A lot of information passes through our brains, and most of us have a natural ability to filter out what is important for us to remember immediately and what can be filed away for later use--or forgotten altogether. However, schizophrenics lack this filter, and scientists hypothesize it is due to excessive release of dopamine into the brain. To find out what exactly this excessive dopamine release might do, scientists at the University of Texas at Austin and Yale University created a computer model, or neural network, to simulate dopamine release. The result was that the network recalled memories in a schizophrenic-like way.
"The hypothesis is that dopamine encodes the importance-the salience-of experience," Texas grad student Uli Grasemann said in a release. "When there's too much dopamine, it leads to exaggerated salience, and the brain ends up learning from things that it shouldn't be learning from."
Many researchers believe that schizophrenics suffer from what is known as "hyperlearning," or an inability to forget or ignore as much as they should. As a result, they learn the ability to take what's meaningful out of the vast amounts of information pumped through our brains every day. So, schizophrenics make connections that are not real or they lose the ability to weave a coherent narrative out of life.
The neural network used by Grasemann and his adviser, Professor Risto Miikkulainen, is called DISCERN. The two were able to simulate excessive release of dopamine by increasing the system's learning rate--or, essentially, telling it to stop forgetting so much. "It's an important mechanism to be able to ignore things," says Grasemann. "What we found is that if you crank up the learning rate in DISCERN high enough, it produces language abnormalities that suggest schizophrenia."
Then, like those who suffer from schizophrenia, the neural network began putting itself at the center of delusional stories that incorporated elements from others it had been told to recall. In one answer, for instance, DISCERN claimed responsibility for a terrorist bombing. In other instances, it replied to requests for a specific memory with a jumble of dissociated sentences and abrupt digressions.
The DISCERN experiment does not completely prove hyperlearning hypothesis, Grasemann said, but it does offer good support for it. And it shows how neural networks can be useful in simulating the human brain.
"We have so much more control over neural networks than we could ever have over human subjects," he said in the release. "The hope is that this kind of modeling will help clinical research."
- read the release from the University of Texas at Austin
- and Slashdot has a discussion going on the subject