The Berkeley Science Review: Articles

Zap!
Magnets trip up brain function. (view PDF)
by Colin Brown

The machine makes a loud clicking sound, and a powerful pulse of magnetic energy is zapped across the skull, scrambling the signals in the brain and changing how the subject thinks. At first glance, it might seem a bit like mad scientist technology or the plot of a James Bond film, but Transcranial Magnetic Stimulation (TMS) is very real, and is being used in hundreds of universities and psychiatric clinics around the globe. Far from a malevolent mind-control device, it is allowing neuroscientists to tease apart how the brain works with a level of safety, control, and precision never before possible. Originally developed in the mid-eighties by neurologists as a diagnostic tool for patients with nerve damage, it has since been developed into a powerful method for probing the wiring of the human brain. Unlike brain imaging technologies such as functional Magnetic Resonance Imaging (fMRI), TMS allows neuroscientists to directly manipulate, rather than simply observe, brain function. At UC Berkeley and elsewhere, TMS is being used to address basic questions about the neural basis of memory, attention, and movement, and it has shown promise as a treatment for a wide variety of disabling psychiatric disorders. But many questions still remain about the technology, not the least of which is: why exactly does it work?

A stimulating experience
According to neuroscience lore, the Italian physician Luigi Galvani first discovered that nerves transmit electrical signals in 1783 when his assistant touched a charged scalpel to an exposed nerve in a dissected set of frogs legs, causing them to twitch. TMS operates on a similar principle, but instead of applying a current to neurons directly, it uses a magnetic field to generate electrical flow in the cells by electromagnetic induction, the same principle used in electrical generators. In the simplest application, a single pulse is applied over a part of the brain for a few milliseconds, causing nerve cells in the region to shoot off randomly. The subject might experience a muscle twitch or see a brief flash of light, depending on the location of the pulse.

However, when a pulse is applied over a region involved in more complicated cognitive processes, the haphazard neuronal firing induced by TMS can scramble the carefully orchestrated sets of signals present during normal function. The result is a very brief, targeted disruption of information flow through that region. If the area targeted by the pulse is involved in a particular cognitive task, this disruption will manifest itself through subtle behavioral effects, such as an increased error rate when reading letters from a screen. By pairing behavioral measures with carefully timed TMS pulses, researchers can establish a causal role for a region in a particular task.

"Essentially, you're introducing transient noise to a region, and if that region is critically involved at that time, it's not going to perform that function as efficiently", said TMS researcher Brian Miller, a recent graduate from Mark D'Esposito's lab at Berkeley.

The disruption of a region after a single TMS pulse lasts for a very short time, usually just a fraction of a second after the pulse is applied. While even these short pulses can be useful for some applications, a far more common and powerful technique is repetetive TMS, or rTMS, where a train of pulses is applied over several minutes. This leads to more extended "offline" disruptions that can last as long as half an hour after the treatment. Depending on where the pulse is applied, the effects of rTMS treatments can mimic those seen in so-called "lesion" patients who have experienced damage to the region from stroke or other brain injuries. Since the effects of rTMS are controllable and fade after a few minutes, researchers can create temporary "virtual lesions" to disable almost any brain region at will. Examining real lesion patients can yield profound insights into the function of healthy brains, but the size and severity of the damage can vary widely among individuals, and the reorganization that takes place as the brain heals itself following an injury can obscure its effects. According to Miller, "with lesions you have no control over the location or the extent of the damage. TMS gives you that very focal control over the region of disruption. You can take someone who is otherwise totally normal, and very transiently disrupt the region you're interested in."

Virtual lesions
Miller's work used this rTMS lesioning approach to study a brain process called "refreshing", which allows the brain to pick out interesting bits of information from the swamp of sensory data that passes through its visual and auditory systems. When Miller and his colleagues used fMRI, which maps regions of neural activation, to scan the brains of people during a task that involved refreshing, they ran into a problem—they found two different regions, one in the prefrontal cortex and one in the parietal lobe, that lit up almost simultaneously.

"The key complement [TMS] makes to techniques like fMRI," said Miller, "is that you can really get at the causality. You can see different areas are active in an fMRI scan, but you can't really say that those regions are necessary for a cognitive function."

Using rTMS like a temporary scalpel, they were able to focus on each region individually. When they disabled the prefrontal cortex, subjects showed significant defects in refreshing, but when they disabled the parietal region the subjects were essentially normal. The conclusion was that the prefrontal cortex was not only involved in refreshing, but was actually the source for the refreshing signals.

Another UC Berkeley neuroscientist, psychology professor Lynn Robertson, used rTMS to examine how people with a perceptual quirk called synesthesia process visual information. These individuals see certain letters or numbers as always having a specific color. The letter A, for instance, might always appear as purple, even when printed on a black-and-white newspaper page. While the causes of synesthesia are complex, Robertson suspected that it might have something to do with a brain process called "binding." The first thing the brain does when receiving a raw image from the eyes is break it down into its most basic features, such as motion, color, and shape. Before binding occurs, a basketball flying toward the rim is just an unassembled medley of concepts: orange, round, movement. When we focus attention on the ball, the brain's binding circuits reassemble these separate parts to form a complete object (the ball) at a specific position in the visual field. Robertson suspected that synesthesia results from a permanent, incorrect binding of a color that is not present in a visual field to letters or numbers that are. To test this, she and her team used rTMS on two synaesthesic individuals to disable a part of the brain known to be critical for binding in normal people. Remarkably, this caused them to temporarily lose their synaesthesic associations, confirming that normal binding and synesthesia share a common mechanism in the brain.

The excitement of competition
Inducing lesions in healthy patients, however, is only a part of what TMS can do. By applying a brief TMS pulse over a region of the brain called the motor cortex, researchers can trigger electrical signals that travel down the nerves of the spinal cord and produce a brief muscle twitch, called a motor evoked potential. By measuring the strength of these twitches in various muscles as pulses are applied over different parts of the cortex, researchers can generate maps of the complex wiring that connects the brain to the body. Although these connections are generally hard-wired, the regions of the motor cortex that connect to a particular muscle can change in both size and "excitability"—how likely they are to trigger a movement—over time. For example, the area in the cortex corresponding to the hands often becomes larger in musicians or athletes who spend a lot of time practicing fine-scale movements. Researchers can map these changes by applying TMS pulses over a region and measuring the strength of the muscle twitches that result—the stronger the twitch, the greater the excitability of the stimulated region.

This mapping technique is useful in clinical settings, for example when doctors want to know the extent of nerve damage in stroke or Parkinson's patients. It is also used by cognitive neuroscientists like Professor Rich Ivry in the psychology department to study how the motor cortex is involved in conceptualizing movements. When we think about reaching for a coffee mug or we watch Tiger Woods swing a golf club on TV, the same parts of the brain that would be involved in executing those movements light up with neural activity. However, since we don't jump up and start swinging every time Tiger does, some process must block the signals that are generated when we are only thinking of a movement rather than actually doing it. Researchers in Ivry's lab were able to use the strength of muscle twitches generated by TMS to measure the extent of this blockage. When a subject thought about making a movement, they found that the strength of the twitches generated by regions involved in that movement went down. Based on fMRI, they suspected that a particular region of the brain's prefrontal cortex might be involved in generating the blocking signal. When they measured muscle twitches in people who had had this region disrupted by rTMS, they found that the blockage was lifted.

Ivry's lab has also investigated how the brain selects whether to use the left or right hand for tasks where a choice is possible. For instance, what if that coffee mug we're reaching for is directly in the middle of the table? "It seems to us like it's an automatic, effortless choice," says Ivry, "but really there's a sort of competition there that gets resolved."

Ivry and his lab peered into this battle by using TMS to measure excitability in the right and left cortices as subjects were asked to plan movements using one or both of their hands. The hypothesis was that the side responsible for controlling the selected hand would stifle the other. But surprisingly, they found more inhibition in the side involved in the actual movement.

An ongoing project in Ivry's lab is even trying to use rTMS to influence how subjects make these left-versus-right choices. By creating a lesion in the region containing information about the right half of a subject's visual space, the researchers are able to skew their preference toward the left.

"We think we're disrupting the representation of where the object is [in space] in the left hemisphere, which is needed for a person to plan a right-hand movement," said Ivry.

How does it work?
Although TMS is growing rapidly in popularity, both in basic research and in clinical settings, there are still many open questions about exactly how it works, and why it has the effects it does.

"TMS, in the grander scheme of things, is still a young technology," said TMS researcher Brian Miller. "People are still trying to figure out exactly what it's doing to the underlying physiology."

In some sense, the greatest strength of TMS—its ability to easily and reversibly affect brain function in human subjects—has also been a barrier to understanding more about its basic mechanisms. Since most studies have focused on humans, there has been little drive to do experiments in animal models that allow much closer inspection of the physiological, neural, and molecular changes that occur following a TMS pulse.

As the uses of TMS in both the clinic and lab have become more complicated, the need for better information about the specific physiological effects of TMS has become even more pressing. Brian Pasley and Elena Allen, graduate students in Ralph Freeman's lab in the UC Berkeley School of Optometry, began to clarify some of these issues in a study published in the journal Science in 2007. One fundamental question in the field is whether the alterations in brain function induced by TMS can be measured directly by imaging technologies like PET scans and fMRI, which detect changes in oxygen usage and blood flow. If this is the case, it would allow researchers to know not just the location where a pulse was applied, but also the specific effect it had on the network of nerve cells in that region. Pasley and Allen used anaesthetized cats as an animal model, and simultaneously measured electrical activity, blood flow, and oxygen levels in the animals' visual centers as rTMS was applied at several different frequencies. They found that all three measures were highly correlated, confirming what most in the field had suspected and opening up the possibility of combining powerful brain imaging methods with TMS.

Pasley and Allen also found that the physiological effects of the rTMS pulse lasted longer than expected in comparison to the known behavioral effects. They saw cellular disruptions lasting several minutes after a pulse of just a few seconds. In addition, they found peculiarities in the measured input and output of brain regions affected by TMS, pointing the way for others to delve deeper into its cellular mechanisms.

TMS in the clinic
The ability of rTMS to noninvasively alter brain function with few apparent side effects makes it particularly appealing as a treatment for psychiatric disorders. By far the most successful example of this strategy is for depression, especially in patients who don't respond well to antidepressant drugs. TMS may even be able to take the place of electroconvulsive treatments (commonly known as "shock therapy") for some. Most of these treatments try to correct imbalances between the left and right halves of the brain's prefrontal cortex. Functional MRI studies have shown that the right half of this region is often hyperactive in depressed patients, while the left half often shows decreased activity. In order to produce an elevated mood in depressed patients, rTMS treatments try to either impair function of the right side by using low frequency pulses, or boost function of the left side using higher frequency pulses that can in some cases excite neural function. The effects of stimulation applied to the cortex can then spread to other brain regions also involved in mood regulation through neurons connecting them to the cortex. How the short-term changes induced by TMS treatment translate into a lasting elevation of mood is still unclear, however.

Although the antidepressant properties of TMS have been a topic of research since the 1990s, its widespread use as a therapy was hampered by questions about exactly how it works and how effective it actually is. A major problem has been the difficulty in creating properly controlled and blinded trials. The coils used for TMS produce a loud click and a noticeable tingling sensation when a pulse is applied, both of which are difficult to mimic in a double blind trial where both the patient and the researcher can't know whether a placebo is used. This problem was so pervasive that special "sham" coils had to be designed before proper trials could be conducted. Also, many early trials showed inconsistent results, likely due to problems with targeting the proper region of the brain and differences in the effectiveness of the treatment among patients with different types of depression. As more trials were performed, however, a consistently beneficial effect became apparent, and last fall the FDA finally approved a TMS device, marketed by the company Neuronetics, for general use as an antidepressant.

The therapeutic usefulness of TMS for other psychiatric disorders, however, is still unclear. Since TMS is relatively easy to apply and has few side effects, researchers have tried using it to treat all types of disease, including mania, obsessive-compulsive disorder, post-traumatic stress disorder, and schizophrenia. Results from most of these trials have been mixed, however, and it is likely that a deeper understanding of both the neural basis of these disorders and the physiological mechanisms of TMS will be needed before it is widely used to treat these conditions.

A better brain by magnetization
One of the most intriguing possible uses for TMS is as a tool for neuroenhancement, that is, boosting mental function in otherwise healthy individuals. In fact, a small number of studies have already claimed substantial improvements in cognitive function using TMS. The best known of these, and also one of the most controversial, was conducted by an Australian neuroscientist named Allan Snyder. By applying rTMS to a region of the prefrontal cortex, Snyder claimed to actually induce enhanced artistic abilities in otherwise normal subjects (including a New York Times reporter), essentially turning them into what he referred to as "savants" for the duration of the treatment. While many in the field are skeptical of these particular results, the thorny ethical implications raised by the possibility of using TMS for neuroenhancement are still there.

In fact, many researchers who use TMS in the lab are cautiously optimistic about the possibility of enhancement. Some have pointed out that there is little physiological difference between the changes in the brain resulting from TMS and those resulting from more "natural" methods such as learning. In a 2003 article, TMS researcher Professor Alvaro Pascual-Leone argued that "all environmental interventions, and certainly educational approaches, represent interventions that mold the brain of the actor. Given this perspective, it is conceivable that neuromodulation with properly controlled and carefully applied neurophysiological methods could be potentially a safer, more effective, and more efficient means of guiding plasticity and shaping behavior." The UC Berkeley researchers consulted for this story agreed that studies of neuroenhancement should be pursued, if for no other reason than to allow these techniques to be applied safely. "We shouldn't close the door out of hand," says Ivry. "If these things do turn out to have some truth to them, people are going to use them. Science's best guide here is to make sure that they're properly evaluated."

Where does it go from here?
The UC Berkeley researchers agree that the future of TMS as a tool for cognitive neuroscience seems bright. New advances in coil design are allowing TMS to be joined with imaging technologies like fMRI to give researchers unprecedented access to the inner workings of the brain. Special ceramic coils can now even be used while a subject is inside the giant magnets of an fMRI scanner, to image the changes induced by TMS in nearly real time. This should allow for better targeting of TMS pulses, and also opens up entirely new types of experiments for researchers interested in understanding the connections between the brain's various processing centers and subnetworks. By applying a pulse to activate one structure using TMS and then using fMRI to follow the tracks of activation through the brain, researchers hope to be able to construct incredibly detailed maps of entire networks of neurons. Better coils may also allow disruptions to be generated deeper inside the brain, allowing researchers and clinicians to study structures controlling memory and emotion.

Much of the future of the technology, especially in the clinical arena, may hinge on how well scientists are able to understand its fundamental mechanisms. An improved understanding of the principles of TMS, combined with better knowledge of how treatment outcomes vary among patients, may allow doctors to predict a specific treatment regimen's effects on a given person. This could reveal an unprecedented array of treatment options for people suffering from psychiatric disease.

Colin Brown is a graduate student in molecular and cell biology.

Comments on this article? Drop us a line at with 'letter to the editor' in the subject!