Fabrizio Benedetti – Placebo Effects

Hier meine Unterstreichungen aus dem schönen Buch.

Not only do animals scratch, rub, and lick themselves, but they scratch and rub their companions as well, devoting as much as 20% of their total day to this one activity, depending on the animal species.

Massage-like stroking in rats raises plasma oxytocin, and increases pain tolerance, which indicates that the mechanical stimulation of the skin affects oxytocin secretion (Agren et al. 1995).

It belongs to a Homo erectus dating back 1.7 million years who had lost all but one tooth several years before death. This specimen represents the earliest case of severe masticatory impairment which raises questions about subsistence strategies in early Homo. In fact, it may attest to evolution’s oldest known example of some kind of altruism and compassion for the elderly and handicapped in the social group.

Most, if not all, of these drugs and procedures were ineffective, with only a few possible but unlikely speculative exceptions. For example, if opium was added to the concoction, it was likely to produce specific analgesic effects. Likewise, bleeding was likely to have specific effects in some circulatory diseases (Shapiro and Shapiro 1997; Benedetti 2008).

Similarly, although the use of salicylic acid as a painkiller is very old, the discovery of the peripheral effects of aspirin (i.e. acetylsalicylic acid) is very recent. In 1964, Robert Lim showed that the analgesic effects of aspirin are not attributable to an action on the central nervous system but rather to a peripheral effect (Collier 1984). To demonstrate this, Lim performed harsh experiments in awake dogs. The spleen of dog B was surgically connected to dog A, so that the blood of dog A supplied the spleen of dog B (Figure 2.2). Then the spleen of dog B was injected with a pain-inducing substance so as to induce a painful sensation in animal B. Lim showed that, by injecting aspirin in dog A, an analgesic effect was induced in dog B, which unequivocally indicates that aspirin does not act on the brain but rather on the spleen itself. Following this experiment, John Vane showed that aspirin inhibits prostaglandins in rodents (Collier 1984).

Today, anaesthesia represents the basis for all surgical operations. Without anaesthesia, no surgery would be possible, particularly lengthy operations. Many sophisticated surgical procedures were developed because patients could be anesthetized, thus allowing the surgeon to carry out complex procedures. Interestingly, one of the main causes of the slow evolution of surgery in the past was the lack of anaesthesia. When anaesthesia was developed, surgery procedures and techniques developed very quickly.

Kirsch and Sapirstein (1998) also assessed natural history effects, in order to evaluate how much of the 75% was attributable to the real placebo responses and how much to other factors. To do this, another 19 trials of psychotherapy, whereby the use of no-treatment groups is more common, were analysed. Natural history accounted for the 23.87%, drug effect for 25.16%, and placebo effect for 50.97%. Therefore, in clinical trials for major depression, one-quarter is due to the specific action of the active medication, one-quarter is due to other factors, like spontaneous remission, and one-half is a placebo effect. Despite the many controlled clinical trials that have been performed in the past, the real effectiveness of many antidepressant treatments still remains an open question, and the underlying mechanisms are still virtually unknown.

Therefore, although only 0.3% of animals are used for biomedical research, up to 63% of the literature criticizes animal research. In Germany, nine animals per citizen are killed yearly for food, whereas 0.03 animals per citizen are killed for scientific purposes. Nevertheless, 80% of the time devoted to coverage of animals by the mass media deals with biomedical research (Singer 1993).

The following example is interesting and instructive and shows the greater emotional impact of biomedical animal research compared with other situations. Ehinger (1986) describes a case of rod and bait fishing. A sports fisherman caught a 16 kilogram river salmon and this took him a full hour to land the fish. A local newspaper depicted him proudly showing his catch. Most interesting, the article approvingly noted that the fisherman was exhausted by the long fight and that he was going to return for more fishing at the same river in coming years. Ehinger (1986) points out how this appealing little piece of local news can very easily be changed into a quite appalling story without changing anything about the actual event. Change the man’s profession from fisherman to scientist and the headline would read ‘Scientist struggled full hour to catch and kill his experimental animal’: The public reaction to such a title would certainly be strong. In fact, whereas hunting and catching animals for pleasure are activities which are acceptable to society, even if they involve prolonged pain and agony to the prey, animal use in biomedical research is readily condemned; it can hardly be viewed as a ‘pleasurable’ activity, no matter what its benefits to mankind. What is important in this example is that most people are not greatly concerned with the real pain and agony of the animal, because the two situations are exactly the same from the point of view of the animal, but rather with human side of the activity.

Expectations about positive therapeutic outcomes may activate specific brain regions, and such expectations may sometimes be the real cause of the clinical improvement. This is shown by some clinical trials in which expectations were assessed. Bausell and colleagues (2005) conducted a study on acupuncture in which patients were asked which group they believed they belonged to (either placebo or real acupuncture), and found that those patients who believed they belonged to the real treatment group experienced larger clinical improvement than those patients who believed they belonged to the placebo group.

In a study, experimentally infected mice were allowed to consume food ad libitum, whereas others were force-fed to the level of uninfected controls. Infected mice that regulated their own intake ate only 58% as much as the controls, and showed longer survival compared to mice that were force-fed (Murray and Murray 1979).

Large placebo responses were as. sociated with greater dopamine and opioid activity in the nucleus accumbens. Interestingly, nocebo responses, which are opposite to placebo responses, were associated with a deactivation of dopamine and opioids. The release of dopamine in the nucleus accumbens accounted for 25% of the variance in placebo analgesic effects. Therefore, placebo and nocebo effects seem to be associated with opposite responses of dopamine and endogenous opioids in a distributed network of regions that form part of the reward and motivation circuit.

Willis and Todorov (2006) investigated the minimal conditions under which people make such inferences. To do this, these authors manipulated the exposure time of unfamiliar faces, and found that judgments made after a 100-ms exposure were highly correlated with judgements made in the absence of time constraints, thereby suggesting that this exposure time was enough for subjects to form an impression. Five judgements were tested: trustworthiness, attractiveness, likeability, competence, and aggressiveness. When the exposure time was increased from 100 ms to 500 ms, subjects’ judgements became more negative, response times for judgements decreased, and confidence in judgements increased. When the exposure time was increased from 500 ms to 1000 ms, overall trait judgements and response times did not change significantly, but confidence increased for some of the judgements. Therefore, additional time may simply boost confidence in judgements. Interestingly, trustworthiness judgements showed the highest correlation, supporting the notion that people can be especially efficient in making inferences of trustworthiness because, from an evolutionary perspective, the detection of trustworthiness is essential for survival (Cosmides and Tooby 1992). This very short period of time, one-tenth of a second, for trustworthiness judgements shows that face exploration is not necessary, and the first impression of whether or not an individual is deemed to be trustworthy does not requires eye exploratory movements. In fact, a time lag of 100 ms is not sufficient for saccadic eye movements (Todorov 2008).

In the light of the importance of these subtle differences in verbal communication between doctors and patients, Pollo and colleagues (2001) carried out a similar study in the clinical setting in order to investigate the differences between the double-blind and the deceptive paradigm. Postoperative patients were treated with buprenorphine on request for three consecutive days, and with a basal infusion of saline solution. However, the symbolic meaning of this saline basal infusion varied in three different groups of patients. The first group was told nothing (natural history or no-treatment group), the second was told that the infusion could be either a potent analgesic or a placebo (classic double-blind adminis-tration), and the third group was told that the infusion was a potent painkiller (deceptive administration). The placebo effect of the saline basal infusion was measured by recording the doses of buprenorphine requested over the three-day treatment. It is important to stress once again that the double-blind group received uncertain verbal instructions (It can be either a placebo or a painkiller. Thus we are not certain that the pain will subside?), whereas the deceptive administration group received certain instructions (It is a painkiller. Thus pain will subside soon.). A decrease in buprenorphine intake was found with the double-blind administration and even more with the deceptive administration of the saline basal infusion. The reduction of buprenorphine requests in the double-blind group was as large as 20.8% compared with the natural history group, and the reduction in the deceptive administration group was even larger, reaching 33.8%. It is important to point out that the time course of pain was the same in the three groups over the three-day period of treatment. Thus the same analgesic effect was obtained with different doses of buprenorphine.

Another important example of the powerful effects of verbal communication is represented by hypnosis. ‘Through hypnotic verbal suggestions it is possible to obtain impressive effects, such as reduced colour-naming conflict in a Stroop task with an associated decrease in functional magnetic resonance imaging signal in the anterior cingulate cortex and visual areas (Raz et al. 2005, 2007), and activation of colour-processing regions following the verbal suggestions to see black-and-white pictures in colour (Kosslyn et al. 2000). In add-ition, in highly hypnotizable subjects, verbal suggestions to perceive pain in a hand that was not touched by any means activates part of the pain-processing circuitry (Raij et al. 2005). Placebo and nocebo verbal suggestions represent another example of the powerful effects of the doctor’s words,

A study investigating the biological effects of hand-holding was performed on married women who were subjected to the threat of electric shock in three different conditions: while holding their husband’s hand, while holding the hand of an anonymous male experimenter, or holding no hand at all (Coan et al. 2006). Whereas holding the spouse’s hand produced a decrease in unpleasantness ratings compared to no hand-holding, holding the stranger’s hand did not decrease unpleasantness (Figure 5.11A).

responsible for their stigmatized condition (in-fected with AIDS as a result of intravenous drug use). It was found that the participants vere significantly more sensitive to the pain of AIDS transfusion targets as compared with heathy and AIDS drug targets, together with greater haemodynamic activity in areas associated with pain processing, such as the right anterior insula, anterior midcingulate cortex, and periaqueductal grey. Conversely, lower activity was observed in the anterior mideingulate cortex for AIDS drug targets as compared with healthy controls. Interestingly the more the participants blamed these targets, the less pain they attributed to them as compared with healthy controls. In another study by Xu and colleagues (2009)

As emphasized by Hein and Singer (2008), it is important to note that empathy is distinguished from sympathy or empathic concern or compassion (see also Batson et al. 2007; Eisenberg 2007). An emotion produced by empathy is isomorphic with the other’s emotion. This is not necessarily true for sympathy or compassion (Eisenberg 2007). Nor is empathy necessarily linked to prosocial motivation, namely, the concern about the others’ well-being. By contrast, prosocial motivation is involved in both sympathy and compassion. In fact, compassion enables individuals to enter into and maintain relationships of caring and tends to motivate us to help people who are emotionally suffering, thus it represents an important aspect of prosocial behaviour. Furthermore, it has also been known that a positive intrinsic reward feeling may occur as a result of experiencing compassion for others (Sprecher and Fehr 2006). Therefore, having compassion towards a sad person goes beyond the simple sharing of the person’s sadness, involving motivational and reward components that lead to prosocial behaviours.

The positive effects of a good relationship are of particular relevance in psychotherapy. Indeed, several authors claim that psychotherapy works only through a benign human re. lationship. For example, as we will see again in section 12.5, it is worth noting that there are more than 400 types of psychotherapy, each with its theory and working hypothesis and, sur prisingly, all of them effective (Parloff 1986; Moerman 2002). In a widely known and influential review of about 40 studies that used different psychotherapeutic approaches, Luborsky and colleagues (1975) found that all the psychotherapies were effective, even those in which minimal treatment was carried out. In a different analysis of 375 studies of different kinds of psychotherapy by Smith and Glass (1977), negligible differences were found in the effects produced by different therapy types. In 1982, Landman and Dawes (1982) went through these 375 studies and found that many of them did not have random allocation of patients to the different groups. Therefore, they re-analysed the 375 studies plus 60 they added, and were able to extract 42 studies that used true random assignment. Surprisingly, Landman and Dawes (1982) found that the results were similar to those of the 375 studies analysed by Smith and Glass (1977), which indicates that it did not matter whether the studies had a true random assignment: psychotherapy was effective regardless of its theories. Therefore, all psychotherapies work more or less pretty well and there are little differences across the dif ferent therapeutic approaches. In other words, psychotherapy might be nothing more than good human interaction between patient and therapist, so that trust, belief, expectation, mo-tivation, and hope, that are common in all types of psychotherapy, would be the factors responsible for the successful therapeutic outcomes.

Indeed, words and attitudes by the provider may induce nega. tive expectations in the patient and may lead to clinical worsening. Interestingly, the primary reason for lawsuits in the United States is not medical injury itself but the failure of communication between doctors and their patients (Beckman et al. 1994; Levinson 1994), thereby pointing out the key role of a good patient-provider interaction.

In a study, the meaning of pain was changed from negative to positive in healthy subiects through verbal suggestions (Benedetti et al. 2013). The subjects had to tolerate ischaemic arm pain as long as they could. However, whereas one group was informed about the aver. sive nature of the task, as done in any pain study, a second group was told that the ischaemia would be beneficial to the muscles, thus stressing the beneficial nature of the pain endurance task. In this latter group, pain tolerance was significantly higher compared to the first one, an effect that was partially blocked by the opioid antagonist naltrexone alone and by the can-nabinoid antagonist rimonabant alone. However, the increased tolerance was antagonized completely by the combined administration of naltrexone and rimonabant, which suggests that a positive approach to pain reduces the global pain experience through the co-activation of the opioid and cannabinoid systems. These findings show that the way patients interpret their own symptoms may have a dramatic effect on their emotional experience.

In their book The Powerful Placebo: From Ancient Priest to Modern Physician, Shapiro and Shapiro (1997a) assert that the history of medical treatment is essentially the history of the placebo effect (see also Shapiro and Shapiro 1997b). Indeed, by considering the very definition of placebo as an ineffective treatment for the symptom or disorder being treated, most of the therapies that were developed over the past centuries were actually placebos.

In the first half of 1800s, Armand Trousseau tested the efficacy of homeopathy, a novel therapeutic approach introduced by Samuel Hahnemann based on the belief that a disease can be cured by very small amounts of the same substances that cause it. Trousseau probably used the first inert substances (placebos) in the history of medicine for assessing the effectiveness of a medical treatment (Trousseau and Gouraud 1834; Kaptchuk 1998). In fact, he used bread pills and told the patients that they were a homeopathic treatment. What the investigators found was a positive effect of bread pills and they attributed this to the natural course of disease and to imagination.

Placebos are traditionally used in clinical practice to please or placate anxious and complaining patients. Indeed, the word placebo (‘I shall please’) was introduced in the medical dictionary to mean an inert treatment that is given to patients more to please than to cure (de Craen et al. 1999; Wall 1999).

For example, in some indigenous tribes of Australia, pointing a bone at someone may induce negative outcomes; and in Latin America and Africa, someone believing that he is bewitched may result in a sort of ‘voodoo’ death (Cannon 1942). Although many of these phenomena are certainly anecdotal (Lewis 1977), it is not surprising that Cannon (1942) explains them as a stress-induced activation of the sympathetic nervous system. Similar explanations were put forward by Lex (1974), who interprets some forms of voodoo death as the imbalance between the sympathetic and parasympathetic nervous system.

If someone takes a placebo just before his symptom starts decreasing, he may believe that the placebo is effective, although that decrease would have occurred anyway. Clearly, this is not a placebo effect but a misinterpretation of the cause-effect relationship, due to a spontaneous remission of the symptom.

Expectations may also enhance the identification of somatic information. Indeed, Geers and colleagues (2011) studied three expectation groups. In the first group, participants ingested a placebo and were told it was caffeine (deceptive expectation). In a second group, participants ingested a placebo and were told it may be either caffeine or placebo (double. blind expectation). Participants in the third group were given no expectation. All participants then tallied the placebo-relevant and placebo-irrelevant sensations they experienced during a seven-minute period. The deceptive expectation group identified more placebo-relevant sensations than placebo-irrelevant sensations, whereas the no-expectation group identified more placebo-irrelevant sensations than placebo-relevant sensations. The double-blind expectation group identified an equal amount of placebo- relevant and irrelevant sen-sations. The amount of both placebo-relevant and placebo-irrelevant sensations detected mediated the relationship between the expectation manipulation and subsequent symptom reports. Therefore, in this study, expectations caused placebo responding, in part, by altering how one identifies bodily sensations.

Subjects who suffer from a painful condition, such as a headache, and who regularly consume aspirin, can associate the shape, colour, and taste of that pill to a decrease in their pain. After repeated associations, a sugar pill that looks like aspirin can also decrease their pain.

If the placebo is administered after two prior administrations of an effective painkiller, the placebo analgesic response is much larger (Amanzio and Benedetti 1999), thus indicating that the placebo effect is a learning phenomenon. This holds true for non-pharmacological treatments as well (André-Obadia et al. 2011).

As pointed out many times (Ader 1997; Siegel 2002; Enck et al. 2008), many placebo effects can be explained in the context of conditioning theories. In fact, a placebo is by definitiona neutral stimulus with no therapeutic effects, in the same way that a conditioned stimuls is by definition neutral. Likewise, a placebo response is by definition elicited by a neutrad stimulus, in the same way that a conditioned response is induced by a neutral stimulus.

In a different study, Peciña and team (2013) assessed psychological traits in healthy controls as to their capacity to predict placebo analgesic effects, placebo-induced activation of H-opioid neurotransmission and changes in cortisol plasma levels during an experimental pain challenge with and without placebo administration. The results showed that an aggregate of scores from Ego-Resiliency, Altruism, Straightforwardness (positive predictors), and Angry Hostility (negative predictor) scales accounted for 25% of the variance in placebo analgesic responses.

In a study, placebo analgesia has been found to be enhanced by oxytocin, an hormone known to be involved in empathy, trust, and social learning (see Box 5.1). Subjects who received intranasal administration of oxytocin showed larger placebo analgesic responses compared to controls, suggesting an important role of prosocial behaviour in the placebo effect and the possibility to exploit this effect in clinical practice (Kessner et al. 2013).

If an injection of saline solution is performed in a rat that had previously received an injection of scopolamine, the saline solution can mimic the motor changes produced by scopolamine (Herrnstein 1962).

More recently, open-label placebos better emphasize this concept, namely that patients’ expectations are not always necessary to induce an effect, as the mere ritual of administration, along with the instructions it is fake, may lead to a clinical improvement. Although larger samples and more clinical trials are necessary to confirm the effectiveness of open-label placebos (Charlesworth et al. 2017), these effects have been shown for irritable bowel syndrome (Kaptchuk et al. 2010), depression (Kelley et al. 2012), pain (Carvalho et al. 2016; Locher et al. 2017), itch (Meeuwis et al. 2018), and cancer-related fatigue (Hoenemeyer et al. 2018).

Box 8.6 Movies and open-label placebos Cinema represents an excellent example that helps understand how and why open-label placebos may work. In cinema everything is fake, none the less movies induce emotions and physiological responses. ‘The crucial point here is that the viewer knows everything is fake. A horror movie is a straightforward example, whereby the viewer does know the victims are actors, blood is tomato juice or a dye, and knives and blades are made of plastic. In spite of this knowledge and awareness, the viewer is scared and shows plenty of physiological responses, such as heart rate increase, sweating, and shivering. A romantic movie is no different as it can induce powerful emotional responses despite the viewer knowing it is entirely fictitious. Likewise, a healing ritual may induce powerful emotional and physiological responses (placebo effects) even though the patient knows everything is fake.

The reasons why social cooperation and, in general, the social environment affects the health status of the single individual are many. First, according to Cassel (1976), the social environment alters host susceptibility by affecting neuroendocrine function, and the most feasible and promising interventions to reduce disease are to improve and strengthen the social supports rather than reduce the exposure to stressors. Second, political and economic processes in different societies are also involved in the health status of the population, and this may lead to social inequalities in health (Krieger 2001). Third, the so-called neighbour-hood effect has emerged as a potentially relevant group or context effect whereby neighbour-hoods possess both physical and social attributes which could plausibly affect the health of individuals (Diez-Roux 2007).

characteristics have reported what has been called a ‘group density’ effect on health, such that members of low-status minority communities living in an area with a higher proportion of their own racial or ethnic group tend to have better health than those who live in areas with a lower proportion. In addition, members of ethnic minorities who live in areas where there are few like themselves are likely to be materially better off than those who live in areas with a higher concentration. However, they may be made more aware of belonging to a low-status minority group by the majority community, and the psychosocial effects of stigma may offset any advantage. Therefore, the psychological effects of stigma are sometimes powerful enough to override material advantage (Pickett and Wilkinson 2008).



Suppression of physical and psychological discomfort by general social interactions (with the members of the group) and by specific interactions (with a single ‘specialized’ member, the shaman) warrants a powerful means to recover, at least in part, from illness. From a natural selection perspective, only those social groups in which shamans were present and trust/hope were the key elements of the members warranted well-being and longer survival.

Therefore, the healer is the environmental variable that triggers those endogenous mechanisms of self-cure, for example the release of endogenous opioid painkillers. This is similar to the above example of stress-induced an-algesia: being chased by a lion is the environmental variable that triggers the endogenous painkilling systems. Within the context of Humphrey’s conceptualization of costs and benefits (Humphrey 2002), there are thus benefits in the doctor-patient encounter, because it triggers plenty of beneficial mechanisms, such as positive expectancies, hope, and related neurochemical systems. In other words, these endogenous mechanisms of healing are based on, and actually triggered by the figure of the doctor.

Wall (1999) claimed that pain is a need state, thus it can be terminated by specific consummatory acts, like hunger and thirst. According to Wall (1999), the consummatory acts that terminate pain can be either withdrawing one’s hand from a noxious stimulus or care and attention from others. It is this purely social event that represents the evolutionary novelty in mankind (Evans 2002). For example, a person whose brain is capable of shutting down pain when the presence of medical help is detected may have an advantage over someone whose brain lacks this capacity (Evans 2003).

In a more recent analysis of five studies in which 130 patients received a placebo treat. ment, 7% to 37% of patients in the placebo group showed greater than 50% of the maximum possible pain relief (McQuay et al. 1995).

The efficacies of rizatriptan labelled as placebo and placebo labelled as rizatriptan were similar. Therefore, increasing positive information incrementally boosted the efficacy of both placebo and medication in headache, which indicates that the information provided to patients and the ritual of pill taking are important components of care.

In another study (Benedetti et al. 2003b), one group of subjects was pharmacologically preconditioned with ketorolac (a non-opioid analgesic) for two days in a row, then the ketorolac was replaced with a placebo on the third day along with verbal suggestions of an-algesia. This procedure induced a strong placebo analgesic response. In order to see whether this response was due to the pharmacological preconditioning, a second group of subjects was also preconditioned with ketorolac, but the placebo was given on the third day along with verbal suggestions that the drug was a hyperalgesic agent. These verbal instructions were sufficient not only to block placebo analgesia completely but also to produce hyper-algesia. This study clearly shows that placebo analgesia depends on expectation of a decrease in pain, even though analgesic preconditioning is performed.

In postoperative pain following the extraction of the third molar (Levine et al. 1981; Levine and Gordon 1984), a hidden intravenous injection of a 6-8 mg morphine corresponded to an open intravenous injection of saline solution in full view of the patient (pla-cebo). In other words, telling the patient that a painkiller was being injected (with what was actually a saline solution) is as potent as 6-8 mg of morphine. The investigators concluded that an open injection of morphine in full view of the patient is more effective than a hidden one because in the latter the placebo component is absent.

The difference between open and hidden injections was investigated in a laboratory setting using the experimental model of ischaemic arm pain in healthy volunteers (Amanzio et al. 2001). As in the clinical setting, a hidden injection of the non-opioid painkiller, ketorolac, was less effective than an open one. Interestingly, when the opioid antagonist, na-loxone, was added to the open injection of ketorolac, the effect was the same as that produced by a hidden injection. This suggests that an open injection in full view of the patient activates endogenous opioids that enhance the effects of the injected painkiller.

it was found that proglumide potentiated the placebo analgesia, and presented a novel and indirect way to test the opioid hypothesis (Benedetti et al. 1995; Benedetti 1996).

In fact, if the placebo response is induced by means of strong expectation cues, it can be blocked by the opioid antagonist naloxone. Similarly, if a placebo is given after repeated administrations of morphine (preconditioning procedure), the placebo response can be blocked by naloxone. Conversely, if the placebo response is induced by means of prior conditioning with a non-opioid drug, it is naloxone-insensitive (Amanzio and Benedetti 1999).

If mice were conditioned with morphine, placebo analgesia was completely antagonised by naloxone, whereas if mice were conditioned with aspirin, placebo analgesia was naloxone-insensitive, as already found in humans by Amanzio and Benedetti (1999).

On the basis of these considerations, Benedetti and team (2011b) induced opioid or nonopioid placebo analgesic responses and assessed the effects of the CB1 cannabinoid receptor antagonist rimonabant. Differently from naloxone, rimonabant had no effect on opioid-induced placebo analgesia following morphine pre-conditioning, whereas it completely blocked placebo analgesia following non-opioid pre-conditioning with the non-steroid anti-inflammatory drug (NSAID) ketorolac.

They found that placebo responsiveness was related to the activation of dopamine in the nucleus accumbens, a region involved in reward mechanisms (Ikemoto and Panksepp 1999; Schultz et al. 2000; Schultz 2002; Knutson and Cooper 2005), as assessed by using in vivo receptor-binding positron emission tomography with raclopride, a D2/D3 dopamine receptor agonist. The very same subjects were then tested with func-tional magnetic resonance imaging for monetary responses in the nucleus accumbens. What these investigators found was a correlation between the placebo responses and the monetary responses; the larger the nucleus accumbens responses to monetary reward, the stronger the nucleus accumbens responses was to placebos.

Large placebo responses were associated with greater dopamine and opioid activity in the nucleus accumbens. Interestingly, nocebo responses were associated with a deactivation of dopamine and opioid release. The release of dopamine in the nucleus accumbens accounted for 25% of the variance in placebo analgesic effects. Therefore, placebo and nocebo effects seem to be associated with opposite responses of dopamine and endogenous opioids in a distributed network of regions that form part of the reward and motivation circuit.

On the basis of these considerations, Benedetti and others (2006b) studied Alzheimer patients at the initial stage of the disease and after one year in order to see whether the placebo component of the therapy was affected by the disease. The placebo component of the analgesic therapy was found to be correlated with both cognitive status and functional connectivity among different brain regions, according to the rule ‘the more impaired the prefrontal connectivity, the smaller the placebo response.

be quite powerful, and sometimes pain can be generated from a non-painful stimulus. For example, a study by Colloca and others (2008a) used a nocebo procedure, in which verbal suggestions of painful stimulation were given to healthy volunteers before administration of either tactile or low-intensity painful electrical stimuli. This study showed that these anxiogenic verbal suggestions were capable of turning tactile stimuli into pain, as well as low-intensity painful stimuli into high-intensity pain.

Colagiuri and Quinn (2018) found that nocebo hyperalgesia persists longer than placebo analgesia and this difference can be explained on the basis of elevated autonomic arousal.

Therefore, observation and social interaction is important in the overall placebo/nocebo phenomenon, and some personality traits may play a key role, because a correlation was found between the magnitude of the nocebo hyperalgesic responses and empathy and catastrophizing.

In addition, Swider and Babel (2013) found a gender effect, whereby nocebo hyperalgesia was greater after a male model was observed compared to a female model, regardless of the sex of the experimental subjects.

cyclooxygenase, that is prostaglandins and thromboxane (Benedetti et al. 2014; see also section 19.2). In this experimental model, a subject (the trigger) receives negative information about the risk of headache at high altitude, and disseminates this negative information across a number of other subjects. In one week, this negative information propagated across 36 subjects (Figure 10.11). This nocebo group showed a significant increase in headache and salivary prostaglandins and thromboxane when at high altitude compared to the control group. In this novel experimental model, negative information propagated across 36 subjects in one week (Benedetti et al. 2014). Over longer periods of time, hundreds or even thousands of subjects might get ‘socially infected’, thus emphasizing the possible important role of negative social communication in the dissemination of symptoms and illness across the population.

It is worth noting that this interindividual communication may have a crucial role in the outcome of a clinical trial. In fact, the same authors (Benedetti et al. 2014) ran two ‘aspirin versus placebo’ clinical trials at high altitude for the control of high altitude headache. The first trial was performed in the control subjects, whereas the second trial was performed in the socially infected individuals. Aspirin was effective in reducing both pain and prostaglandins synthesis in the control subjects, whereas placebo was totally ineffective. Conversely, both aspirin and placebo reduced pain and prostaglandins in the socially infected individuals. In other words, whereas no placebo effect was present in the controls, a placebo effect occurred in the socially infected subjects, a difference that is attributable to the differem baseline levels of prostaglandins induced by the spread of negative information. The placebo effect occurred only in the socially infected individuals because the placebo acted only on the nocebo component of the prostaglandins and pain increase.

worsening. It was found that proglumide prevented nocebo hyperalgesia in a dose-dependent manner, even though it is not specifically a painkiller, thus suggesting that the nocebo hyperalgesic effect is mediated by CCK.

responses in some brain regions. Overall, negative expectations may result in the amplification of pain (Koyama et al. 1998; Price 2000; Dannecker et al. 2003) and several brain regions, like the anterior cingulate cortex, the prefrontal cortex and the insula, have been found to be activated during the anticipation of pain (Chua et al. 1999; Hsieh et al. 1999; Ploghaus et al. 1999; Porro et al. 2002, 2003; Koyama et al. 2005; Lorenz et al. 2005; Keltner et al. 2006).

nocebo effects have also been found to be associated with a decrease in dopamine and opioid activity in the nucleus accumbens, thus underscoring the role of the reward and motivational circuits in nocebo effects as well (Scott et al. 2008).

All domains of Parkinsonism were subject to the placebo effect, but bradykinesia and rigidity were more susceptible than tremor, gait, or balance.

Substantial improvements in patients who receive placebo are also present in surgical treatments of Parkinson’s disease. For example, in a study on the effect of intrastriatal implantation of foetal porcine ventral mesencephalic tissue to treat Parkinson’s disease (Watts et al. 2001), the degree of motor performance improvement at 18 months was substantial in both the real surgery group and the sham surgery group. In one multicentre randomized double-blind sham-surgery-controlled study of human foetal transplantation (Olanow et al. 2003) there was no difference between the transplant and the sham surgery group, although pilot studies that used an identical technique had demonstrated substantial benefit (Hauser et al. 1999).

In another study by Benedetti and others (2003), patients implanted for deep brain stimulation were tested for the velocity of movement of their right hand according to a double-blind experimental design in which neither the patient nor the experimenter knew if the stimulator was turned off. The velocity of hand movement was assessed with a movement analyser (Figure 11.1A). The stimulator was turned off several times (at four weeks and two weeks) before the test session (Figure 11.1C). Each time, the velocity of movement was measured just before the stimulator was turned off and 30 minutes later. On the day of the experimental session, the stimulator was kept on but patients were told it had been turned off, so as to induce negative expectations of motor performance worsening (nocebo procedure). Although the stimulator was on, motor performance worsened and mimicked the worsening of the previous days (Figure 11.1C). This nocebo bradykinesia could be prevented completely by verbal suggestions of good motor performance (placebo procedure). Therefore, as with pain, motor performance can be modulated in two opposite directions by placebos and nocebos, and this modulation occurs on the basis of positive and negative expectations about motor performance.

Indeed, fatigue has been found to respond to placebo treatments much better than other symptoms, such as pain. For example, in hypoxic conditions at high altitude, differently from headache pain, it is not necessary to perform a preconditioning procedure with oxygen to obtain robust placebo responses, verbal suggestions alone being sufficient (e.g. Benedetti et al. 2018 and section 19.2.2).

To resolve the issue of subjective versus objective measurements, Fratello and col. leagues (2005) compared subjective, behavioural, polysomnographic, and quantitative electroencephalographic measurements after administration of a placebo (two 50 mg lactose pills along with the verbal suggestion that they were hypnotics). This study was designed to compare no-treatment and placebo within the same individual. These investigators found that their nights were self-rated as more restful, and the number of nocturnal awakenings decreased. In addition, there was an improvement in behavioural tests (e.g. reaction times) on the morning following the placebo night. This study also showed specific electroencephalographic changes following placebo administration. In particular, there was an increase of 0.5-4 Hz power during non-rapid eye movement (REM) sleep and a decrease in the -frequency band during REM sleep. These changes occurred only at central sites.

These are only a few of examples of many neurological conditions in which placebo treatments have been found to induce improvements. As control groups are not included in these neurological trials, real placebo effects are not guaranteed; certainly many improvements can be attributable to the natural history of the disease and/or to selection biases (regression to the mean). Interestingly, Papadopoulos and Mistikostas (2010) found a high rate of nocebo responses in multiple sclerosis. They conducted a systematic search for all randomized, placebo-controlled multiple sclerosis trials published between 1989 and 2009, and found that no-cebo responses are substantial and appear to have increased significantly in recent years. In addition, nocebo responses were found to exhibit an association with medication and trial-related factors.

The response rate in the placebo groups in antidepressant clinical trials is very high. In 1998, a meta-analysis was conducted by Kirsch and Sapirstein (1998) in 19 double-blind clinical trials which included 2318 patients. These investigators found that 75% of the response to the active drug is attributable to a placebo effect, thus the specific pharmacodynamic effect of the drug would account for only 25%. In addition, a high correlation was found between the placebo effect and the drug effect, which indicates that virtually all the variation in drug effect size was due to the placebo component. In the same study, Kirsch and Sapirstein (1998) also assessed natural history effects, in order to evaluate how much of the 75% was attributable to real placebo responses and how much to other factors. To do this, another 19 trials of psychotherapy, where the use of no-treatment groups is more common, were analysed. Natural history accounted for 23.87%, drug effects for 25.16%, and placebo effect for 50.97%.

Interestingly, in Walsh’s study (Walsh et al. 2002) both the proportion of patients who responded in the placebo groups and the proportion who responded to medication were significantly positively correlated with the year of publication, with a more statistically robust correlation for placebo than medication. From this study it appears that, in the two decades between 1981 and 2000, the proportion of patients who show symptom reduction in the placebo group has increased at a rate of approximately 7% per decade, and a similar increase has occurred in the fraction of patients responding to the active drug.

In a subsequent study, Leuchter and others (2004) analysed the neurophysiological, symptomatic, and cognitive characteristics of subjects who were likely to respond to placebo in clinical trials for major depressive disorder. It was found that placebo responders had lower pre-treatment frontocentral electroencephalographic cordance in the 0-frequency band than all other subjects, particularly the medication responders. Placebo responders also showed faster cognitive processing time and lower reporting of late insomnia. A logistic regression analysis showed that these three pre-treatment measures (cordance, cognitive processing time, and late insomnia) accurately identified 97.6% of eventual placebo responders.

Compelling evidence that placebo-induced expectation plays an important role in anx-lety is shown by the hidden administration of anti-anxiety drugs. By using the open-hidden approach (see section 8.2.4) the effectiveness of diazepam, one of the most frequently used benzodiazepines for treating anxiety, was assessed after overt and covert administration in postoperative patients with high anxiety scores (Benedetti et al. 2003; Colloca et al. 2004; Benedetti et al. 2011). In the open group there was a clear-cut decrease of anxiety, but in the hidden group diazepam was totally ineffective, which indicates that anxiety reduction after the open diazepam was a placebo effect (Figure 12.4A). Open-hidden interruption of a diazepam treatment has also been investigated (Benedetti et al. 2003; Colloca et al. 2004; Benedetti et al. 2011). In the open condition anxiety increased significantly after four and eight hours; in the hidden condition it did not change (Figure 12.4B). Therefore, the anxiety relapse after the expected (open) interruption of diazepam could be attributed to the negative expectation of anxiety relapse (nocebo-like effect).

In animals, the ability of drugs to activate the reward circuits (e.g. the nucleus accumbens) is larger when they are given cocaine in a context where they had previously received it than in a context where they had not (Duvauchelle et al. 2000). Today we know some mechanisms, both psychological and biological, underlying placebo and placebo-like effects in addiction.

Smoking is a good example of how both pharmacological and non- pharmacological factors are at work in tobacco dependence. Although nicotine is necessary for maintaining tobacco dependence (Jaffe 1990, Benowitz and Henningfield 1994; Stolerman and Jarvis 1995) it is generally not sufficient. In fact, despite the fact that nicotine replacement therapy is capable of delivering as much nicotine as a cigarette (and almost as rapidly), many smokers relapse to smoking (Perkins et al. 2003).

regular coffee drinkers completed the Caffeine Withdrawal Symptom Questionnaire (CWSQ) before and after receiving decaffeinated coffee. However, whereas one-half of the participants were led to believe the coffee was regular caffeinated coffee (the ‘Told Caffeine’ condition), another one-half were told that it was decaffeinated (the ‘Told Decaf’ condition). Participants in the Told Caffeine condition reported a significantly greater reduction in the factors of cravings, fatigue, lack of alertness, and flu-like feelings than those in the Told Decaf condition.

Moore and Aubin (2012) assessed the strength of addictions by measuring cessation rates with placebo or no treatment controls, and found that a weaker addiction has a higher cessation rate than a stronger addiction. Moore and Aubin analysed several systematic reviews and meta-analyses of cessation trials, using randomized or quasi-randomized trials and reporting objectively-measured abstinence. The outcome for comparison was the percentage of participants abstinent according to an objective test of abstinence at six months or longer (quit rates). A total of 28 cessation reviews with 139,000 participants were analysed. Most of these (127,000) came from trials of smoking cessation. Cessation rates with placebo in randomized trials using objective measures of abstinence and typically over six months duration were 8% for nicotine, 18% for alcohol, 47% for co-caine, and 44% for opioids. Therefore, evidence from placebo cessation rates indicates that nicotine is more difficult to give up than alcohol, cocaine, and opioids. Although this study needs further confirmation, it represents an excellent application of placebos in a specific clinical condition, in which placebo quit rates are a useful proxy marker for the strength of different addictions.

Psychotherapy is quite interesting in the context of placebo and placebo-related effects because both psychotherapy and a placebo procedure use verbal suggestions. The old debate about whether or not psychotherapy and placebos have similar mechanisms consists of ascertaining whether psychotherapy is nothing but a placebo effect, and thus whether a placebo procedure is a very simple form of psychotherapy.

The conclusion appears to be that all psychotherapies work more or less fairly well and that there are little differences across the different therapeutic approaches. However, the explanation might be that psychotherapy is nothing more than a good human interaction between patient and therapist, so that trust, belief, expectation, motivation, and hope common elements in all types of psychotherapy–are the factors responsible for the successful therapeutic outcomes (Moerman 2002).

In this context another study is very instructive. In 1979, Strupp and Hadley (1979) conducted an interesting experiment whereby ‘disturbed’ college students were allocated to either psychotherapy carried out by practising psychotherapists with experience of more than 20 years each on average, or interaction with professors of English, philosophy, history, or mathematics with a renowned reputation for warmth and trustworthiness but with no previous experience as therapists. Both groups improved. Furthermore, there was no significant difference between the improvement of those who interacted with psychotherapists or with the professors.

The fact that all sorts of psychotherapy show negligible differences between each other (Luborsky et al. 1975; Smith and Glass 1977) and that successful outcomes can be obtained by inexperienced therapists (Strupp and Hadley 1979) suggests that factors other than specific elements of each psychotherapy are at work. This concept was expressed in the 1960s by Frank (1961) who proposed that all forms of psychotherapy work because they contain similar elements, such as a ritual to reinforce the therapist-patient relationship and the presence of a thoughtful listener. All these considerations lead to the question of what placebo and placebo-related effects have to do with psychotherapy.

Recent studies of clinical response in major depression to either cognitive-behavioural therapy (Goldapple et al. 2004) or interpersonal psychotherapy (Brody et al. 2001; Martin et al. 2001) demonstrate very different regional brain change patterns from those seen with placebo (Mayberg et al. 2002). Both cognitive-behavioural therapy and interpersonal psychotherapy are associated with substantial prefrontal decreases with other regional effects specific to each psychotherapy strategy. With clinical response to cognitive-behavioural therapy, Mayberg and colleagues described additional changes in regions not affected by antidepressant drugs, like fluoxetine, including the orbital frontal and medial frontal cortex and dorsal anterior cingulate cortex (Benedetti et al. 2005).

andler and co-workers (2010) studied 99 children with ADHD who were randomized to one of three treatments of eight-week duration: (i) conditioned placebo dose reduction condition (50% reduced dose + placebo) or (ii) a dose reduction only, or (iii) a no reduction condition. The conditioned placebo dose reduction procedure involved daily pairing of mixed amphetamine salts dose with a visually distinctive placebo capsule administered in open label, with full disclosure of placebo use to subjects and parents. Seventy children completed the study. Most subjects in the reduced dose + placebo group remained stable during the treatment phase, whereas most in the reduced dose group deteriorated. There was no difference in ADHD symptoms between the reduced dose + placebo group and the full dose group, and both groups showed better ADHD control than the reduced dose group. Side effects were lowest in the reduced dose + placebo group. This is a nice clinical application of placebo conditioning mechanisms, whereby pairing placebos with stimulant medication elicits a placebo response that allows children with ADHD to be effectively treated on 50% of their optimal stimulant dose.

•The placebo effect in the immune and endocrine system is basically a conditioned response, whereby classical, or Pavlovian, conditioning plays a key role. Indeed, both immune mediators and hormones can be conditioned by administering a drug several times in a row and then replacing it with a placebo.

An intriguing observation was reported in 1896 by MacKenzie (1896) who showed that some people who are allergic to flowers show an allergic reaction when presented with something that superficially looks like a flower but contains no pollen (an artificial flower). In this case, the association between the features of the flower (colour and shape, and the like) and the pollen antigen inside the flower may lead to a conditioned response, whereby the colour and the shape of the flower may induce a conditioned allergic response.

Opioids seem to play a role as well. Morphine decreases natural killer cell activity (Mellon and Bayer 1998, 1999; Gomez-Flores and Weber 1999) and opioids in general suppress many immune responses, like antibody production, and delayed-type hyper-sensitivity. This occurs in part through desensitization of chemokine receptors on neu-trophils, monocytes, and lymphocytes (Grimm et al. 1998; Rogers et al. 2000). Pavlovian conditioning of morphine-induced immunomodulating effects has been described (Coussons-Reed et al. 1994a). The opioid antagonist, naltrexone, when injected before conditioning, has been found to prevent acquisition of the conditioned suppression of splenic proliferation of lymphocytes, of natural killer cell activity, and of IL-2 production (Coussons-Reed et al. 1994a).

timuli that signal food, such as the sight and smell of food (Sjostrom et al. 1980; Johnson and Wildman 1983; Simon et al. 1986) and imagination or thinking about food (Goldfine et al. 1970) are all capable of inducing insulin secretion in the beta cells of the pancreas. This is called the cephalic phase of insulin release (see Stockhorst et al. 2000 for a review). Although cephalic-phase insulin secretion, like many other functions described in this book, has never been conceptualized in terms of a placebo effect, it is a good example of how contextual cues can trigger the activation of an endogenous system. Cephalic-phase insulin secretion is mediated by the vagus nerve, as it is abolished by vagotomy (Louis-Sylvestre 1976; Woods and Bernstein 1980). In addition, the blockade of vagal activity with atropine in obese humans abolishes the cephalic-phase insulin response (Sjostrom et al. 1980).

Asthma has been reported to be influenced by many psychological factors (McQuaid et al. 2000; Liu et al. 2002; Sandberg et al. 2004; Chen et al. 2006). In 1968, Luparello and team (1968) showed that administration of a placebo, along with verbal suggestions that the inhaled substance contained irritants, induced a dramatic increase in airway resistance and breathing difficulty in asthmatic patients. However, if placebo was administered again, but the patients were told it was therapeutic, airway resistance decreased and easy breathing returned.

Clinical trials comparing active treatments for cough with placebo treatments show that the difference between the two therapies is small, with medication a little more effective than placebo (Schroeder and Fahey 2001, 2002; Eccles 2002, 2006, 2009, 2010).

•Current evidence shows that cancer progression is not affected by placebo treat-ments, however symptoms can be reduced by placebos. This emphasizes the differential effects of placebos on objective and subjective outcomes in some conditions.

A first point that should be emphasized is that there is no real evidence that placebos can cure; rather, they may sometimes improve quality of life, as described throughout this book, for example for the symptom pain’ (chapter 10) and for the symptoms of Parkinson’s disease (chapter 11). There is plenty of confusion on this point and unfortunately many claim that they can cure virtually all illnesses with placebos. Hard science tells us that placebos can reduce symptoms like pain as well as muscle rigidity in Parkinson’s disease, yet the progression of the disease is not affected; for example, in Parkinson’s disease, neurons keep degenerating even though some symptoms can be reduced for a short time.

For example, Clark and colleagues (2000) investigated the placebo effect of a carbohydrate supplement on endurance performance in 43 cyclists in a simulated 40-kilometre timed trial. They were subdivided into three groups: the first received just water and was told nothing; the second ingested a drink containing 7.6 g per 100 mL of carbohydrate; and the third received a non-caloric placebo. The second and third groups were further subdivided into those who were told the drink contained nothing, those who were told it contained carbohydrate, and those who were told it contained placebo. The difference in mean power between the ‘told-carbohydrate and the ‘told-placebo’ group was 3.8%.

In a subsequent study, six cyclists in a laboratory 10-kilometre timed trial were told that they would receive either placebo or 4.5 mg/kg caffeine or 9 mg/kg caffeine, but in fact they always received placebos (Beedie et al. 2006). A dose-response effect was found, whereby the cyclists who believed they had ingested placebo showed a decrease in power of 1.4%, those who believed they had ingested 4.5 mg/kg caffeine showed an increase in power of 1.3%, and those who believed they had received 9 mg/kg caffeine showed a 3.1% increase in power. Interestingly, some caffeine- related side effects were reported. By contrast, in a study on placebo responses associated with caffeine suggestions, only a limited effect of placebo was found (Flaten et al. 2003).

in a competition of pain endurance, Benedetti and colleagues (2007) found that placebo administration on the day of competition induced longer pain tolerance compared to an untreated group (Figures 19.1A and 19.1B). However, if pharmacological preconditioning was performed with morphine in the pre-competition phase, the replacement of morphine with a placebo on the day of the competition induced an increase in pain endurance and physical performance that was significantly larger than placebo without prior morphine preconditioning (Figure 19.1C). The placebo effect after morphine preconditioning could be prevented by administration of the opioid antagonist, naloxone, which suggests that this placebo response is opioid-mediated (Figure 19.1D). These placebo analgesic responses were obtained after two morphine administrations that were separated as much as one week from each other.

A placebo, which the subjects believed to be caffeine at high doses, was administered twice in two different sessions. Each time the weight to be lifted with the quadriceps was reduced surreptitiously so as to make the subjects believe that the ‘ergogenic agent’ was effective. After this conditioning procedure, the load was restored to the original weight, and both muscle work and fatigue were assessed after placebo administration. A robust placebo effect occurred,with a significant increase in muscle work and a decrease in muscle fatigue. These findings suggest a central mechanism of top-down modulation of muscle fatigue, which seems to be more important or as important as peripheral mechanisms arising from muscles.

f

Muscle fatigue has also been found to be affected by a central governor (Box 19.1) (St Clair Gibson et al. 2003, 2006; Lambert et al. 2005). For example, St Clair Gibson and colleagues (2001) found that as fatigue increased in cycling, the electromyographic activity in leg muscles declined, even during sprint, thus indicating that fatigue is not due to muscle fibres hitting some limit. In many of the above reported studies, athletes were asked to perform at their limit, in an all-out effort. Placebos apparently acted by pushing this limit forward. Therefore, it can be speculated that they could impact on a central governor of fatigue. The output of this centre would continuously regulate exercise performance to avoid reaching maximal physiological capacity. This would provide protection against damage on one hand, and constant availability of a reserve capacity on the other (Hampson et al. 2001; Lambert et al. 2005). By altering expectations, placebos could then represent a psychological means to signal the central governor to release the brake, allowing an increase in performance in a manner not dissimilar from that achieved by pharmacological means (e.g. by amphetamines decreasing perceived fatigue).

Can oxygen be partially replaced with a placebo? Today there is experimental evidence that by using the high-altitude model of hypoxia, placebos can mimic the effects of oxygen for oxygen reductions up to 50% compared to sea level. As anticipated in the previous section, most of the evidence comes from studies at 3500 m (64% oxygen compared to sea level), 4500 m (57% oxygen), and 5500 m (50% oxygen) (Benedetti et al. 2018), as well as from high-altitude simulations in the lab at sea level (Torres-Peralta et al. 2016). These studies show that placebos can affect critical life functions like ventilation, circulation, perfusion, as well as overall performance, but not oxygenation. In fact, in no case a change in SO, was found after placebo administration (see the studies described below), which indicates that placebos cannot in any way affect blood oxygenation. This is not surprising, as blood oxygenation is due to the direct diffusion of oxygen from the lungs to the blood stream, without any involvement of the central nervous system. Indeed, it is unlikely that a placebo can change the way oxygen diffuses from air to blood. As it will be discussed in detail below, this is one of the most interesting findings of all these studies because it shows that placebos may affect many body functions without any change in oxygen content of the body.

oxygen preconditioning is crucial for a placebo respiratory effect to occur (Benedetti and Dogue 2015). Therefore, the placebo respiratory effect is basically a conditioned hypoventilation response. Second, this effect on ventilation can be seen with an oxygen reduction up to 50%.

However, SO, is not completely responsible for overall performance, at least within some limits, central nervous system mechanisms, including complex psychological factors, playing an important role. ‘These factors are related to the so-called central command’ or ‘central governor’ (see Box 19.1), which implies that central mechanisms are more important than peripheral mechanisms in the generation of fatigue.

The perceived fatigue has a powerful influence on overall performance: the time taken to complete 3000 steps with a stepper lengthens from sea level to high altitude, and performance returns to that of sea level following administration of both oxygen and placebo (Benedetti et al. 2018). Importantly, this improvement occurs when placebo is given without any oxygen preconditioning, and this ‘first-time’ placebo administration is sufficient to reduce perceived fatigue and increase performance. Therefore, the increase in SO, after oxygen delivery is not completely responsible for the decrease in time to complete 3000 steps, as performance improves also with a placebo and without any SO, increase.

Interestingly, glucose administration was found to improve recognition memory times and performance on the vigilance task, but only in sessions where subjects were informed that they would receive glucose and not when they were told that they would receive aspartame. Therefore, expectation contributes to the positive effects of glucose on cognition. In another study, Oken and colleagues (2008)

Crum and colleagues (2011) conducted a study in which participants consumed a 380-calorie milkshake under the pretence that it was either a 620-calorie ‘indulgent’ shake or a 140-calorie ‘sensible’ shake. Ghrelin (see Box 20.1) was measured via intravenous blood samples at three time points: baseline (20 min), anticipatory (60 min), and post-consumption (90 min). During the first interval (between 20 and 60 min) participants were asked to view and rate the (misleading) label of the shake. During the second interval (between 60 and 90 min) participants were asked to drink and rate the milkshake. The mindset of indulgence produced a dramatically steeper decline in ghrelin after consuming the shake, whereas the mindset of sensibility produced a relatively flat ghrelin response. Participants’ satiety was consistent with what they believed they were consuming rather than the actual nutritional value of what they consumed. This is an example of how the effect of food consumption on ghrelin may be psychologically mediated, depending on subjective expectations. This confirms a previous study, whereby underestimation of caloric intake occurs when a restaurant is perceived as healthy (Chandon and Wansink 2007).