A
game conceived by Smart Studio.
Tweaks by Matt Adereth and Dan Moskowitz.
Two
players are hooked up to
EEGs, which feeds the
data to a
computer. The
object of the game is to
minimize your
brain activity. The computer
displays the
difference of the
integrals of the
brain activities of the two players as a
ball that moves towards your
opponent's
goal if you are winning. The game ends after a
fixed amount of
time.
For one
minute before the
start of the game, each player tries to
maximize his brain activity. Then, during the game, we
express the player's
score (the integral of their brain activity) as a
percentage of the
average of the maximized brain activity.
Because of this calibration
technique, each player must be good at both maximizing and minimizing his brain activity.
Also, during the game, each player may decide to devote a certain small amount of his brain activity to tricking the other player into increasing his brain activity, e.g., by asking him questions or getting him to think about something.
So far, the best strategy we've come up with is asking the other player, "so what's your brainball strategy?"
Notice that we haven't
defined what "brain activity" is, but it really doesn't matter. We can define it to be anything we want (the EEG readout may not
correspond exactly to our
notion of "brain activity"), and the brain will learn what it means to
win, via
positive feedback of seeing where the ball is on the screen.
We figure that after about a month of playing brainball, we'd all be so good that our brains would have learned how to voluntarily slip into a coma... not such a good thing. Note that this would not be a conscious learning process, it's more like the brain learning how to use itself -- and when you think about it (and even when you don't), we really don't have conscious control over our brains (like we do over our muscles). In fact, I'm not sure if "to have conscious control over one's brain" is possible or even meaningful.
Also, since good brainball players would be able to minimize and maximize their brain activity at will, they could easily have their brain "speak" in binary (minimize=0, maximize=1) to a computer, albeit rather slowly.