1: Automobile drivers have been shown to be more aggressive around cyclists with helmets (and interestingly, males) than those without helmets. This increases risks for the cyclists and partially offsets the benefit of wearing the helmet.
2: Mandatory helmet laws significantly decrease bicycle ridership. This has two negative effects
2A: Cycling is a very healthy activity. Fewer cyclists implies more heart disease, diabetes, etc
2B: Fewer cyclists on the road increases the risk for those who remain due to some combination of inferior infrastructure and lack of attention by automobile drivers, who are not used to looking for cyclists in areas with low ridership
The combination of 1 and 2B leads to the somewhat counter-intuitive result that while helmets have been clearly shown to be effective in preventing injury, mandating them does not reduce the injury rate on the population level. Add in the costs associated with 2A, and these laws actually appear to be a net negative for the public. It is generally my opinion (one that I hope most people share) that the government should only restrict peoples' freedom when there is a compelling case to do so. I just don't see how you can make a compelling case here, as it appears the non-cycling portion of the public is actually worse off, not better off, due to these laws.
One other point in the article that I found interesting was the similarity in injury rates per hour for walking, cycling, and driving. It's actually rather hard to defend a cycling helmet law without implying that pedestrians and drivers should have to wear helmets as well. Can you really imagine demanding drivers where helmets in the car, or pedestrians don them before crossing the street?
A typical Japanese bicycle ride