researcher, rather than a clinician, Mechelli is not allowed to break the bad news. "I tell them the scan hasn't worked properly and I can't show them it," he says. "And then I contact their GP and tell them to arrange a medical scan."
Of course Mechelli is not giving me an absolute all clear (the scan is not as detailed as a clinical one), but he does give me a mug and a £10 thank you for helping out, as well as a CD with rolling 3D electronic images of my brain on it. "Have another scan in 20 years and compare your brains," he suggests. "It's quite depressing." DA
To take part in the UCL neuroscience research call the imaging laboratory on 020-7833 7472, or go to www.fil.ion.ucl.ac.uk