Slashdot Mirror
Think And Click
cecil36 writes: "Yahoo! has reported that scientists have discovered neural technology that allowed a monkey to control a mouse cursor using the brain. ... Further reading states that this technology can be used with the paralyzed or those with Lou Gehrig's Disease to allow them to use their computers."
This is really nothing new.
The Cyberlink Interface: Hands Free Brain-Body Actuated Control for Augmentation and Enhancement of Human Computer Interaction(produced in 1999), and their website
And an article from last year about a similar device.
The Air Force must not have gotten too far because DARPA is currently requesting proposals for research leading to a Brain-Machine interface.
The problem with most brain-machine interfaces is the skull and fluid surrounding the brain. Both of these elements serve as spatial and temporal filters degrading the usefullness of electrodes placed outside the skull as control sensors.
As the article states, the monkeys controlled a screen cursor, not a mouse cursor. There was no mechanical pointing device (mouse, trackball, etc.) involved. What sets this research apart from other work done in this area is that electrodes were implanted directly into a specific, targeted, area of the brain. Other research had been done with either generalized implanted electrodes or surface electrodes (like are used on an EEG machine).
Chaos maximizes locally around me.
These guys have had a commercially available
brain actuated mouse cursor gizmo out for years.
http://www.brainfingers.com/cyberlink.htm
This press release is kinda interesting. First, the work was presented over two months ago. Second, the work was carried out in the lab of Richard Andersen. Yet Andersen, who has spent a lot of the last decade reorganizing his lab around such efforts, was not mentioned. Instead, Meeker, a graduate student on the project was named. In addition, all this work in the Andersen lab was spearheaded by a person who is now at Stanford, Krishna Shenoy who recently left the lab. In addition, the intellectual property for the project, the patents, are co-authored by Andersen and Shenoy.
It's kinda weird when you know a bit about the work behind these press releases, and then see how it is actually presented to the "lay" public.
Personally, I think the project has a low probability of success. A neural prosthetic device should be interfaced with as peripheral part of the nervous system as possible. This group has chosen to use as abstract a part of the nervous system as possible. But maybe they'll prove me wrong.
The actual reseacrh described in the Yahoo article using implanted electrodes seems a bit strange - though the claim to have identified a few individual neurons is interesting.
Most of the other groups are working with stick-on electrodes. At the moment all they can do is move a mosue around a screen and click, but progress seems to be good - Correct recognition is around 70% after 5 one-hour sessions, which sounds impressive to me. The big obstacle to getting this into service for real people with disabilities is that the hardware is currently a bit chunky, especially the EEG machine. But we all know what happens to hardware, very, very quickly.
Oh - and, yes, the guy i talked to says the thing that secretlty drives him is eventually using it to play Quake. (Wonderful thing, altrusim)
Now wouldn't that be cool.(Unfortuantely you have to shave your head, I think!)
Daniella is part of Richard Anderson's lab at Caltech. They research motor planning and spatial orientation. It is a very interesting place.
As pointed out in the article, the area from which they record makes this experiment significantly different from previous ones. Several lab have done similar work, but they were less sure of the origin of their signal. Much of the sensory and motors areas of the cortex are right next to each other. It was not clear whether the recorded signals were motor signals or sensory signals driven by stretch sensors within muscles or something similar. The area Daniella records from is fairly far away from sensory cortex. There is much less chance that they are recording feedback from the sensory side. For comparison, examine an older story from a team of competitors.
Take a look at the Cyberlink You wear a headband and control the mouse by thinking. Sounds like the same thing.
Indeed. Prosthetics are the most plausible useful application of this technology for the near future. I still puzzle at the claim that several neurons were identified in the initiation and control of volutary movement, though. Every voluntary muscle command channel contains several feedback loops that essentially dampen and modify the signal to ensure that the movement doesn't overshoot or fall short of the target. These signals come from cells within muslces that inform the brain of the amount of current tension in the muscle and from specialized neurons within the joints that report flexion angle with one-degree accuracy.
People without the most important of these channels basically have Parkinsons. They can initiate action, but the action is delayed because they have a difficult time ending their previous action. Similarly, they shake violently the entire time, because of the grossly overestimated signals coming from their motor cortex. This makes me wonder how fine the mouse movement control could have been on the part of the monkey. People with parkinsons are severaly disabled, but might still be able to complete such a large-scale task with 90% accuracy, too.
To really use this for prosthetics, you'd have to not only detect the impulse to commit the action, but you'd also probably have to send signals back up via the remaining somatic nerves [in the case of amputation] or directly into the brain [in the case of degenerative disease.] Mental signals are not a matter of On or Off, they are on a sliding scale from strongly inhibitory to strongly excitatory. Signal regulation is the golden egg.
If these are the few neurons responsible for initiation of reaching action, how can the rest of the system determine when the monkey means to reach normally or reach virtually? I'd like to know whether there was any twitch in the arm when the mental cursor was moving.
There is a fantastic difference between "up, down, left & right" and "reach for and grasp object 3 feet from here." It's nice to see enthusiasm, but it's a little premature.
Philip Kennedy did something similar in 1998 with humans. I witnessed an operation where he implanted an electrode in a monkey over 10 years ago. He also received a 1999 Discover Technical Innovation Award for the same work. Check out: http://www.emory.edu/WHSC/HSNEWS/releases/feb99/02 2399brain.html