I am currently performing experiments on the placebo effect in collaboration with Axel Cleeremans et Jérôme Sackur. Here, I present our hypotheses on the subject.
Placebos constitute a standard control in clinical trials since decades, but it is only fairly recently that they are gaining the status of a neuropsychological and physiological phenomenon in its own right, worthy of a systematic attention by fundamental research . The interest of scientists in placebos has definitively been renewed by the development of brain imaging technologies, which make the action of placebos ‘visible’ . Since then, the powerful psychobiological effect of placebos has been shown in the context of many conditions (e.g., Parkinson disease , pain ). Importantly, most of real not-inert treatments actually incorporate a placebo component; culminating with antidepressants the action of which has been shown to be attributable at 80% to a placebo effect . Nonetheless, if the placebo effects are now well acknowledged, the mechanisms underlying these effects are highly controversial and still poorly understood.
By ‘placebo effect’ or ‘placebo response’, we mean the psychobiological effects (e.g., analgesia) following the administration of any pharmacologically ‘neutral’ treatment (e.g., saline injection). Pain, for its clinical significance, is a crucial context when investigating placebos. Psychosocial and cultural factors  have been shown to be important in triggering and modulating placebo analgesia—pain reduction following a sham procedure. At the neuropsychological level, though, the large majority of theories consider expectations as the central factor for its production [7,8,9,10]. ‘Expectation’, however, is a blanket and vague term that leaves unspecified the neurocognitive mechanisms underlying placebos.
In view of neuroimaging studies showing dorsolateral prefrontal cortex (DLPFC)  activity during placebo analgesia, it has been proposed that the latter works by recruiting executing functions . For instance, the expectation of pain reduction might lead the subject to redirect her attention away from pain. Recent behavioural evidence, however, demonstrates that the effect of placebo analgesia does not interact with a concurrent attentionnally-demanding working memory task . This strongly suggests that placebos do not recruit online executive processes. That being said, such evidence does not tell us whether placebo analgesia reflect sensory or decisional processes. Fortunately, early data reveal that placebo analgesia mainly affects response bias, rather than sensory sensitivity, making participants more conservative after than before placebo analgesia administration . In sum, placebo analgesia does not seem to be mediated by attentional/executive processes but, at the same time, seems to affect high-level cognitive rather than low-level sensory processes.
Given these different elements, we hypothesise, first, that most placebos have their effect by altering participants’ metacognitive processes. Metacognition refers to the ability to evaluate, monitor and control one’s first-order cognitive states and processes . Our hypothesis is thus that the expectation of pain reduction would alter participants’ secondary evaluation of their pain state without affecting sensory processes on which pain emerges. Put simply, the confidence with which tactile painful stimuli, for instance, are evaluated would be less accurate in discriminating “correct” from “incorrect” answers after placebo administration. Specifically, placebos might affect such metacognitive accuracy by generating specific metacognitive biases . Participants might become more confident after placebo administration when categorising a stimulus as less painful than another and less confident when categorising a stimulus as more painful than another. This hypothesis would explain why participants become more conservative after placebo administration, as well as the reduction in pain ratings reported by most of placebo analgesia studies: being less confident as to whether a stimulus is painful, participants need a more intense stimulus to make pain judgments . In addition, the DLPFC has been shown to mediate the accuracy of retrospective metacognitive judgments; therefore, previous evidence showing DLPFC activities during placebo analgesia might actually reflect metacognitive rather than other executive processes.
Our second hypothesis is that different placebo conditions involve different mechanisms. The vast majority of studies use deceptive placebos. However, a recent study has shown that patients who received an open- (i.e., non-deceptive) placebo showed improvement of their irritable bowel syndrome condition . Participants were told that placebos had had positive effects on other participants (note that the study did not include the critical comparison between open- and deceptive-placebo). To the extent that in an open-placebo condition participants are aware of the condition, it is unlikely that open-placebos affect metacognitive processes. We hypothesise that an open-placebo analgesia leads the participant to adopt strategies such as redirecting her attention away from pain. Therefore, we predict that in contrast to deceptive-placebo, open-placebo analgesia should affect participants’ sensory sensitivity leading them to become more conservative. In sum, both deceptive- and open-placebo would lead to response bias but so by different routes.
jeanremy08 [at] gmail [dot] com
I am a Cognitive Scientist with a background in Cognitive Science and Philosophy.
My current research focuses on the way social influence affects human cognitive functioning and, more specifically, how it modulates certain aspects of consciousness.