Neurobiology of placebo effects

Converging evidence from research since the 1970‘s substantiates that placebo responses are not merely a psychological, but a complex psycho-neuro-biological phenomenon that involve the activation of distinct brain areas as well as peripheral physiology including the release of endogenous substrates. Placebo analgesia represents one of the best studied placebo responses. Neuropharmacological evidence for the involvement of endogenous opioids in placebo analgesia dates back to the 1970‘s when Levine and his colleagues discovered that placebo analgesia can be blocked by administering the opioid antagonist naloxone (Levine, Gordon & Fields, 1978). In the past decade, functional brain imaging studies have contributed substantially to our knowledge of the brain mechanisms that mediate placebo phenomena.

Neuroimaging studies confirmed related mechanisms of opioid and placebo analgesia by revealing shared neural networks underlying both – opioid and placebo related analgesia (e.g., (Petrovic, Kalso, Petersson & Ingvar, 2002)). Since then, the relevance of this brain network has been substantiated by several studies using different procedures to induce placebo analgesia (including fake analgesic creams, sham acupuncture and others). All these studies demonstrated that placebo analgesia involves the activation of cingulo-frontal brain regions together with subcortical structures such as the PAG, hypothalamus and the amygdala. Connectivity analyses further showed that the behavioural placebo analgesic effect depends on an enhanced functional coupling of the frontal lobes and rACC with brainstem areas such as the PAG (Bingel, Lorenz, Schoell, Weiller & Buchel, 2006; Eippert, Finsterbusch, Bingel & Buchel, 2009; Kong et al., 2006; Sarinopoulos, Dixon, Short, Davidson & Nitschke, 2006; Wager et al., 2004). The opioidergic nature of this pain modulating system was supported by both pharmacological studies using the opioid antagonist naloxone and in vivo receptor binding approaches (Eippert, Finsterbusch, Bingel & Buchel, 2009; Levine, Gordon & Fields, 1978; Zubieta et al., 2005).

All these studies support the notion that placebo analgesia involves a top down activation of endogenous analgesic activity via the descending modulatory system, the very same system targeted by exogenous opioid administration (Fields, 2004). Recently, functional neuroimaging of the spinal cord was used to investigate the involvement of spinal cord processes in a model of local placebo analgesia (fake analgesic cream applied to the left forearm). It was found that that pain-related activity in the ipsilateral dorsal horn, corresponding to painful stimulation, is strongly reduced under placebo. These results provided direct evidence for spinal inhibition as one mechanism of placebo analgesia and highlighted the fact that psychological factors can act on the earliest stages of pain processing in the central nervous system (Eippert, Finsterbusch, Bingel & Buchel, 2009).

The above mentioned neural circuitry underlying placebo analgesia, involving cingulo-frontal brain regions interacting with subcortical nuclei, does not seem to be specific for the generation of placebo responses in the pain domain. Similar patterns of brain activations have also been observed in brain-imaging studies of emotional placebos using placebo-anxiolytics or placebo-antidepressants. Petrovic et al. found a shared modulatory network involving the rostral anterior cingulate cortex and the lateral orbitofrontal cortex during both emotional placebo and placebo analgesia (Petrovic et al., 2005). These effects were correlated with the reported placebo effect and were predicted by the amount of treatment expectation. Similarly, therapy with placebo or the SSRI fluoxetine in patients with major depression had similar effects on brain metabolic changes in the anterior and posterior cingulate cortex and prefrontal cortex after 6 weeks of treatment (Mayberg et al., 2002).

Similarly, studies that investigated placebo responses in Parkinson‘s disease demonstrate that the expectation-induced improvement of motor functions following a sham treatment is associated with the release of endogenous dopamine in the basal ganglia. de la Fuente-Fernandez et al. used raclopride-PET to study involvement of the endogenous dopamine system for placebo responses in Parkinson patients (de la Fuente-Fern├índez, Schulzer & Stoessl, 2004). Following the administration of a placebo that the patients believed to be apomorphine (a powerful anti-parkinsonian treatment), they observed increasing dopaminergic neurotransmission in the striatum, which corresponded with clinical improvement. Similarly, Benedetti and colleagues recorded changes in the firing patterns and bursting activity of neurons in the subthalamic nucleus and associated motor-circuitry during the placebo response in Parkinson patients undergoing the implantation of deep brain stimulation (Amanzio, Corazzini, Vase & Benedetti, 2009; Benedetti et al., 2004).

However, the dopaminergic system is a complex system involved in a variety of brain functions, including reward mechanisms. This might be the reason why dopamine release during a placebo response is not specific for Parkinson‘s disease. Rather, dopaminergic mechanisms have been associated with placebo responses or placebo-like phenomena in different systems, including the pain system (Schweinhardt, Seminowicz, Jaeger, Duncan & Bushnell, 2009; Scott et al., 2008). The detailed context and condition-specific role of dopamine for placebo responses needs to be identified in future studies. Taken together, exemplary available clinical and experimental data, for three different systems, strengthen the notion that placebo-induced clinical benefits involve neuro-biological responses which are the very same as those targeted by specific pharmacological treatments.

The specificity of the involved brain circuitry and neurotransmitter systems is far from clear. Further, in contrast to the expectation-induced placebo effect, less data are available on the neurochemistry and involved brain areas during behavioral conditioning of placebo responses (Enck, Benedetti & Schedlowski, 2008). In rodents it has been demonstrated that conditioned immunosuppression is mediated via the basolateral nucleus of the amygdala and the insular cortex (Pacheco-Lopez et al., 2005). However, it is unknown whether these areas also mediate placebo or nocebo effects in other experimental or clinical conditions. Finally, little is known about activation processes of brain circuitries during the nocebo response. To summarize, experimental data indicate that single neurobiological or psychobiological mechanism cannot explain placebo and nocebo phenomena in general. Instead, common and system-specific mechanisms may coexist by which placebo or nocebo responses are mediated across diseases and experimental conditions. These mechanisms will be highlighted in the current projects.