Dr. Zuckerman offers some observations on the state of the art of the brain-computer interface.
Those loyal followers of this blog (thanks, Mom) realize that much of the material discussed has concerned the use of technology in the practice of neurology. Well, I am going to begin discussing more clinical topics in the next few installations. However, in order to ease the transition, I would like to offer some observations on the state of the art of the interface between the brain and controlling external devices, such as computers. This is known as the brain-computer interface (BCI), and there have been excellent discussions of this field at the Annual AAN meetings. For those interested, I would recommend purchasing the audio file of the course.
As an academic pursuit, BCI research has focused on the potential to utilize the analog signals produced by the cortex as a means for communication or improving the mobility of those patients with paralyzing disorders. In essence, the electrical output of the brain is no longer directed toward the motor system, but rather toward the cortex to produce signals that can be processed in such a fashion as to communicate intent to some external device. The device takes those signals and transforms the intent into the necessary output format. For example, patients can be trained to control their EEG patterns to rapidly (about one-third to one-half as quickly as using a joystick) and accurately control a cursor on a computer screen.
The training process enables the patient to learn how to produce the optimal signals, and the machine adapts to the best signals produced to provide the desired effect. These signals can be recorded in one of three methods:
· From surface electrodes, like standard EEG leads
· From electrodes surgically placed on the surface of the brain- e-Cog
· From implanted electrodes
Each of these methods has its advantages and disadvantages, but certainly the use of surface electrodes would be the most practical long-term solution if possible. The most useful location for the recording of signals is over the somatosensory cortex. Different frequencies can be produced, and patients are trained to produce more or less of those frequencies in specific areas to control the output device. For example, increasing a Mu rhythm in the left cortex may move the cursor upward (and conversely, decreasing those rhythms creates downward movement). A Beta rhythm from the right hemisphere may control horizontal movement. Three separate locations can provide the information (via algorithms) to move a cursor in three-dimensional space.
The description of the training process is very interesting. To provide the movement in one dimension, the patient is instructed to visualize a movement that eventually enables them to produce the desired control rhythms. However, by the time the patient can control the cursor in two dimensions, apparently adding the third dimension of control no longer requires the intermediate step of movement visualization. The control just takes place without conscious input- just as we use our own muscle system.
As fascinating as this topic is, there is very little practical application existent to date. The price of these systems is currently cost prohibitive, and there is not a sufficient potential market to justify commercial development of BCI products. It is sort of like an orphan drug, lacking in capital to develop the product. However, some companies have produced a BCI and are using that device in a commercial venture. One such company is NeuroSky, which produces a BCI called XWave, an interface for the iPhone family (retailing at $99.00) and Mindset ($200.00), which interfaces with your PC. Both devices use a single channel of EEG recorded from a dry electrode (no gel required) placed over the forehead. According to Engadget, a version of this type of headset will retail around $100.00 in the near future.
NeuroSky used this type of apparatus to help manufacture two toys. One, produced by Uncle Milton, coupled the control of brain wave activity (labeled attention or meditation) with a fan, which would then make a ball float. Of course, the ability to use just your minds activity to make an object float is reminiscent of “The Force” in the Star Warsmovies, and in fact the game was called “Force Trainer.” Mattel produced a similar game called “Mind Set.” Apparently the hardware supplied by NeuroSky for these games is very similar to the XWave and Mindset apparatus, but access to the output data is restricted. For a technical article, please see: “How to Hack Toy EEGs.”
So, we have a science fiction-grade technology that is in its infancy. Some commercial ventures are using this technology and are looking for the “killer app” to make it indispensible. Of course, using BCI as the means to control next-generation video games seems like a natural evolution. NeuroSky has several free apps available in the Apple Store, including the provocatively named NeuroBoy.
Other proposed applications could monitor a driver’s alertness while operating a car, teach children with ADD to concentrate better, or even allow athletes to train in a more focused fashion. However, since this is a technology still searching for applications, how would you apply a BCI to improve our world? Let us know if you have any notions as to how BCI can be used to our advantage.