Originally Posted by New Scientist

A spliff test for science

It's the oldest but most important scientific question when two phenomena appear related: does one cause the other, vice versa, or is the apparent relationship pure coincidence? The question came up again this week when an Australian study demonstrated that 15 men who had all smoked marijuana heavily for at least 10 years had shrunken brain structures compared to those in non-users.

So was it the cannabis that on average shrank their hippocampuses by 12% and their amygdalas by 7%? Or were these same regions small to start with in these men, and if so, was it something that played a part in their strong liking for cannabis?

Certainly, both these regions are heavily affected by cannabis because they are both unusually rich in molecular receptors for delta-9-tetrahydrocannibol (THC), the psychoactive component in weed. The hippocampus is vital for storing memories and for the perception of time, and marijuana is known to affect both. Likewise, the amygdala is the brain's "fear" centre, and plays a key role in whether we react aggressively to events. Again, this fits with the observation that cannabis users sometimes develop paranoia.

To come back to the Australia study, is it equally possible that such prolonged exposure to cannabis wears out and shrinks these cannabis-sensitive regions? Again, we're back to cause and effect.

The only way to resolve it once and for all (as pointed out by the Australian researchers themselves at the end of their paper) would be to have brain scans of people before and after they began smoking cannabis. That way, you could see whether these regions did actually shrink the more cannabis they were exposed to. Or whether some people with unusually small regions at the outset turned out to be more attracted to the weed.

Unfortunately, a study to a wide population of children and kept as a general resource for researchers. If any of the scanned children subsequentfind out by deliberately giving cannabis to volunteers then following them for many years to see if their brains shrank would be unethical. Ethical comparisons could only be done if scans had been performed randomly on ly became heavy dope users, it would be easy to check back and monitor whether brain regions were changing size. But obtaining the scans would cost a huge amount of money without any guarantee that it would yield any findings of interest.

So for now, we simply don't know for sure whether cannabis is genuinely changing brain architecture. And the same dilemmas apply to study of all addictions. Which is why some researchers contacted by New Scientist cautioned against sensationalising the Australian results.

"You must be very careful looking at this paper in isolation," says Tim Williams, who studies addiction at the University of Bristol. "With this kind of study, you can't tease out cause and effect." Williams also pointed out that a study in 2005 of long-term cannabis users by researchers at Harvard Medical School found that there was no effect on the size of their hippocampuses. "I'm surprised the Australians found an effect where others haven't," he adds.

The take-home message is clear! Be cautious about concluding too much from addiction studies which might confuse cause and effect. Yes, it could be down to the drug, but equally, it could be down to your pre-existing brain architecture, and the effect of that on your personality.

Andy Coghlan, New Scientist reporter