Mind Over Matter: Brain Control Interfaces Become a Reality_2

Current Uses of Brain Control Interfaces

Picture courtesy of Dr. Jonathan Wolpaw, Wadsworth Center, Albany, NY.

ET: Can you give me an example of a recent research application of your technique?
GS: We're about to publish a study now where we are demonstrating by far the fastest system that has ever been demonstrated when working with human subjects—possibly with monkeys too but I'm not sure. They system involves allows a subject to type over 20 characters per minute purely by thought alone. That's still a lot slower than typing, but when compared to a system for example where the subject has to blink their eyes as a cursor slowly scans over the alphabet, it's a big improvement.

ET: Was the subject disabled in some way?
GS: No, the subject involved had epilepsy. Since those patients have to have electrodes implanted anyways as part of the pre-operative surgical protocol for treating epilepsy, it gives us the perfect opportunity to do research with them.
ET: Will you ever get an implant yourself?
GS: I don't think my wife will let me so probably not.
ET: I thought most spouses wanted greater control over their partners?
GS: If you're willing to go off the record then I might have some answers for you!
ET: Tell me about the SIGFRIED project.
GS: SIGFRIED is a technology that allows brain activity to be visualized in real time for clinical uses. It's been acquired by and will be put to use by several clinics both in the United States and in Europe for the purpose of localizing activity in the brain prior to invasive brain surgery.
ET: Are you software programmer or hardware designer yourself?
GS: I've been programming computers for 25 years now.
ET: What language is SIGFRIED written in?
GS: C++.
ET: Are some people naturally more talented at using a BCI than others?
GS: Using a BCI is skill like any other. Some people are naturally better at it, while others never seem to be able to pick it up. Sometimes it's a natural talent issue, while in many others it's more an issue of motivation.
ET: What is the maximum number of disparate signals you can detect reliably with a BCI?
GS: You have to distinguish between the number of signals being extracted in what we call "open loop" fashion where we just monitor what the brain is doing and the number of signals that people are actually using to control a device. When they actually control the device I think the maximum number is three, at least in humans. For example, controlling a cursor in three dimensions simultaneously requires mapping one signal to the control of one axis. In other examples where you simply monitor what the brain is doing, you can record and decode more than three signals at the same time. We just published a study where we showed it's possible to determine from the brain signals how people are moving each of the five fingers on a hand and detect which ones they were flexing. This means you can not only tell which finger is moving but how it's moving just by looking at brain signals.
ET: If only three signals are currently being detected, how did those people select letters for "thought typing" in the system you mentioned?
GS: That system was actually quite simple and uses only one signal. The subject watches a matrix of the characters. Whenever the right row or column is flashed—the one that contains the character the subject wants—the brain produces a response that's different than when the subject sees when one of the unwanted rows or columns flash. Of course, it could potentially be faster if it wasn't a binary system and you could detect what word or character you were thinking of. That's something we are working on at the moment and we are having somewhat encouraging results.