SAM can’t quite put it into words. It’s not relaxation, exactly, or
concentration. He wouldn’t call it meditation either, though it requires
conscious effort and he sits still while he’s doing it. It’s just a state he
puts himself into whenever he senses the aura that heralds an epileptic fit. And
it seems to work. Ever since he learned to do it, he’s suffered fewer
seizures.
It took practice, though. The doctors began by sticking electrodes to his
head and feeding the signals into a computer. Then they asked him to play a
computer game where he had to raise a rocket ship to the top of a screen by the
power of thought alone. He doesn’t quite know how, but eventually he learned
what it felt like to keep the ship afloat. Now, even without his computer, he
can recreate the feeling at will, and it somehow seems to prevent the massive
discharge of electricity through his brain that accompanies a seizure.
The treatment is called neurofeedback, or neurotherapy, and it first became
popular in the 1960s when biofeedback was all the rage. It came to be seen as a
slightly wacky relaxation technique, and its unfortunate association with
mind-expanding drugs and mysticism earned it a bad reputation in the medical
community. Nonetheless it was credited with helping dozens of patients with
epilepsy, and adapted for all sorts of other conditions. Hundreds of US clinics
routinely offer neurofeedback to treat attention deficit hyperactivity disorder
(ADHD) as well as alcohol and drug addictions and post-traumatic stress
disorder—with little evidence that it works.
Advertisement
But despite the controversy, some researchers have stuck with it, insiting on
more rigorous testing. And their persistence may be paying off, as recently
there has been a steady trickle of promising findings, from thought exercises
for improving musicians’ performances to suggestions that we might one day be
able to train away some of the effects of schizophrenia. “A great deal more is
claimed than is credible,” says Peter Rosenfeld from Northwestern University in
Evanston, Illinois. But as methods of measuring brain activity patterns improve,
it’s hoped that many more conditions will soon be amenable to neurofeedback.
The idea behind neurofeedback is that certain brain states, such as
particular emotions or disorders, generate distinctive patterns of electrical
activity in the brain. These complex patterns of brain waves can be displayed in
an electroencephalogram, or EEG, produced from a set of electrodes attached to
the scalp. Scientists think that abnormal rhythms or patterns reflect
abnormalities in the underlying interactions between brain cells, and that if
you could correct the wave pattern you should be able to correct or somehow
compensate for symptoms. The trick is knowing how to modify the wave
pattern.
The theory is that if you can see some kind of representation of your own
brain activity, you should be able to learn to modify it. So neurofeedback
training involves filtering the abnormal signal out of the EEG using specialised
software, and transforming it into, say, a cursor or a rocket ship on a screen.
Pleasant lights, sounds or points reward the patient who sends it in the right
direction (see Diagram).
The earliest forms of neurofeedback were techniques for boosting the
alpha-wave component of the EEG—a signal associated with drowsiness.
Supporters claim these methods have honed people’s sporting skills, and even
helped Olympic champions, although many researchers are sceptical that
neurofeedback had much to do with it. The first person to explore the potential
of neurofeedback as a treatment was a psychologist at the Veterans
Administration Hospital in Sepulveda, California, who serendipitously discovered
a link between epilepsy and a component of the EEG he called the sensorimotor
rhythm (SMR).
Barry Sterman, who is now at the School of Medicine at the University of
California, Los Angeles, was asked by the US Air Force to look into the
possibility that a rocket fuel then in use could cause seizures. He designed
some experiments where he exposed cats to the fuel. By chance, some of the cats
had been involved in an earlier, unrelated series of experiments to investigate
the threshold between sleep and wakefulness, in which they had been trained with
food rewards to adopt a special drowsy state. It turns out that in this state
the level of SMR in the EEG is always high.
What was interesting for Sterman was that these cats were also highly
resistant to seizures when they encountered the fuel, and he managed to
attribute the resistance to the high level of SMR. In the early 1970s Sterman
went on to show that patients with epilepsy could also reduce their risk of
having a seizure by learning to boost the SMR component.
But technique became controversial almost immediately. On the basis of very
little controlled testing, people began to use it to treat other conditions.
Joel Lubar, a psychologist at the University of Tennessee in Knoxville, adapted
the idea for children who were hyperactive. He showed that using neurofeedback
to suppress the SMR and other, low-frequency brain signals seemed to help
children with the hyperactive component of ADHD, while boosting higher
frequencies improved their concentration. The effects can last ten years or
longer, claims Lubar, and between 40 and 60 per cent of patients who have tried
neurofeedback can reduce their dose of the controversial drug Ritalin, or even
stop taking it completely.
The problem, say some researchers, is that the treatment of ADHD has become
the biggest single clinical application of neurofeedback today, but without
proper placebo-controlled trials. “My personal opinion is that proof is still
lacking that neurofeedback is efficient for hyperactive children,” says Niels
Birbaumer from the University of Tübingen in Germany. “But it’s widely sold
in the United States, and that bothers me.”
Tests of another form of neurofeedback began in the 1980s, when Rosenfeld,
working with Robert Dowman also at Northwestern University, showed that the
technique could be used to manipulate pain thresholds in rats and in people.
Instead of looking at the brain’s spontaneous activity pattern, as others were
doing, they set out to mould the bursts of activity produced by specific brain
cells in response to stimulation.
They pinpointed cells in the rat’s brains, known as somatosensory cells.
These cells usually sense touch, but as the signal gets larger, the sensation
turns to pain. The researchers stimulated these cells with implanted electrodes
to test this pain threshold. Then they rewarded the animals with food when they
did things that modified the responses of the cells, and showed that it was
possible to increase or decrease the rats’ sensitivity to pain. The rats were
essentially performing the rocket ship trick, but being rewarded for achieving
the right “feeling”.
While the researchers turned their attention to altering pain thresholds in
people, they spotted another potential application for neurofeedback. Richard
Davidson of the University of Wisconsin had discovered an asymmetry in the
activity of the left and right frontal lobes of depressed patients that was
distinctively different from the normal pattern. This abnormal asymmetry, as
measured by the ratio of alpha activity, could predict a person’s emotional
responses to happy and sad film clips, and fluctuated systematically in response
to reward or punishment.
Rosenfeld realised that here was something he could work with. By applying
neurofeedback techniques, he might be able to correct the waveform, and perhaps
even alleviate the symptoms. Sure enough, by training patients with a
neurofeedback regime that rewarded them with a pleasant tone when the ratio of
activity on the two sides became normal, Rosenfeld and his colleagues Elsa and
Rufus Baehr found their patients’ depression lifted.
In the most dramatic case a woman had suffered from periodic bouts of
depression and had been treated unsuccessfully with antidepressants and
psychotherapy for 12 years, as well as other types of neurofeedback. But she
recovered after 34 hour-long sessions of the asymmetry training, combined with
psychotherapy. And when the researchers followed her up this year, six years
after the original sessions, she had suffered no more bouts of depression. Of
course she might have recovered naturally, says Rosenfeld. Twenty patients have
tried the treatment so far, and though 18 improved two did not. But, he warns,
he has yet to carry out a proper controlled study.
Birbaumer firmly agrees with Rosenfeld that controlled trials are the only
way to show whether the benefits of neurofeedback are real. He has spent years
developing ways to test the technique for epilepsy.
In one study he compared a group of people who got real feedback from their
brain signals with a group given totally random feedback. But he found that it
didn’t take long for a person suffering from severe epilepsy to work out if
their symptoms were improving or not. As the training proceeded, it became
increasingly difficult to fool the human guinea pigs as to who was receiving the
experimental treatment and who the equivalent of a sugar pill. Although this
seems promising evidence that neurofeedback works, he never finished the study
because his control group wouldn’t continue.
Birbaumer and his colleagues tried to get round the problem by comparing the
effects of neurofeedback with another form of biofeedback. He taught one group
of people with epilepsy to manipulate a particular EEG component called the slow
cortical potential (SCP). This is thought to be a gauge of how excitable
cortical neurons are—how likely they are to be triggered into action. Just
before a seizure, large SCPs can be seen near the focus of the epilepsy, so by
using neurofeedback to teach patients to stabilise their SCPs, the researchers
hoped to stave off seizures.
To provide some sort of placebo control, they also studied a biofeedback
group who learned to avoid hyperventilation—a known risk factor for
seizures—by controlling the carbon dioxide content of their exhaled
breath. Both groups were trained for the same length of time, and a third
received new anti-convulsive drugs. While the groups on medication and
neurofeedback had significantly fewer seizures, the respiration group did not.
Birbaumer’s studies suggest that two-thirds of patients respond to SCP
neurofeedback with a reduction in seizures that compares favourably with the
latest drugs.
Meanwhile John Gruzelier of Imperial College Medical School in London is
beginning to think about applying neurofeedback to schizophrenia. One theory
about the condition is that there may be some abnormal asymmetry in the way the
brain develops, and, just as Davidson found with depressed patients, Gruzelier
reasoned that this abnormal asymmetry should show up in the EEG. If he and his
colleagues could find this pattern, then maybe they could use neurofeedback
techniques to modify it.
In the past, people with schizophrenia were considered poor candidates for
neurofeedback because they were not thought to be strongly motivated enough to
concentrate on the task. But now Gruzelier, working with Birbaumer, has shown
that patients with schizophrenia can concentrate hard enough to shape their
brain waves—though not yet in a manner directly relevant to their
symptoms. In December 1999 in the International Journal of
Psychophysiology he showed that patients with schizophrenia can regulate
their own SCP patterns and equalise them between the two hemispheres.
In another study with Tobias Egner, Gruzelier has wandered into territory
that even the neurofeedback fans of the 1960s would have found incredible. In
preliminary findings last year, the pair said that enhancing the SMR—the
component that is prominent in drowsy states—could improve musicians’
performances. Using a method similar to that used for hyperactive children
seemed to reduce the number of impulsive mistakes they made (New
Scientist, 16 September 2000, p 19).
Other than the “alpha” claims, this was the first time researchers had
demonstrated positive effects of neurofeedback in healthy people—in this
case students at the Royal College of Music in London. Since then, sponsored by
the Leverhulme Trust, Egner has collected more data indicating that boosting the
SMR can improve attention in healthy people too.
This opens up a huge range of exciting possibilities for the future of
neurofeedback. Egner says that SMR enhancement might help people who need a
calm, steady hand, or perhaps better impulse control. But although he would love
to see the technique used to help reduce impulsive aggression in sociopathic
prison inmates, he fears that in the near future it is much more likely to be
used to try to reduce somebody’s golf handicap. Whether its future lies in
treating illness and enhancing the skills of healthy people, or whether it
remains largely a gimmick, will depend on researchers finding successful ways to
test it, with proper placebo-controlled studies.
But if it seems to work, does it really matter if the effect is purely
placebo? “One wants to know what is happening, at the level of mechanism, so as
to improve the treatment,” says Rosenfeld. If there is a genuine mechanism
involved, then there may be better ways to activate it—direct stimulation
of the brain, perhaps. “If it is placebo,” he adds, “maybe we can design more
impressive-looking machines to enhance the effect.”