Controlling computers by thought

Toby Howard

This article first appeared in Personal Computer World magazine, February 1999.

UNTIL RECENTLY, controlling computers by human thought was science fiction, but it's rapidly becoming science fact. Now researchers have succeeded in tapping directly into thoughts, by implanting tiny electrodes into the brain. It's called cognitive engineering, and it's mind-blowing.

Attempts to use thought-power alone to control computers have traditionally used an electroencephalograph (EEG), to measure the brain's electrical activity. The patient wears a skull-cap studded with electrodes. Because the skull muffles much of the neuronal chatter, the EEG records only large-scale activity, when large groups of neurons fire together. It's rather like listening to the neighbours through the wall -- you get the general gist, but you can't hear all you'd like to. And this is the crux of the brain-computer interface problem -- how can individual thoughts be picked up? Much research focuses on signal processing techniques to search the EEG traces for subtle changes in the cranial cacophany, but it's a far from exact science.

Then, just a few weeks ago, came the startling announcement of a huge leap forward. Neurosurgeon Roy Bakay and his team at Emory University in Atlanta, Georgia, have developed a brain implant that can monitor extremely small-scale activity in the brain's motor area.

The implant is a hollow glass cone the size of a ballpoint pen's tip. Called a "neurotrophic electrode", it's inserted through a hole drilled in the skull, into the cerebral cortex just above the ear. The placement is crucial. Bakay's team scans the patient's brain using magnetic-resonance imaging, a non-invasive technique which displays computer graphics images of patterns of blood flow. When the patient is asked to think about moving a limb, the motor area of the brain becomes active, and from its increased blood flow the precise location of the active region can be identified. This is where the electrode is implanted.

Inside the glass cone is a microscopically-thin gold wire, surrounded by nerve tissue extracted from the patient's leg, which stimulates neurons from the surrounding cortex to grow into the cell. Over a period of months, the neurons fuse with the wire. "It's like having a little piece of brain inside the electrode", says Bakay. Unlike previous generations of brain electrodes, the implant needs no cabling. It receives its power from an induction coil sown into in a baseball cap worn by the patient. Any signal picked up from motor neuron activity is detected and amplified by a tiny receiver placed just under the skull.

The patient subsequently undergoes a training programme using biofeedback. The electrical activity recorded by the implant controls the sound of a buzzer, and the patient gradually learns which thoughts make the buzzer sound louder and faster. Later, the buzzer is replaced by a cursor on a computer screen, and the patient learns to "think" the cursor from side to side.

So far, two patients have received Bakay's implants. The first, who suffered from a fatal degenerative motor neuron disease, received her implants in 1997, and lived long enough to take part in preliminary experiments. "We learned a lot about the basic principles from her", says Bakay. "She helped us identify the brain cells we were looking for". The second patient is a 53-year old stroke victim paralysed from the neck down, known as "J.R.". He has two implants, enabling him to separately control the horizontal and vertical movements of a cursor, and select icons which trigger synthesised speech. For the first time since his stroke, J.R. can communicate. "If you can use a computer, you can talk to the world", says Bakay.

While they admit that the research is in its infancy, Bakay and his team are confident that the brain-computer interface will one day transform the lives of severely handicapped people, eventually enabling them to operate artificial limbs as easily as if they were their own.

To many observers, cognitive engineering is by turns fascinating and frightening. Where does the human stop and the machine start? If the brain-machine communication is two-way, could the machine control the person? Hypothetical questions for now, perhaps, but they might -- one day -- become very real.

Toby Howard teaches at the University of Manchester.