Skip to content

Opening the mind to machines

Scientists connect the brain to computers

Cyborgs may soon be reality. As scientists learn to connect the human brain directly to machines, the possibility of mechanical limbs and cognition-enhancing computers is increasingly plausible.

This summer, Andrew Schwartz, a neurobiologist at the University of Pittsburgh, enabled a monkey to dexterously control a mechanical arm using only its thoughts. To do this, Schwartz decoded brain activity by identifying the neurons responsible for certain types of movements. He started by recording from neurons using electrodes, and identified a “code” by matching the firing rates of certain neurons to the monkey’s specific arm movements. Next, the monkey’s arm was gently restrained, and the signals from the brain were fed into a computer attached to a robotic arm. The computer turned the signals into arm movements, and with a little practice, the monkey learned to use the arm to feed itself.

The arm was simple, consisting of a basic shoulder, elbow, and two-fingered hand. But Schwartz hopes to one day make a more advanced limb, – one with many joints and capable of fine object manipulation.

“We are currently working to extend this to the wrist and hand with anthropomorphic actuators. This is a big challenge, as the control problem – with the big increase in the number of joints to control – grows dramatically,” Schwartz said.

As the technology advances, brain signals could also be used to activate a paralyzed individual’s muscles, allowing them to live normal lives.

Recent studies by Dr. Lee Miller, a physiology professor at Northwestern University, have shown that paralyzed muscles contract when stimulated electrically. Miller hopes the technology will route signals from the brains of paralyzed people to the spinal chord and muscles.

Creating a system precise enough to detect brain signals and convert them into electric current is a major hurdle. For the brain to activate the right muscles at the right time, with just the right force, it needs to send precise signals.

According to Schwartz, another challenge in applying this technology to humans includes the unknown interactions of brain tissue and electrodes.

“There is a lack of basic scientific knowledge about the cascade of events that take place in the brain when an electrode is introduced and kept there permanently,” he said.

Rüdiger Krahe, a neuroscientist at McGill, expects progress on these issues will come from interdisciplinary efforts.

”Neural prosthetics is very strongly interdisciplinary. There are collaborations between neuroscientists, engineers, and mathematicians. In general, this is where you get the most exciting approaches – when people from different disciplines get together, cross boundaries, and combine ideas, and new things come out,” he said.

Recent successes in decoding the brain activity cause one to wonder how good we could get at understanding the code. Will we one day be able to read thoughts from a pattern of brain activity? If it is possible to make neural prosthetics that aid motor function, what about cognitive prosthetics to help areas that control memory and other cognitive functions? Will we one day be able to download information into our brain, Matrix-style?

With respect to understanding the “code” for abstract thoughts, Krahe is pessimistic.

“There are neurons in many different parts of the brain whose activity may be relevant at one time. To understand what’s going on, we really would have to extract the whole pattern and that’s such a huge combinatorial problem, that I don’t see us getting there,” Krahe said.

Schwartz also quickly extinguished dreams of cyborgs in the near future.

“It is important to remember that we cannot build devices to do things we don’t understand. There is very little about higher-level brain function – decision-making, learning, memory, language – that we do understand, so making claims about devices that work in these realms, seems to me, to be unwarranted,” he said.