Before I begin this, i'd like to note that I am not a neurobiologist, and its a weak area of mine, so please be gentle and tell me if a make a glaring error.
A strand of placebo research which has become more and more important with time has been the increasing focus on brain correlates of placebo responses. The biochemical history of placebo begins with Levine and his demonstration that naloxone blocks many placebo pain responses. Induced from this is the notion that placebo pain is mediated by the endogenous opioid system. This is not always true, and seems to vary based on the type of response which is wanted. . The lasting contribution of this research is that it paved the way for the placebo to come in from the fringes of medical science, as even the most dogmatic materialist could not deny the biochemical evidence as demand characteristics.
A recent meta-analytic review seems to argue that placebo effects in pain are quite large (d=.89) and that naloxone is quite effective in reducing them (d=.55), pointing towards an interpretation of placebo effects in pain being substantially mediated by endogenous opioids.
In this area, the work of Benedetti and his colleagues has been instrumental in unveiling the biochemical pathways through which placebos exert their effects, and much of this work is summarised in his book along with the work of others. It appears that both the opioid and dopaminergic systems are involved in the placebo effect. While Benedetti and others have done much of the research into the opioid system, De La Fuente Fernandez \cite{DeLaFuente-FernAƒA¡ndez2002} has blazed a trail in looking at the dopaminergic system.
It has been observed that the dopamine system activates not just to reward, but rather the expectancy of reward, and that this release varies as a factor of the certainty of the expectancies. In one study, the activation of the dopaminergic systems during placebo analgesia was correlated with activity observed during a monetary reward task, suggesting that the mechanisms of reward are a common feature of placebo effects\cite{Scott2007a} It has been argued that there is a descending link from the OFC to the Periaqueductal Gray Area (PAG) and from here to the amygdala, and that this link is responsible for the observed placebo effects.
Another interesting suggestion was that placebo analgesia experiments which show altered brain activity in the rACC and OFC demonstrate the existence of a generalised expectancy network. This hypothesis received some support from a recent experimental study which used either true or false sound cues to create expectancies for particular aversive tastes. This study showed that the rACC and OFC and to a lesser extent, the DLPFC activated in response to these expectations, suggesting that these parts of the brain may well be associated with the expectancies
An interesting finding arose from an experimental study into patients suffering from Irritable Bowel Syndrome (IBS) \cite{Lieberman2004}. This study looked at placebo using a disruption theory account, which accounts for neural changes due to placebo in terms of inhibition. The authors found that although the right ventro-lateral pre frontal cortex was activated by expectancies of analgesia, this activity was totally mediated by the dorsal anterior cingulate cortex which argues that this part of the brain is more foundational to the placebo response. Another study which looked at patients suffering from IBS found that naloxone did not reduce the size of placebo effects, which would suggest that these were not opioid mediated \cite{Vase2005}.
A further discovery around placebo analgesia is that it can be directed at specific sites in the body \cite{Benedetti1999} . This study induced expectancies of placebo responses at either the right or the left hand, and demonstrated the expected placebo effects. These effects were completely antagonised by naloxone, which suggests that they were mediated by the endogenous opioid system. This finding is interesting, as it suggests that the opioid systems can be activated at specific parts of the body, and not just globally as some former theorists have claimed. A more recent finding \cite{Watson2006} found that perhaps 50\% of participants in a placebo analgesia study generalised a placebo response across both arms, even though cream was only applied to one arm for each person. This study would suggest that the placebo analgesia phenomenon is quite malleable and subject to individual interpretation.
There is some evidence to suggest that some of the effects may involve both descending and ascending pathways within the brain, judging from the results of a study on mechanical hyperalgesia \cite{Goffaux2007}.This study used a counter-irritation technique and the use of a basin of water to act either as a placebo or nocebo. The authors suggested that the reflexes in the arm should not change if the placebo effect was completely cortically mediated, but the results suggest that descending pathways are equally as important in placebo analgesia. These pathways are controlled from the mid-brain and these findings suggest that the placebo effect exerts changes in large portions of the body, and is not exclusively a cortical phenomenon.
Further evidence in favour of this idea comes from the study of Matre \cite{Matre2006a} who noted large differences between placebo and control areas of mechanical hyperalgesia, again suggesting the involvement of the whole body in the response. In this context, the results of Roelofs et al. \cite{Roelofs2000} are worth considering. Using similar techniques to the two other studies referenced in this paragraph, they found no evidence that placebo effects cause changes in spinal reflex activity. However, this study also found no evidence for a placebo effect in general, which weakens their conclusions. It is worth mentioning that even though they found no significant effects, they did find a correlation between the brain activity and spinal reflexes, which suggests that they found an effect, but their study was either underpowered or used a badly designed expectancy manipulation (most likely the latter) \cite{Goffaux2007}.
An interesting finding which has come about through placebo research is what is known as the uncertainty principle in analgesia \cite{Colloca2005} , where they argue that the effects of any analgesic can not be accurately measured in a clinical situation as the awareness of being given this substance will activate the opioid system which will further reduce pain. This finding arises from work done previously, where it was shown that open injections of painkillers or placebo registered far more variability than hidden injections, suggesting that while physiological responses to analgesia may be similar across people, the awareness of treatment may invoke differential activation of endogenous painkilling systems which cause the total effects to appear to vary quite substantially \cite{Amanzio2001} . Some research has also confirmed that placebo and opioid analgesia share the same neural patterns of activation in the brain \cite{Petrovic2005}.
Some sterling work has also been done in the area of depression and placebo response. A fascinating study suggests that prior to treatment, placebos may induce changes in neurophysiology which predict later treatment response. This is an extremely interesing finding, however the authors used a new measure (that of EEG cordance) developed by themselves and to date, there have been no replications of the study. Another useful study of placebo neural activity in depression has also been conducted comparing the activation of particular brain regions following treatment with either prozac or placebo \cite{Mayberg2002}. . One fascinating finding of the Mayberg et al study is that areas of the striatum were activated, and this region of the brain is known to be rich in dopamine receptors, which may suggest that while the placebo response in depression is primarily opioid mediated, the effects of SSRI's may also influence the dopamine systems, which may account for their (slightly) superior effectiveness overall. However, some research shows that psychotherapy activates different brain regions in the treatment of depression, which argues against the existence of a common depression treatment pathway in the brain
So far, so interesting. My one complaint about this stream of research is that much of it reiterates old findings backed up with correlational evidence of which parts of the brain are involved. To the extent that it furthers understanding, its great, but to the extent that it substitutes for it, its worthless.
Weeks, S., & Tsao, J. (2009). Fabrizio Benedetti, Placebo effects: Understanding the mechanisms in health and disease , Oxford University Press (2009) ISBN: -13: 978-0-19-955912-1, 310 pages, $59.95. Journal of the Neurological Sciences, 281 (1-2), 130-131 DOI: 10.1016/j.jns.2009.03.013
