Please Scroll Down to See Forums Below 
redguru
New member
University of Texas
AUSTIN, TX - A University of Texas scientist has grown a living "brain" that can control the video game Quake 3: Arena made by id Software, allowing scientists to study how brain cells compute and deal with sensory input.
The "brain," a group of 50,000 nerve cells taken from a dog and raised in a Petri dish, lets scientists observe brain functions at the cellular level in real time. Scientists hope to better understand how the brain deals with sensory input which may lead to treatments for certain types of blindness and brain trauma.
As a living computer, the cells may one day be used control robots to handle dangerous tasks such as navigating complex paths to let bomb squads remotely diffuse a bomb or even to build soldier robots designed to take out entrenched enemies.
"I'm interested in how the brain makes decisions based on enormous amounts of real time information," said John Thomson, the UT professor of biomedical engineering who designed the system. "Computers today can do a certain set of things very well, but getting a computer to perform simple functions that animals and humans do everyday is incredibly difficult. If we can understand how the brain functions in better detail, we can design adaptable computer systems capable of better pattern recognition."
Thompson's experimental "brain" controls the common game Quake 3: Arena through a grid of 120 electrodes that interact with a layer of living dog neurons. The electrodes are interfaced with a standard desktop computer in order to control the player character.
The electrodes both sense the interactions of the neurons and provide feedback about the environment. By looking at how the cells interact with the game, scientists can determine how the neural network establishes connections and begins to compute.
When the neurons are first placed in the dish, they are separate and unconnected. After a while, they being to extend lines toward each other to makes connections. "It takes time, but they naturally make connections and begin to compute," said Thompson.
Location awareness, collision detection, and eventually enemy location data is fed to the "brain" through the feedback system. The neurons analyze the data and send signals to the game controls. After several trials the neurons progress to the point where they can actually navigate a level and engage an enemy, usually a simple computer controlled character, called a bot, included in the game.
"When the brain is first hooked up to the game, it doesn't know what it is doing so it randomly moves around," Thompson said. "After enough data flows to the brain it begins to change the neural network, it starts to control the character in a progressively more intelligent way."
Though the brain can successfully compete with a bot, the goal is to gain better understanding of how brains function, Thompson said.
AUSTIN, TX - A University of Texas scientist has grown a living "brain" that can control the video game Quake 3: Arena made by id Software, allowing scientists to study how brain cells compute and deal with sensory input.
The "brain," a group of 50,000 nerve cells taken from a dog and raised in a Petri dish, lets scientists observe brain functions at the cellular level in real time. Scientists hope to better understand how the brain deals with sensory input which may lead to treatments for certain types of blindness and brain trauma.
As a living computer, the cells may one day be used control robots to handle dangerous tasks such as navigating complex paths to let bomb squads remotely diffuse a bomb or even to build soldier robots designed to take out entrenched enemies.
"I'm interested in how the brain makes decisions based on enormous amounts of real time information," said John Thomson, the UT professor of biomedical engineering who designed the system. "Computers today can do a certain set of things very well, but getting a computer to perform simple functions that animals and humans do everyday is incredibly difficult. If we can understand how the brain functions in better detail, we can design adaptable computer systems capable of better pattern recognition."
Thompson's experimental "brain" controls the common game Quake 3: Arena through a grid of 120 electrodes that interact with a layer of living dog neurons. The electrodes are interfaced with a standard desktop computer in order to control the player character.
The electrodes both sense the interactions of the neurons and provide feedback about the environment. By looking at how the cells interact with the game, scientists can determine how the neural network establishes connections and begins to compute.
When the neurons are first placed in the dish, they are separate and unconnected. After a while, they being to extend lines toward each other to makes connections. "It takes time, but they naturally make connections and begin to compute," said Thompson.
Location awareness, collision detection, and eventually enemy location data is fed to the "brain" through the feedback system. The neurons analyze the data and send signals to the game controls. After several trials the neurons progress to the point where they can actually navigate a level and engage an enemy, usually a simple computer controlled character, called a bot, included in the game.
"When the brain is first hooked up to the game, it doesn't know what it is doing so it randomly moves around," Thompson said. "After enough data flows to the brain it begins to change the neural network, it starts to control the character in a progressively more intelligent way."
Though the brain can successfully compete with a bot, the goal is to gain better understanding of how brains function, Thompson said.
