Addiction's grip now seen as 'extreme memory'

Explain this: An addict sweats through withdrawal. He commits to staying sober. With years of effort, he builds a life he loves. And then, one day, he passes his old shooting alley or gets pain pills from the dentist, and boom. Relapse. It all comes crashing down. By all accounts, something similar may have hit US Representative Patrick Kennedy earlier this month.

Old theories of addiction seem to fall short here. If the essence of addiction is dependence on a drug and fear of withdrawal symptoms, then why should this happen to a man who long since went through withdrawal? Or if addiction is about pleasure, why should a man embark on a course that will surely bring nothing but pain?

Last week, brain scientists gathered at the Massachusetts Institute of Technology batted around a newer theory that could fit a few more pieces into the puzzle and is already spurring experiments on new potential treatments.

The idea is that addiction may be a form of ''extreme memory" or ''pathological learning." And that addiction can be so potent and persistent because it takes over learning processes in the brain that are central to humans' very survival.

The brain evolved to identify essentials such as food or water and to record exactly how they can be reached, said Dr. Steven E. Hyman, a neuroscientist who is provost of Harvard University. And when it finds -- or even only expects to find -- such essentials, the chemical messenger dopamine is released. Its job in the brain is to say: ''This is very important; let's remember exactly how we did this."

Even though they provide no benefit to the body, drugs can usurp that system, he said, by releasing dopamine -- so much dopamine, in fact, that little else can compete, leading addicts' brains to ''overlearn" the false message that drugs are good.

Those surges of dopamine, the theory goes, contribute to the laying down of long-term memories and associations that remodel the connections in the brain and can last forever.

Last year Hyman published a review article called ''Addiction: A Disease of Learning and Memory" in the American Journal of Psychiatry and concluded that neuroscience was reaching ''a far more accurate and robust picture of addiction than we had a few short years ago."

Scientists who spoke at the MIT forum, held at the Picower Institute for Learning and Memory, cautioned that addiction is hideously complex and the memory aspect is surely only a piece of it.

But the ''extreme memory" idea has been gaining attention among researchers in the last several years, they said, raising hopes that they may be able to apply some of the recent scientific progress on understanding memory at a molecular level to the conundrum of addiction.

Already, researchers working with animals are trying out drugs that affect memory to see whether they can help treat addiction, said Dr. Nora Volkow, director of the federal National Institute on Drug Abuse.

The institute is funding several studies that treat ''addicted" rodents with drugs that allow the replacement of old memories with new and different ones. The idea, Volkow said, is to break the power of cues that can cause relapse.

For example, she said, a rat can be conditioned to associate a certain place with cocaine. But if it is given a drug that ''breaks" old memories as they come up again, a new no-cocaine memory may replace the cocaine one, breaking the association in the animal's brain between the place and the drug.

Researchers hope that such memory-altering medications might even ''erase the memory of the pleasure associated with the drug," she said.

One medication that has shown promise in breaking associations, even in humans, is D-Cycloserine, which is commonly used at higher doses for tuberculosis. Experiments have shown that it might be an effective treatment for phobias. At this point, however, the studies are too new to tell whether memory drugs will work for addictions, Volkow said.

Recent research on the genes associated with addiction also seems to support the theory that learning and memory play an important role, Volkow said.

In animals, when researchers manipulate genes involved in long-term learning, they create changes in the animals' response to drugs of abuse, she said, ''suggesting to investigators that these genes may be important in regulating the vulnerability to addiction."

Other animal studies have shown that addictive drugs have dramatic effects on learning circuits, researchers said.

A key to understanding addiction is to understand why and where such changes occur in learning circuits, said Dr. Robert C. Malenka, a neuroscientist at Stanford University. ''And if we understand that, maybe we'll be able to develop therapies that could begin to reverse them."

As this ''pathological learning" model gains force, it may help reduce some of the stigma of addiction, some at the Picower conference said. Addiction, according to this model, is nothing so simple as a weakness of will, they said.

The new model also provides a possible biological underpinning for aspects of the disease that addicts and those who treat them have long observed, said Dr. Shelly F. Greenfield, associate clinical director of the Alcohol and Drug Abuse Treatment Program at McLean Hospital.

''We often think of treatment as quitting and staying quit," she said. ''Most people, I think, would agree that the quitting part is easier than the staying quit."

Carey Goldberg can be reached at goldberg@globe.com.