In the first type of placebo response – designated 'non-pharmacological' – we feel better because we consciously expect that we should, perhaps because we have received a dummy pill, or experienced some other ritual recognised as medicine in our culture. They demonstrated the existence of two different types of placebo response, each mediated by a different biochemical process.

PCDR exploits a curious phenomenon first described by Martina Amanzio and Fabrizio Benedetti in 1999. PCDR would, they suggest, be a way to minimise the side effects of long-term medication. In this month's issue of Trends in Pharmacological Sciences, Bettina Doering and Winifred Rief argue for the incorporation of placebo into general medicine, as part of placebo-controlled dose reduction. So there's plenty of evidence that the placebo effect is real, and not 'all in the mind'. And researchers have demonstrated that the same regions of our brain are activated during both Prozac and placebo treatment. Placebo treatments have also been found to stimulate dopamine release in the brains of patients with Parkinson's. Studies of pain relief have found that patients given placebo treatment show increased levels of β-endorphin, a natural painkiller, in their cerebrospinal fluid.

We know that our belief in a treatment's effectiveness can induce measurable physiological changes. How can that be? And, perhaps more importantly, why would we do it?

Pavlov effect