Robots Read Human Thoughts

Author:  Ghatalia Pooja

Institution:  Molecular Biology
Date:  June 2007

Uncertain whether to bubble the "A" or "C" on a multiple choice test, students dig into their pockets to find a coin to toss. But, the new "Brain-Machine Interface" machine can interpret the mixed brain waves in indecisive minds to help make the right choice.

Using functional magnetic resonance imaging (fMRI) machine, the joint project by Kyoto-based Advanced Telecommunication Research (ATR) Institute International and Honda Research Institute provide an interface between the human mind and a robot.

The fMRI records data on the changes in blood flow connected with neural activity in the motor cortex (part of the brain associated with movements). A machine algorithm analyzes the fMRI data and the decoded data is transmitted to a robot.

For example, imagine that a subject in an fMRI machine forms the rock, paper or scissors with his hand (Rock-paper-scissors is a hand game often played by children to select their teams randomly). Based on the subject's finger movements, when he forms a rock, paper or scissors with his hand, the robot hand reproduces the choice within about seven seconds with 85 percent accuracy.

"A brain-machine interface is only one of the many possible applications of the decoding technique," said cognitive neuroscientist Yukiyasu Kamitani of ATR. "For instance, if you could decode a person's attitude towards a product, you could use that for marketing."

Honda Research Institute's Tatsuhiko Sekiguchi envisions a brain analyzer that would monitor drivers' mental states and warn of drowsiness or inattention.

This technology is potentially applicable to other types of non-invasive brain measurements such as the brain's electric and magnetic fields and brain waves.

- By Pooja Ghatalia.