Researchers have created a model of the human brain that could help researchers study the mind in ways that would be unethical in humans. However, it is not simply its virtues that make the model brain so appealing; its mistakes are fascinating as well.
Called "Spaun", the model uses 2.5 million virtual neurons - an impressive number, but it pales in comparison to a human brain, which uses 100 billion neurons. Though it is powered by a computer and software, its mistakes are surprisingly human.
If you ask Spaun a question, it will pause the same amount of time that humans do. If you give Spaun a list to memorize, it will stumble if the list becomes too long. If Spaun is given a list, it remembers the first and the last numbers best - just like humans. It is the first model of a human brain that can perform a variety of tasks and has behaviors, which makes it ideal for studying brain processes as diverse as aging and mental illness.
While Spaun is surely not the first or most powerful brain model, it is the only one that can perform an assortment of tasks. Spaun can perform eight tasks, to be precise. It can copy what it sees, recognize numbers written in different handwriting, and complete patterns after giving examples.
That makes Spaun strikingly like the human brain, which can use the same faculties to perform activities as diverse as riding a bike and solving a puzzle. Spaun can make it possible for researchers to create robots that perform various tasks. Currently, robots are only able to perform a single task, like playing chess.
Because Spaun acts like a human brain, researchers are able to perform experiments that would be entirely unethical in human participants. For example, one scientist deleted neurons for a test in an effort to study the aging process.
Unlike humans, though, Spaun's computer takes two hours to run a single second of Spaun simulation. Researchers hope to improve Spaun enough so that the computer can operate in real time.
In the future, researchers hope to input Spaun's faults into robots to make the assistants more human-like.
"Those kinds of features are important in a way because if we're interacting with an agent and it has a kind of memory that we're familiar with, it'll more natural to interact with," Chris Eliasmith from the University of Waterloo said to LiveScience.
Researchers wrote about Spaun in the journal Science.