The Tools of Neuroscience

Speakers:
Marcus E. Raichle, Washington University in St. Louis School of Medicine
Miguel Angelo Laporta Nicolelis, Duke University

Highlights

• Conventional neuroimaging studies focus on changes in brain activity as subjects perform specific tasks, but often ignore underlying, intrinsic activity that accounts for the vast majority of the brain's energy usage.
• The "noise" that most neuroscientists subtract when they analyze brain images may actually contain important information.
• New, computational brain-machine interfaces can process brain signals to enable monkeys to operate robotic arms and legs.
• Preliminary tests of similar interfaces in humans suggest that such technologies might one day be used to create prostheses capable of restoring motor abilities to people with disabling neurological injuries and diseases.

Enough about you

In recent years, the field of neuroscience, driven by a bumper crop of powerful new imaging technologies, has risen to a level of cultural prominence few scientific endeavors achieve. Newscasters and editors of popular scientific magazines never tire of stories about scientists diagnosing mental illnesses definitively, confirming or refuting some long-held cultural belief, or uncovering a new strategy to treat widespread neurological diseases such as Alzheimer's or Parkinson's. The beautiful full-color brain scans that inevitably accompany these discoveries are one reason for their popularity—they look great on a magazine page—but perhaps the biggest reason for neuroscience's appeal is simple human egotism.

"When I wrote for Discover magazine, we discovered ... that readers loved learning about the brain, but what they really loved most was learning about their brain," says Eric Haseltine, who organized the appropriately titled "Your Brain and Yourself" track at the 2007 Aspen Ideas Festival in Aspen, Colorado. Synthetic chemistry may drive the development of new drugs, and nuclear physics may revolutionize electronic engineering, but neither one helps us talk about ourselves.

Attendees of this conference track not only learned new ways to think about thinking, they also heard many of the world's top neuroscientists ponder where these thoughts are leading us.

Tuning in the noise

Marcus Raichle, one of the inventors of modern brain imaging, began his presentation with a primer on the two main techniques that now dominate the field: positron emission tomography (PET) and functional magnetic resonance imaging (fMRI).

In PET, subjects are injected with a harmless radiolabeled dye, usually fluorodeoxyglucose, and a scanning system detects its uptake in the brain. MRI, in contrast, uses the magnetic resonance of the molecules already in the brain to paint a picture of blood flow to different areas. As blood flow and glucose utilization change in different brain regions, the amount of oxygen fluctuates accordingly. This blood oxygen level dependent (BOLD) change underlies most modern fMRI studies.

The raw data from a typical PET or fMRI study produce images showing lots of activity throughout the brain. To focus on the most important areas, researchers superimpose scans from a subject performing a particular task over scans of the same subject at rest, and then use software to subtract the "noise" and highlight only the areas that differed.

Identity may lie in the brain's background noise.

While this focus on specific areas has produced important breakthroughs, Raichle argues that the "noise" researchers are filtering out may actually contain important information. "The underlying issue here is that most of what you're looking at is a teeny fraction of what's there," says Raichle.

Drawing an analogy with astronomy, Raichle explains that new neuroimaging techniques operate much like the Hubble space telescope, looking at an enormous field of information from which researchers can study only isolated fragments. Just as the universe appears to be full of "dark energy," whose existence astronomers can only intuit, the brain's activity appears to be dominated by yet-unseen processes.

The brain makes up just 2% of an adult's body weight, but consumes 20% of the body's energy, an equation that wouldn't make much evolutionary sense if only a few patches of the cerebrum need to be active at any time. To explore this conundrum, Raichle and a few other neuroscientists have been looking beyond the evoked activity of the brain and into its intrinsic activity—they're listening to the noise.

What they're hearing so far is tantalizing. In one experiment, for example, researchers looked at the brain areas that show decreased activity in subjects who are focusing so intently on a task that they have become "lost in their work." The resulting images highlight areas involved in memory, emotional salience, and self-referential judgments, suggesting that the intrinsic activity of these areas in resting subjects form an important part of their identities.

In another experiment, scientists found ongoing fluctuations in the BOLD signal in patients who were lying still. A closer look revealed that this seemingly random noise was actually synchronized between the two sides of the brain. "So the background ... for how this system is working is already in synch, it isn't waiting to be turned on," says Raichle, adding that in general, "the brain isn't activated and turned on—it's modulated, and it's always on."

Monkey think, monkey do

Duke University's Miguel Nicolelis explained how he and his colleagues have begun to understand some of the brain's modulations. The team's principal goal is to learn how to decode specific types of brain signals, such as those that drive limb movement, and use that information to develop bionic prostheses for amputees and paralyzed patients. It sounds like science fiction, but the researchers seem very close to achieving it.

The business end of Nicolelis's technique is an array of tiny wire filaments that can be implanted in an animal or human brain. Each filament samples a single cell, and the arrays can record the activity of 300–500 cells in a monkey, or 50 cells in a human. Sending the signals into an audio circuit and a pair of speakers, the scientists can "listen to a neural orchestra of a conscious human brain that is talking to you," says Nicolelis.

After enjoying a few neural symphonies, the investigators fed the signals from the motor control area of a monkey's brain into a digital signal processing algorithm, and began analyzing the data while the animal played a simple video game. In the game, the monkey uses a joystick to move a cursor and hit a target that appears in random locations on a video screen. Each successful hit yields a fruit juice reward, and the monkey appears to enjoy playing. "She can stay in this game for hours and drink the entire allotment of liquid [for the day]," says Nicolelis.

A monkey can control a robot arm using only its brain.

Meanwhile, a cluster of computers processes the monkey's brain signals, eventually developing an optimal model of the motor signals controlling the joystick. Having reverse-engineered the appropriate hardware driver, the researchers moved the monkey's joystick to another room, where a robotic arm controlled the stick in response to the monkey's brain signals. After a short learning period, the monkey began playing the game successfully by controlling the robotic arm with her brain. "She ... realized that she had to win by thinking, something that doesn't happen very often in our society," says Nicolelis.

Interestingly, muscle monitors reveal that the monkey's actual arm no longer makes joystick-like movements when the animal is using its mind to control the cursor. Indeed, in video footage from the experiment, the monkey appears to use both hands for scratching, gesturing, and other unrelated tasks while controlling the game mentally. In effect, the monkey has acquired a third, robotic arm.

The team is now working on decoding the motor signals that drive a more complex behavior: walking. With the brain signals of a monkey walking on a treadmill in their laboratory in North Carolina, the investigators can drive a bipedal robot on a treadmill in Kyoto, via the Internet. Nicolelis says the walking algorithm is currently about 90% accurate.

With some modifications, a similar system could eventually operate prosthetic limbs in humans, who appear to use similar motor control signals. In one recent study, for example, the researchers found that conscious patients undergoing brain surgery can play a video game similar to the one the monkey played, using nothing but brain signals fed through the same algorithm that worked in the monkey study.

Neuroscience and the Law

Speakers:
Jeffrey Rosen, George Washington University
Michael Gazzaniga, University of California, Santa Barbara

Highlights

• Courts are already using neuroscience evidence in the courtroom, setting disturbing precedents.
• Much of the current neuroscientific evidence being used in courts probably doesn't belong there.
• Findings from neuroscience could have implications for our understanding of volition, free will, and how we understand personal responsibility.
• Some have postulated that the tools of neuroscience could have potential applications in lie detection and in predicting the likelihood of criminal behavior in specific individuals.
• Multidisciplinary efforts are underway to assess what current neuroscience theory tells us, and how it might properly inform legal practice.

Thought crimes and misdemeanors

While cautions like Andreasen's were scattered throughout the conference presentations, the two speakers who began and concluded the neuroscience-oriented track focused on a much more ominous and imminent challenge. The word "neurolaw" remains on the periphery of public consciousness, but a disturbing series of recent scientific and legal developments could soon change that.

Legal scholar Jeffrey Rosen set the tone during a well-attended breakfast session. According to Rosen, as neuroscientists discover more about the formation of intent and the brain's processing of higher concepts, such as morality, lawyers and courts have taken a keen interest.

He cites the 2005 U.S. Supreme Court decision in Roper v. Simmons as a watershed moment. In that case, a majority of justices overturned a death penalty conviction of a minor, partly based on research showing that the human brain is not fully myelinated before age 18, and is therefore incapable of fully adult morality. In another development, several recent neuroscience experiments have shown that a person's intent to do something actually forms before the subject is even aware of it, undermining the fundamental concept of free will. "Maybe we will have to re-examine the whole idea that we're intentional, thinking human beings ... rather than automatons who are driven here by forces beyond our control," says Rosen.

We make decisions before we understand them.

Michael Gazzaniga, delivering the final talk of the conference, echoed the same theme. He cites several studies, including one in which researchers were able to identify a subject's intention to choose a particular response about 300 milliseconds before the patient actually responded. "We're running sort of on tape delay with what the brain is actually doing," says Gazzaniga, adding that "by the time you know something consciously, your brain has already done the work."

In Gazzaniga's own work, so-called split-brain patients—people with epilepsy who undergo a procedure to sever the connections between the two halves of their brains to prevent seizures—reveal an even more disturbing phenomenon. The patients can match pictures of different objects correctly when using either eye, but the stories they concoct to explain their choices differ radically.

In one experiment, a patient looked at a picture of a chicken on his right-hand side, and a picture of a house covered in show on his left. He was then presented with a series of additional pictures to match with what he was seeing. His right hand pointed to a chicken foot, and his left hand to a shovel. The choices made sense, but the explanation was somewhat surprising.

"His left hemisphere, that's doing the talking, says 'the chicken goes with the chicken claw,' and then he's looking down at his left hand, and he says, 'and you need a shovel to clean out the chicken shit.'" The right hemisphere had in fact seen and processed the snow scene, and presumably matched it with the snow shovel, but the patient could not articulate the logic of that choice because the two sides of his brain were not connected. He automatically concocted a different explanation.

The amusing response conceals a thorny problem: the brain appears to be making a decision unconsciously, then inventing a rationale for it after the fact. If we can't control our decisions, should the law hold us accountable for them? Throughout their talks, Rosen and Gazzaniga explored multiple facets of this issue, ranging from the possibility of exonerating defendants on the basis of minor brain abnormalities, to using neuroimaging technologies as lie-detection tools, to the spectre of squads of "thought police" preventively detaining citizens whose brain scans reveal guilty knowledge or predelictions for crime.

Will we jail suspects because of their thoughts?

Pointing out that historically, American courts have abetted such ethical disasters as McCarthyism and eugenics, Rosen issued a call for action. "When I look at the history of the privacy debates over the course of the 19th and 20th century, I'm struck by how little salvation has come from the courts and how much from the political system," he says.

Gazzaniga not only concurs, but is also taking concrete steps to push the debate forward. In a new initiative sponsored by the MacArthur Foundation, which will be launched formally in October, neuroscientists like Gazzaniga will collaborate with prominent judges, lawyers, and ethicists to place neurolaw on a solid philosophical base.

Despite the ominous possibilities of neurolaw, conference attendees were visibly excited about the promise of neuroscience, and continued discussing the presentations over their lunches and dinners. Whether the golden age of neuroscience leads to medical breakthroughs, social disruptions, or both, one thing is certain: we won't soon tire of talking about it.