Nociceptive input can result in highly variable sensations of pain with expectations playing a crucial role in pain modulation (Atlas and Wager, 2012). Positive expectations can lead to hypoalgesia (placebo effect), while negative expectations can increase the perceived intensity of pain (nocebo effect; Kong and Benedetti, 2014). Although many studies have shown that expectations can influence pain perception, the neuronal processes underlying the generation of expectations prior to the appearance of the pain stimulus are not yet fully understood (Benedetti, 2014; Büchel et al., 2014; Koyama et al., 2005; Wager and Atlas, 2015). Several studies have demonstrated that placebo and nocebo effects influence brain activity during pain perception (Egorova et al., 2015; Wager and Atlas, 2015; Zunhammer et al., 2021). This may occur through multiple pathways that are involved in integrating expectations and sensory information (Geuter et al., 2017). Especially the descending pain modulatory system (DPMS; Geuter et al., 2017; Tu et al., 2022) is associated with placebo and nocebo effects and is thought to consist of areas like the periaqueductal gray (PAG), dorsolateral prefrontal cortex (DLPFC), anterior cingulate cortex (ACC), and amygdala (Eippert et al., 2009), as well as frontal areas like the vmPFC (Geuter et al., 2017; Tu et al., 2022). Furthermore, placebo and nocebo effects modulate activity in areas classically associated with pain processing like the thalamus and the insula during noxious stimulation (Atlas and Wager, 2014; Wager and Atlas, 2015; Wager et al., 2004; Zunhammer et al., 2021).

A central feature of expectations is that they are generated prior to the appearance of the stimulus and should, therefore, be reflected in anticipatory neural activity (Kong et al., 2007; Wager et al., 2004). The mere expectation of the appearance of a painful stimulus has been shown to activate regions relevant to subsequent pain processing, such as the insula, DLPFC, and thalamus (Palermo et al., 2015; Ploghaus et al., 1999). Similarly, there is evidence that expecting reduced pain (e.g. via a placebo) modulates activity in parts of the DPMS and the insula already during pain anticipation (Wager et al., 2011). This includes the DLPFC (Amanzio et al., 2013; Geuter et al., 2013; Wager et al., 2004; Watson et al., 2009) and the ACC (Amanzio et al., 2013; Geuter et al., 2013) and takes place prior to the widespread modulation of neural activity in the DPMS during pain processing (Amanzio et al., 2013; Atlas and Wager, 2014). The activation of brain areas prior to the administration of a painful stimulus aligns with the general framework of placebo effects and nocebo effects proposed by Büchel et al., 2014. This framework posits that placebo hypoalgesia and nocebo hyperalgesia can be attributed to predictive coding, suggesting that perception is the result of a constant matching of incoming sensory data with the top-down predictions of an internal or generative model (Büchel et al., 2014). These top-down predictions should be reflected in the expectation generation happening before the stimulus. However, the neural mechanisms of this process remain unclear, with little information about where and how the expectations relevant to these top-down predictions are generated. Furthermore, it is uncertain whether the anticipatory modulation reported in the literature reflects only a pre-activation of later relevant networks or indicates functionally distinct processes in expectation generation. Comparing neural processing during both the anticipatory and pain periods prior to the stimulus is crucial for better dissociating the formation and maintenance of pain-related expectations from their integration with nociceptive input (Wager et al., 2004).

Another important factor when examining the representation of expectations is their valence, i.e., whether they are positive (as in placebo effects) or negative (as in nocebo effects). It remains elusive whether positive and negative expectations share a common neural basis or depend on different networks (Freeman et al., 2015). On the basis of behavioral differences between positive and negative expectations, such as their varying correlations with the amount of prior experience and differences in learning (Colloca et al., 2010; Colloca et al., 2008), it seems reasonable to assume at least partially dissimilar neural representations of positive and negative expectations. Especially during pain perception, there is evidence of distinct neural processes for positive and negative expectations, as well as valence-dependent modulation of similar systems. More specifically, some findings suggest differential modulation of activity in key areas of the DPMS and reward system by the valence of expectations (Benedetti et al., 2020; Crawford et al., 2021; Koyama et al., 2005; Scott et al., 2008). Alternatively, some studies have reported the absence of shared brain activations during the perception of pain (Bingel et al., 2011; Freeman et al., 2015; Fu et al., 2021; Shi et al., 2021; Shih et al., 2019). Similar findings have been reported during pain anticipation, as some areas have been reported to be specifically activated for either placebo or nocebo (Fu et al., 2021; Rossettini et al., 2023) or show opposing valence-dependent responses (Amanzio et al., 2013; Kong et al., 2008; Palermo et al., 2015). However, there is conflicting evidence of shared anticipatory activation for both placebo and nocebo in some areas of the DPMS like the DLPFC and ACC (Amanzio et al., 2013; Amanzio and Palermo, 2019; Colloca and Benedetti, 2007; Frisaldi et al., 2015; Manaï et al., 2019; Palermo et al., 2015; Rossettini et al., 2023; Schmid et al., 2015). This illustrates an ongoing debate about whether there is a common neural basis for positive and negative expectations instead of entirely separated representations (Freeman et al., 2015; Fu et al., 2021). Taken together, the neuronal representations of positive and negative expectations may consist of shared valence-neutral processes (common effects) and different or differentially modulated valence-dependent processes (distinct effects), and this relationship may change over time from the formation of expectations until their integration with the sensory information. To adequately examine common and distinct neuronal representations, it is imperative to compare positive and negative expectations against each other, as well as to an appropriate control condition without any directed expectation, meaning that perception is not biased in any direction in this condition. Hence, in this study, we implemented a within-subjects design in which participants were subjected to positive, negative, or neutral expectations on a trial-by-trial basis, enabling an exploration of common and distinct processes.

Even during the anticipation phase, it is reasonable to presume the involvement of distinct processes unfolding in a temporal sequence. For instance, the visual cue must first be encoded and transformed into an expectation signal that can be interpreted by a ‘pain system’. A potential candidate for this integration process is the insula, which is recognized for its function as a multimodal network hub (Adler-Neal et al., 2019; Dionisio et al., 2019; Lu et al., 2016). The multimodal role of the insula is further reflected in its role in fear conditioning, with the insula being associated with threat (Fullana et al., 2016), while other areas that are important for placebo modulations like the vmPFC are closely connected to the default mode network, potentially reflecting a safety signal (Fullana et al., 2016). Within the DPMS, prefrontal areas have been suggested to provide predictions for downstream pain-sensitive systems, implying an early role within this system (Geuter et al., 2017; Koban et al., 2017; Wager and Atlas, 2015). Subsequently, it appears plausible that other areas of the DPMS would need to be ‘informed’ and activated in close temporal proximity to the pain stimulus in order to exert their influence on pain perception (Amanzio et al., 2013). The assessment of the temporal profile of expectancy generation is beyond the possibility of fMRI. We therefore combined fMRI with simultaneous EEG measurements, allowing us to temporally localize neural activity by utilizing the temporal and spatial advantages of both techniques at the single-trial level.

In this study, we investigated the neural basis of the common and distinct processes underlying positive and negative expectations, and the formation and integration of expectations into pain perception, using a novel paradigm that allowed the manipulation of expectations on a trial-by-trial basis, while EEG and fMRI measures were recorded. We presented cues to induce expectations (positive, negative, or neutral expectations) followed by an anticipatory period in which different expectations emerged. This allowed us to examine the distinct and common effects of placebo and nocebo in the anticipation and pain phase by comparing the different expectation conditions. We focused on the evaluation of the different neuronal processes that contribute to the generation of directed expectations (i.e. positive and negative) in the anticipatory period and the effects during pain perception, using combined EEG and fMRI.

Based on the literature and the assumed theoretical approach of predictive coding, proposing an expectation formation before a stimulus, we expected that representations of pain-related expectations undergo dynamic changes during the anticipation phase and pain phase, reflecting different processes during these phases such as expectation formation, expectation integration, and pain modulation. These processes could involve either separate networks during the anticipation and pain phase or they could take place in the same networks, with the anticipatory activity having preparatory qualities for the later perception. Furthermore, different patterns of activity for positive and negative expectations could arise. If the valences do not differ in their activation patterns either during pain anticipation or pain processing, this would mark similar processes for both positive and negative expectations. On the other hand, different effects would either indicate distinct processes or a different modulation of the same process. We would mainly expect a distinct nature of the valences, but that similar areas would be engaged throughout anticipation and pain processing. Therefore, we hypothesized to see neural representations of expectations in similar areas during the anticipation and pain phase, albeit that those positive and negative valences would be differentially represented, marking dissociable dynamics of positive and negative expectations.

In total, we investigated 50 participants (32 female) in a combined EEG-fMRI paradigm. In short, participants were told that they would be given real-time visual feedback on their current pain sensitivity based on their EEG activity using a Brain-Computer Interface (BCI). The feedback indicated one of three different brain states: either a state of high pain sensitivity (red cue/nocebo condition/negative expectation), low pain sensitivity (green cue/placebo condition/positive expectation), or that the BCI algorithm would not make any prediction (yellow cue/control condition/neutral expectation). In fact, the visual cues were not related to any brain activity but were only used to produce the corresponding expectations. To reinforce expectations, we also performed a learning (i.e. conditioning) phase. Here, red cues were paired with higher pain intensity (VAS level 70), while green cues were paired with lower pain intensity (VAS level 30). In the ensuing test phase (see Figure 1), temperatures were kept constant (VAS level 60). Participants were informed that they would receive different pain stimuli of medium intensity and were unaware that the stimulation temperature was always exactly the same. In each trial, participants were given a BCI-based feedback supposedly related to their current brain state (cue phase) and subsequently rated their pain expectation for the next stimulus (expectation rating). After a fixed anticipation phase, they were presented with a brief heat pain stimulus with a constant target temperature irrespective of condition (pain phase), and lastly had to rate how intensely they perceived the stimulus (pain rating). Apart from EEG and fMRI we also continuously recorded electrodermal activity.

Figure 1

Download asset Open asset

Successful induction of placebo and nocebo effects in behavioral ratings and skin conductance responses

Our data showed successfully induced expectations in line with the cued sham brain states as evidenced by a significant main effect of condition in a repeated-measures ANOVA for mean expectation ratings (F_(2,98) = 86.51, p<0.001, η_P²=0.64; see Figure 2A). Expectation ratings were higher in the nocebo (M=65.80, SD = 15.80) compared to the control condition (M=48.58, SD = 13.79, p<0.001), which in turn were higher than in the placebo condition (M=34.33, SD=17.71, p<0.001). Similarly, mean pain ratings were affected by our manipulation (rmANOVA: F_(2,98) = 63.00, p<0.001, η_P²=0.56; see Figure 2B). Post-hoc Tukey tests revealed higher pain ratings in the nocebo (M=56.80, SD=14.21) compared to the control condition (M=51.40, SD=14.31, p<0.001), which in turn led to higher pain ratings than the placebo condition (M=44.88, SD=15.06, p<0.001). Moreover, placebo (control - placebo) and nocebo (nocebo - control) effects were significantly correlated across subjects for both expectation (r=0.64, p<0.001) and pain ratings (r=0.30, p=0.033), indicating that subjects who experienced stronger placebo effects also experienced larger nocebo effects.

Figure 2

Download asset Open asset

To assess whether ratings within the three conditions were stable or varied over time, we compared the relative variability index (Mestdagh et al., 2018), a measure that quantifies intra-subject variation over multiple ratings, between the three conditions and over the three measurement blocks. We observed differences in relative variance indices between conditions for both expectation (F(2,96) = 8.14, p<0.001) and pain ratings (F(2,96) = 3.41, p=0.037). For both measures, post-hoc tests revealed that there was significantly more variance in the placebo compared to the control condition (both p_holm<0.05), but no difference between control and nocebo. Variance in expectation ratings decreased from the first block compared to the other two blocks (F(1.35,64.64)=5.69, p=0.012; both p_holm<0.05), which was not the case for pain ratings. There was no interaction effect of block and condition for neither expectation (F(2.65,127.06)=0.40, p=0.728) nor pain ratings (F(4,192) = 0.48, p=0.748), which implies that expectations were similarly dynamically updated in all conditions over the course of the experiment.

The expectation manipulation not only affected behavioral ratings but also the skin conductance responses (SCRs) to the pain stimuli (rmANOVA: F(2,50) = 4.33, p=0.018, η_P²=0.15; see Figure 2C). A post-hoc Tukey test showed larger SCRs in the nocebo (M=0.05, SD=0.09) compared to the placebo condition (M=–0.04, SD=0.10, p=0.049). SCRs in the control condition (M=–0.01, SD=0.07) did not significantly differ from neither the nocebo (p=0.072) nor placebo condition (p=0.607).

Successful induction of expectation effects in fMRI pattern

Induction of expectation effects was also tested in functional imaging data. For all fMRI analyses, a finite impulse response (FIR) model was used to characterize BOLD fluctuations over time from cue onset to the pain rating (see Methods for details). The stimulus intensity independent pain signature (SIIPS) has been introduced as a marker for subjective pain perception going beyond intensity differences as it has been reported to be affected by psychological factors such as expectations (Botvinik-Nezer et al., 2023; Woo et al., 2017). To further validate our experimental design, we estimated the SIIPS score for each condition during the pain phase as a marker for differences in pain perception between the three conditions (see Methods). A rmANOVA revealed significant differences between the three conditions within the pain period (F(2,88) = 11.59, p<0.001, η_P²=0.21), with Bonferroni-corrected paired t-tests showing significant differences between the placebo and nocebo condition (t₍₄₄₎ = 4.79, p<0.001) and between the nocebo condition and the control condition (t₍₄₄₎ = 3.36, p=0.002) but not between the control and the placebo condition (t₍₄₄₎ = 1.35, p=0.184, see Figure 3). We therefore conclude that the manipulation of expectations led to significant perceptual differences. Contrastingly, the SIIPS signature failed to discern between conditions during the anticipation period, suggesting fundamentally distinct processes in the two phases (F(2,88) = 0.79, p=0.455, η_P²=0.02).

Figure 3

Download asset Open asset

Neuronal representation of expectations over time

In our main analysis, we found a clear dissociation between the anticipation and pain phases with a predominantly common representation of positive and negative expectations during the anticipation phase and a later shift towards distinct effects during the pain phase (see Figure 4A). In order to investigate how the representations of directed expectations changed over time from the anticipation to the pain phase, we identified common (i.e. positive and negative vs. control; constrained to areas with no statistical difference between positive and negative) and distinct (positive vs. negative) effects of directed expectations in each phase (see Supplementary file 1a-d for all comparisons). During the anticipation phase, common effects of directed expectations were found in several important areas of the DPMS, e.g., in the bilateral DLPFC, bilateral ACC, and right vmPFC, indicating that directed expectations were represented in a rather general and nonspecific way during this period. With the stimulus onset, activity in these areas showed differential activation between positive and negative expectations. Further differential activity was observed in e.g., the left insula, amygdala, thalamus, and hippocampus during the pain phase.

Figure 4

Download asset Open asset

Crucially, the bilateral DLPFC, right vmPFC, left anterior insula, and thalamus were engaged during both the anticipation and pain phases (see Figure 4B). In all these areas, directed expectations shifted from a common (positive and negative > control) towards a distinct representation (positive > negative: bilateral DLPFC, right vmPFC, and left anterior insula; negative > positive: thalamus) over time. In addition to these areas that are frequently related to expectation effects, the right angular gyrus was also engaged throughout the time course, similarly initially showing a common representation of positive and negative expectations during the anticipation phase and differentiating only during pain perception (p<0.05 whole-brain FWE-corrected; see Figure 5).

Figure 5

Download asset Open asset

The differences between the anticipation and the pain phase demonstrate that specific expectations were mediated by different processes during these phases and arise from a dynamic interplay of brain regions such as the DLPFC, vmPFC, anterior insula, and thalamus over time.

Timing of effects during the anticipation phase

To obtain detailed information on the temporal characteristics of the expectation effects during pain anticipation, we performed fMRI-informed EEG analyses. Specifically, we were interested in the temporal sequence of the areas involved. Single-trial estimates of fMRI activity during the anticipation phase were correlated with time-frequency decomposed EEG measures for each participant and then statistically tested at the group level. This analysis was conducted separately for the identified regions of interest that represented directed expectations during both the anticipation and pain phase (left anterior insula, right vmPFC, bilateral DLPFC, and thalamus). We further included the bilateral ACC, as we were interested in all areas that reflected directed expectations during the anticipation phase.

Clear temporal differences between areas were observed. The earliest correlation of fMRI activity with EEG oscillatory power was found in the left DLPFC already during the expectation rating at theta to low gamma frequencies in a negative direction (–6.3 until –4.4 s, 4–45.28 Hz; p=0.007; see Figure 6A; additional information for all regions can be found in Supplementary file 1e). Next, anticipatory activity in the left anterior insula was associated with decreased EEG oscillatory power during the late expectation rating and anticipation phase (−4.4 to –2.5 s) spanning from theta to low gamma frequencies (5.67–32 Hz; p<0.001; see Figure 6B). Lastly, we found a significant positive association of EEG activity with fMRI activity in the right ACC during the anticipation and early pain phase in the gamma frequency range (−3 to 0.3 s, 26.89–128 Hz; p=0.003; see Figure 6C). As expected due to the close spatial proximity, activity in the left ACC was similarly positively correlated with EEG activity in the anticipation and pain phase (cluster 1: –2.3 to –1.3 s, 32–128 Hz; p=0.019; cluster 2: –0.9 to 0.3 s, 32–128 Hz; p=0.021). We did not observe significant correlations of EEG power with fMRI activity in the other ROIs.

Figure 6

Download asset Open asset

In summary, the analyses indicated that areas exhibiting similar fMRI effects for positive and negative expectations during the anticipation phase are linked to distinct temporal and oscillatory patterns. Notably, the left DLPFC and the left anterior insula displayed an early negative correlation between anticipatory activity and EEG power, primarily in the lower frequency range. Conversely, the subsequent effects in the bilateral ACC were associated with an increase in gamma oscillations and were observed at later time points during the anticipation phase indicating different processes.

The manipulation of positive, negative, or neutral expectations on a trial-by-trial basis allowed for a detailed analysis of their neural representations during the generation and integration of expectations with the nociceptive input. Our results revealed fundamental differences between the anticipation and pain phases, indicating the involvement of divergent processes. During the anticipation phase, valence-neutral representations were observed in areas of the DPMS and the anterior insula. After the onset of the nociceptive stimulus, these areas showed differentiated representations indicating separate processes for the formation of expectations and their integration within stimulus processing. The excellent temporal resolution of our fMRI-informed EEG measures further revealed a temporal sequence of expectation processing during the anticipation phase, with an early effect within the DLPFC, followed by activation in the anterior insula, and late effects within the ACC, in line with the occurrence of different time-sensitive sub-processes of expectation generation in this phase.

Our novel within-subjects paradigm was highly effective, as expressed by 47 out of 50 subjects consistently expecting and perceiving the intensity of pain stimuli in line with the cues that induced positive, negative, or neutral expectations. The trial-by-trial manipulation of expectations did not only affect self-reports, but also objective markers such as skin conductance responses and scores of the SIIPS (Woo et al., 2017). The SIIPS is an indicator of neural activity that tracks differences in pain perception that go beyond pain intensity and is classically affected by placebo and nocebo interventions (Botvinik-Nezer et al., 2023; Woo et al., 2017), thus serving as an excellent manipulation check for our paradigm. In line with our expectations, the SIIPS scores discriminated between positive and negative expectations only during the pain phase and not during the anticipation phase. This is a first hint at substantially different processes during the two phases.

The neural representations of directed expectations differed depending on the valence of the expectation, i.e., if they were positive or negative, and the time period. A shift from common (valence-neutral) to distinct (valence-dependent) effects was evident in areas classically involved in placebo analgesia during pain processing, including the bilateral DLPFC, right vmPFC, left anterior insula, and thalamus. The observation of this pattern in key areas of the DPMS indicates that the activation during the anticipation phase is not just mere pre-activation for later pain modulation, but that preparatory processes that are distinguishable from those that occur during pain perception take place. These anticipatory processes may include expectation generation and maintenance, while the focus shifts towards the integration of these expectations into the sensory stream and the evaluation of the percept during pain processing. Importantly, the spatial resolution of fMRI is limited when it comes to discriminating whether the same pattern of activity is due to identical activation or to activation in different sub-circuits within the same area. Nonetheless, the overlap of areas is an indicator of similar processes involved in a more general preparation process. The observed differentiation during pain is in agreement with the few studies that looked into the neuronal representations of positive and negative expectations and mainly found differential effects, e.g., opposite responses or differentially modulated areas during pain (Benedetti et al., 2020; Crawford et al., 2021; Koyama et al., 2005; Scott et al., 2008).

The network of the DPMS around the DLPFC, ACC, and vmPFC appears to play a pivotal role in expectation processes and in the valence-dependent modulation of pain perception. Following the framework of predictive coding, our results would suggest that the DPMS is the network responsible for integrating ascending signals with descending signals in the pain domain and that this process is similar for positive and negative valences during anticipation of pain but differentiates during pain processing. One important node of the DPMS is the DLPFC. The outstanding role of the DLPFC in placebo effects is well-established and supported by numerous reports of increased DLPFC activity during pain anticipation (Watson et al., 2009) as well as during pain processing (Zunhammer et al., 2021). This area has been classically associated with top-down-regulation and expectation modulation (Geuter et al., 2017), as when its neuronal excitability is experimentally manipulated, placebo effects can be diminished or enhanced (Egorova et al., 2015; Krummenacher et al., 2010; Tu et al., 2021). The early onset of DLPFC activity implied by the EEG analysis now suggests that the DLPFC plays an important role in the initiation of both positive and negative expectation effects within the DPMS prior to the noxious stimulation (Frisaldi et al., 2015; Geuter et al., 2017; Rossettini et al., 2023; Wager and Atlas, 2015; Wager et al., 2011). The vmPFC is another important hub within the DPMS which is suggested to integrate the input from the DLPFC and to generate affective meaning in order to maintain expectations (Geuter et al., 2017). It further modulates pain perception by directly affecting brainstem systems (Koban et al., 2017), e.g., by influencing PAG activity (Geuter et al., 2017). The notion that prefrontal areas exert top-down control over other areas of the DPMS is further supported by the earlier timing of DLPFC effects compared to the ACC (Craggs et al., 2014; Watson et al., 2009). The ACC also has a direct influence on the activity of the PAG, suggesting that the ACC is another crucial area for pain modulation (Geuter et al., 2017; Livrizzi et al., 2022). The effect of the bilateral ACC at the transition from the anticipation to the pain phase as indicated by our combined EEG-fMRI analysis could be interpreted as a preparatory mechanism for the modulation of incoming sensory information in downstream areas, consistent with its supposed role in the DPMS (Geuter et al., 2013; Geuter et al., 2017; Kong et al., 2008). Using this framework, reports of ACC activations for both positive and negative expectations during pain anticipation could be understood as a pre-activation of the ACC for pain modulation in either direction (Rossettini et al., 2023). The link between anticipatory activity in the ACC and EEG oscillatory activity was observed in the high gamma band, which is consistent with findings that demonstrate a connection between increased fMRI BOLD signals and a relative shift from lower to higher frequencies (Kilner et al., 2005). Gamma oscillations have been repeatedly reported in the context of pain and expectations and have been interpreted as reflecting feedforward signals of noxious information (e.g. Ploner et al., 2017; Strube et al., 2021). In combination with our findings, this might imply that high-frequency oscillations may not only signal higher actual or perceived pain intensity during pain processing (Nickel et al., 2022; Ploner et al., 2017; Strube et al., 2021; Tu et al., 2016), but might also be instrumental in the transfer of directed expectations from anticipation into pain processing.

Similarly, a shift from valence-neutral towards valence-dependent processing over time was demonstrated in the anterior insula. The insula is a key brain region involved in the network responsible for pain processing (e.g. Atlas and Wager, 2014; Kober et al., 2008) but is also part of several non-pain-related networks and works as a multimodal network hub (Horing and Büchel, 2022). Results of the fMRI-informed EEG analysis indicated early effects in the anticipatory period consistent with an early role of the anterior insula in expectation generation and initiation. Due to the multimodal nature of the insula, it may be speculated that multiple networks interact to integrate relevant information from different domains (e.g. from the visual cue and interoceptive information) into an expectancy signal in the anterior insula. Due to its connections and function within the salience network, this may involve encoding the expected threat level and salience of subsequent pain stimuli, leading to an anticipatory activation prior to the actual perceptual modulation (Taesler and Rose, 2016; Wiech et al., 2010). During pain perception, the anterior insula may then be more engaged in a network responsible for stimulus processing and evaluating actual salience and prediction errors in this perceptual process (Horing and Büchel, 2022), which may imply that the insula is involved in multiple tasks over time.

Similar to the anterior insula, the activation of the angular gyrus during pain processing and pain anticipation could be understood as an indicator for processes related to the formation and integration of directed expectations. The angular gyrus has been connected to expectation-related pain modulation only a few times (e.g. Atlas and Wager, 2014; Tu et al., 2019) but its involvement is prominent in our results. Based on its presumed function in maintaining recollected multimodal representations (Jablonowski and Rose, 2022; Vilberg and Rugg, 2012), the angular gyrus may be engaged in transforming sensory information from the visual cues into expectancy signals that can be processed by e.g., the DPMS and salience network.

With our rather unconventional and new paradigm, we were able to manipulate participants expectations on a trial-by-trial level and derive insights into the neural dynamics of positive and negative expectations. While this may give rise to questions regarding the comparability of our study to previous paradigms and the manner in which our control condition was employed, we would argue that our expectation manipulation falls in line with a manipulation of treatment expectancies, a typical method of expectation manipulation in placebo paradigms (see Atlas and Wager, 2014). Accordingly, it is comparable to previous studies on placebo and nocebo effects. In our study, participants were presented with a cue that induced expectations regarding a ‘treatment,’ although in this case the ‘treatment’ originated from changes in their own brain activity. This is, in a broader sense, comparable to studies utilizing sham TENS-devices that are supposedly altering peripheral pain transmission (Skvortsova et al., 2020).

Moreover, implementing a proper control condition in expectation modulation paradigms is an inherently difficult task as forming expectations about our environment is a natural process. Therefore, we recognize that participants most likely did form expectations of medium pain intensities in the control condition over the course of the experiment. This is in line with previous research on placebo and nocebo effects, in which participants also typically rated control stimuli in between placebo and nocebo conditions (Bingel et al., 2011; Colloca et al., 2010; Shih et al., 2019). However, we would still argue that we can meaningfully compare the placebo and nocebo condition to the control conditions to investigate the neuronal underpinnings of expectation effects. Independently of whether participants build up an expectation of ‘medium’ intensities in the control condition, which caused them to perceive stimuli in line with this, or if they simply perceived the stimuli as they were (of medium intensity) with limited effects of expectations, the crucial difference to the placebo and nocebo conditions is that there was no alteration of perception due to previous experiences or verbal information and no shift of perception from the actual stimulus intensity towards any direction in the control condition. Thus, we were able to identify the effects of directed expectations by comparing positive and negative expectations to neutral expectations as a baseline. Our analysis of within-condition variability further showed that ratings indeed varied within conditions and that the amount of variation was comparable between nocebo and control. Over time, expectations were dynamically updated in all three conditions, speaking against alternative explanations of the rating differences between conditions such as a regression to the mean of ratings in the control condition.

Based on the present results, understanding the processing of expectations requires an examination of its temporal and spatial dynamics from anticipation to pain processing, while comparing positive and negative valences to each other and to a control condition. We found largely comparable activation for positive and negative expectations during the anticipation phase, including regions outside of those classically observed in pain processing and modulation. Based on the observed temporal profiles, the DLPFC and anterior insula may be related to the top-down initiation and generation of expectations, while the ACC is activated in close proximity to pain onset as a direct link to pain modulation. During the pain phase, the focus shifts from expectation generation and maintenance towards pain modulation in either positive or negative direction, leading to distinct effects of positive and negative expectations in many areas that initially encoded expectations independently of their valence. It is not surprising that expectations are not a static process but involve different time-dependent components, as pain perception was also recently described as a complex process of interactions among multiple brain systems that are reconfigured over time (Lee et al., 2022). Expectation generation, integration, and pain perception all appear to be dynamic processes, with both common and distinct routes for positive and negative expectations, depending on the time point of examination.

Derived data that support the findings of this study are available at https://osf.io/3g49v/. Due to data privacy restrictions, further data is only available on request.The only data not publicly available are the unprocessed raw data, which could potentially be used to re-identify participants; this measure is in place to safeguard participant privacy. Researchers interested in accessing the raw data should contact the lead investigator Michael Rose (rose@uke.de), providing a rationale for their request. Upon review, they will be granted access to the raw data.

1. 1. Wittkamp CA
   2. Wolf M
   3. Rose M

   (2024) Open Science Framework

   The neural dynamics of positive and negative expectations of pain.

   https://doi.org/10.17605/OSF.IO/3G49V

1. 1. Adler-Neal AL
   2. Emerson NM
   3. Farris SR
   4. Jung Y
   5. Coghill RC
   6. Zeidan F

   (2019) Brain moderators supporting the relationship between depressive mood and pain

   Pain 160:2028–2035.

   https://doi.org/10.1097/j.pain.0000000000001595

   • PubMed
   • Google Scholar
2. 1. Amanzio M
   2. Benedetti F
   3. Porro CA
   4. Palermo S
   5. Cauda F

   (2013) Activation likelihood estimation meta-analysis of brain correlates of placebo analgesia in human experimental pain

   Human Brain Mapping 34:738–752.

   https://doi.org/10.1002/hbm.21471

   • PubMed
   • Google Scholar
3. 1. Amanzio M
   2. Palermo S

   (2019) Pain anticipation and nocebo-related responses: a descriptive mini-review of functional neuroimaging studies in normal subjects and precious hints on pain processing in the context of neurodegenerative disorders

   Frontiers in Pharmacology 10:969.

   https://doi.org/10.3389/fphar.2019.00969

   • PubMed
   • Google Scholar
4. 1. Atlas LY
   2. Wager TD

   (2012) How expectations shape pain

   Neuroscience Letters 520:140–148.

   https://doi.org/10.1016/j.neulet.2012.03.039

   • PubMed
   • Google Scholar
5. Book

   1. Atlas LY
   2. Wager TD

   (2014) A meta-analysis of brain mechanisms of placebo analgesia: consistent findings and unanswered questions

   In: Atlas LY, editors. Handbook of Experimental Pharmacology. Springer. pp. 37–69.

   https://doi.org/10.1007/978-3-662-44519-8_3

   • Google Scholar
6. Book

   1. Beck AT
   2. Steer RA
   3. Brown GK

   (1996)

   Manual for the Beck Depression Inventory-II

   Psychological Corporation.

   • Google Scholar
7. 1. Benedek M
   2. Kaernbach C

   (2010) A continuous measure of phasic electrodermal activity

   Journal of Neuroscience Methods 190:80–91.

   https://doi.org/10.1016/j.jneumeth.2010.04.028

   • PubMed
   • Google Scholar
8. 1. Benedetti F

   (2014) Placebo effects: from the neurobiological paradigm to translational implications

   Neuron 84:623–637.

   https://doi.org/10.1016/j.neuron.2014.10.023

   • PubMed
   • Google Scholar
9. 1. Benedetti F
   2. Frisaldi E
   3. Barbiani D
   4. Camerone E
   5. Shaibani A

   (2020) Nocebo and the contribution of psychosocial factors to the generation of pain

   Journal of Neural Transmission 127:687–696.

   https://doi.org/10.1007/s00702-019-02104-x

   • PubMed
   • Google Scholar
10. 1. Bingel U
    2. Wanigasekera V
    3. Wiech K
    4. Ni Mhuircheartaigh R
    5. Lee MC
    6. Ploner M
    7. Tracey I

    (2011) The effect of treatment expectation on drug efficacy: imaging the analgesic benefit of the opioid remifentanil

    Science Translational Medicine 3:70ra14.

    https://doi.org/10.1126/scitranslmed.3001244

    • PubMed
    • Google Scholar
11. Preprint

    1. Botvinik-Nezer R
    2. Petre B
    3. Ceko M
    4. Lindquist MA
    5. Friedman NP
    6. Wager TD

    (2023) Placebo treatment affects brain systems related to affective and cognitive processes, but not nociceptive Pain

    bioRxiv.

    https://doi.org/10.1101/2023.09.21.558825

    • Google Scholar
12. 1. Büchel C
    2. Geuter S
    3. Sprenger C
    4. Eippert F

    (2014) Placebo analgesia: A predictive coding perspective

    Neuron 81:1223–1239.

    https://doi.org/10.1016/j.neuron.2014.02.042

    • PubMed
    • Google Scholar
13. 1. Colloca L
    2. Benedetti F

    (2007) Nocebo hyperalgesia: how anxiety is turned into pain

    Current Opinion in Anaesthesiology 20:435–439.

    https://doi.org/10.1097/ACO.0b013e3282b972fb

    • PubMed
    • Google Scholar
14. 1. Colloca L
    2. Sigaudo M
    3. Benedetti F

    (2008) The role of learning in nocebo and placebo effects

    Pain 136:211–218.

    https://doi.org/10.1016/j.pain.2008.02.006

    • PubMed
    • Google Scholar
15. 1. Colloca L
    2. Petrovic P
    3. Wager TD
    4. Ingvar M
    5. Benedetti F

    (2010) How the number of learning trials affects placebo and nocebo responses

    Pain 151:430–439.

    https://doi.org/10.1016/j.pain.2010.08.007

    • PubMed
    • Google Scholar
16. 1. Craggs JG
    2. Price DD
    3. Robinson ME

    (2014) Enhancing the placebo response: functional magnetic resonance imaging evidence of memory and semantic processing in placebo analgesia

    The Journal of Pain 15:435–446.

    https://doi.org/10.1016/j.jpain.2013.12.009

    • PubMed
    • Google Scholar
17. 1. Crawford LS
    2. Mills EP
    3. Hanson T
    4. Macey PM
    5. Glarin R
    6. Macefield VG
    7. Keay KA
    8. Henderson LA

    (2021) Brainstem mechanisms of pain modulation: a within-subjects 7t fmri study of placebo analgesic and nocebo hyperalgesic responses

    The Journal of Neuroscience 41:9794–9806.

    https://doi.org/10.1523/JNEUROSCI.0806-21.2021

    • PubMed
    • Google Scholar
18. 1. Dionisio S
    2. Mayoglou L
    3. Cho SM
    4. Prime D
    5. Flanigan PM
    6. Lega B
    7. Mosher J
    8. Leahy R
    9. Gonzalez-Martinez J
    10. Nair D

    (2019) Connectivity of the human insula: A cortico-cortical evoked potential (CCEP) study

    Cortex; a Journal Devoted to the Study of the Nervous System and Behavior 120:419–442.

    https://doi.org/10.1016/j.cortex.2019.05.019

    • PubMed
    • Google Scholar
19. 1. Egorova N
    2. Yu R
    3. Kaur N
    4. Vangel M
    5. Gollub RL
    6. Dougherty DD
    7. Kong J
    8. Camprodon JA

    (2015) Neuromodulation of conditioned placebo/nocebo in heat pain: anodal vs cathodal transcranial direct current stimulation to the right dorsolateral prefrontal cortex

    Pain 156:1342–1347.

    https://doi.org/10.1097/j.pain.0000000000000163

    • PubMed
    • Google Scholar
20. 1. Eippert F
    2. Bingel U
    3. Schoell ED
    4. Yacubian J
    5. Klinger R
    6. Lorenz J
    7. Büchel C

    (2009) Activation of the opioidergic descending pain control system underlies placebo analgesia

    Neuron 63:533–543.

    https://doi.org/10.1016/j.neuron.2009.07.014

    • PubMed
    • Google Scholar
21. 1. Englert R
    2. Schedlowski M
    3. Engler H
    4. Rief W
    5. Büchel C
    6. Bingel U
    7. Spisak T

    (2023) ALIIAS: anonymization/pseudonymization with limesurvey integration and ii-factor authentication for scientific research

    SoftwareX 24:101522.

    https://doi.org/10.1016/j.softx.2023.101522

    • Google Scholar
22. 1. Frässle S
    2. Aponte EA
    3. Bollmann S
    4. Brodersen KH
    5. Do CT
    6. Harrison OK
    7. Harrison SJ
    8. Heinzle J
    9. Iglesias S
    10. Kasper L
    11. Lomakina EI
    12. Mathys C
    13. Müller-Schrader M
    14. Pereira I
    15. Petzschner FH
    16. Raman S
    17. Schöbi D
    18. Toussaint B
    19. Weber LA
    20. Yao Y
    21. Stephan KE

    (2021) TAPAS: an open-source software package for translational neuromodeling and computational psychiatry

    Frontiers in Psychiatry 12:680811.

    https://doi.org/10.3389/fpsyt.2021.680811

    • Google Scholar
23. 1. Freeman S
    2. Yu R
    3. Egorova N
    4. Chen X
    5. Kirsch I
    6. Claggett B
    7. Kaptchuk TJ
    8. Gollub RL
    9. Kong J

    (2015) Distinct neural representations of placebo and nocebo effects

    NeuroImage 112:197–207.

    https://doi.org/10.1016/j.neuroimage.2015.03.015

    • PubMed
    • Google Scholar
24. 1. Frisaldi E
    2. Piedimonte A
    3. Benedetti F

    (2015) Placebo and nocebo effects: A complex interplay between psychological factors and neurochemical networks

    The American Journal of Clinical Hypnosis 57:267–284.

    https://doi.org/10.1080/00029157.2014.976785

    • PubMed
    • Google Scholar
25. 1. Fu J
    2. Wu S
    3. Liu C
    4. Camilleri JA
    5. Eickhoff SB
    6. Yu R

    (2021) Distinct neural networks subserve placebo analgesia and nocebo hyperalgesia

    NeuroImage 231:117833.

    https://doi.org/10.1016/j.neuroimage.2021.117833

    • PubMed
    • Google Scholar
26. 1. Fullana MA
    2. Harrison BJ
    3. Soriano-Mas C
    4. Vervliet B
    5. Cardoner N
    6. Àvila-Parcet A
    7. Radua J

    (2016) Neural signatures of human fear conditioning: an updated and extended meta-analysis of fMRI studies

    Molecular Psychiatry 21:500–508.

    https://doi.org/10.1038/mp.2015.88

    • PubMed
    • Google Scholar
27. 1. Geuter S
    2. Eippert F
    3. Hindi Attar C
    4. Büchel C

    (2013) Cortical and subcortical responses to high and low effective placebo treatments

    NeuroImage 67:227–236.

    https://doi.org/10.1016/j.neuroimage.2012.11.029

    • PubMed
    • Google Scholar
28. 1. Geuter S
    2. Koban L
    3. Wager TD

    (2017) The cognitive neuroscience of placebo effects: concepts, predictions, and physiology

    Annual Review of Neuroscience 40:167–188.

    https://doi.org/10.1146/annurev-neuro-072116-031132

    • PubMed
    • Google Scholar
29. Book

    1. Hautzinger M
    2. Keller F
    3. Kühner C

    (2006)

    Beck Depressions-Inventar

    BDI-II). Revision. Harcourt Test Services.

    • Google Scholar
30. 1. Hipp JF
    2. Engel AK
    3. Siegel M

    (2011) Oscillatory synchronization in large-scale cortical networks predicts perception

    Neuron 69:387–396.

    https://doi.org/10.1016/j.neuron.2010.12.027

    • PubMed
    • Google Scholar
31. 1. Horing B
    2. Büchel C

    (2022) The human insula processes both modality-independent and pain-selective learning signals

    PLOS Biology 20:e3001540.

    https://doi.org/10.1371/journal.pbio.3001540

    • PubMed
    • Google Scholar
32. 1. Jablonowski J
    2. Rose M

    (2022) The functional dissociation of posterior parietal regions during multimodal memory formation

    Human Brain Mapping 43:3469–3485.

    https://doi.org/10.1002/hbm.25861

    • PubMed
    • Google Scholar
33. 1. Kasper L
    2. Bollmann S
    3. Diaconescu AO
    4. Hutton C
    5. Heinzle J
    6. Iglesias S
    7. Hauser TU
    8. Sebold M
    9. Manjaly Z-M
    10. Pruessmann KP
    11. Stephan KE

    (2017) The physio toolbox for modeling physiological noise in fmri data

    Journal of Neuroscience Methods 276:56–72.

    https://doi.org/10.1016/j.jneumeth.2016.10.019

    • Google Scholar
34. 1. Kilner JM
    2. Mattout J
    3. Henson R
    4. Friston KJ

    (2005) Hemodynamic correlates of EEG: A heuristic

    NeuroImage 28:280–286.

    https://doi.org/10.1016/j.neuroimage.2005.06.008

    • PubMed
    • Google Scholar
35. 1. Koban L
    2. Jepma M
    3. Geuter S
    4. Wager TD

    (2017) What’s in a word? How instructions, suggestions, and social information change pain and emotion

    Neuroscience and Biobehavioral Reviews 81:29–42.

    https://doi.org/10.1016/j.neubiorev.2017.02.014

    • PubMed
    • Google Scholar
36. 1. Kober H
    2. Barrett LF
    3. Joseph J
    4. Bliss-Moreau E
    5. Lindquist K
    6. Wager TD

    (2008) Functional grouping and cortical-subcortical interactions in emotion: A meta-analysis of neuroimaging studies

    NeuroImage 42:998–1031.

    https://doi.org/10.1016/j.neuroimage.2008.03.059

    • PubMed
    • Google Scholar
37. 1. Kong J
    2. Kaptchuk TJ
    3. Polich G
    4. Kirsch I
    5. Gollub RL

    (2007) Placebo analgesia: Findings from brain imaging studies and emerging hypotheses

    Reviews in the Neurosciences 18:173–190.

    https://doi.org/10.1515/REVNEURO.2007.18.3-4.173

    • Google Scholar
38. 1. Kong J
    2. Gollub RL
    3. Polich G
    4. Kirsch I
    5. Laviolette P
    6. Vangel M
    7. Rosen B
    8. Kaptchuk TJ

    (2008) A functional magnetic resonance imaging study on the neural mechanisms of hyperalgesic nocebo effect

    Journal of Neuroscience 28:13354–13362.

    https://doi.org/10.1523/JNEUROSCI.2944-08.2008

    • Google Scholar
39. 1. Kong J
    2. Benedetti F

    (2014) Placebo and nocebo effects: an introduction to psychological and biological mechanisms

    Experimental Pharmacology 225:3–15.

    https://doi.org/10.1007/978-3-662-44519-8_1

    • PubMed
    • Google Scholar
40. 1. Koyama T
    2. McHaffie JG
    3. Laurienti PJ
    4. Coghill RC

    (2005) The subjective experience of pain: Where expectations become reality

    PNAS 102:12950–12955.

    https://doi.org/10.1073/pnas.0408576102

    • Google Scholar
41. 1. Krummenacher P
    2. Candia V
    3. Folkers G
    4. Schedlowski M
    5. Schönbächler G

    (2010) Prefrontal cortex modulates placebo analgesia

    Pain 148:368–374.

    https://doi.org/10.1016/j.pain.2009.09.033

    • PubMed
    • Google Scholar
42. Book

    1. Laux L
    2. Hock M
    3. Bergner-Koether R
    4. Hodapp V
    5. Renner KH

    (2013)

    Das State-Trait-Angst-Depressions-Inventar: STADI

    Göttingen: Hogrefe.

    • Google Scholar
43. 1. Lee JJ
    2. Lee S
    3. Lee DH
    4. Woo CW

    (2022) Functional brain reconfiguration during sustained pain

    eLife 11:e74463.

    https://doi.org/10.7554/eLife.74463

    • PubMed
    • Google Scholar
44. 1. Livrizzi G
    2. Lubejko ST
    3. Johnson DA
    4. Weiss CE
    5. Patel J
    6. Banghart MR

    (2022) Prefrontal input to the periaqueductal gray controls placebo analgesia

    The Journal of Pain 23:19–20.

    https://doi.org/10.1016/j.jpain.2022.03.076

    • Google Scholar
45. 1. Lu C
    2. Yang T
    3. Zhao H
    4. Zhang M
    5. Meng F
    6. Fu H
    7. Xie Y
    8. Xu H

    (2016) Insular cortex is critical for the perception, modulation, and chronification of pain

    Neuroscience Bulletin 32:191–201.

    https://doi.org/10.1007/s12264-016-0016-y

    • PubMed
    • Google Scholar
46. 1. Manaï M
    2. van Middendorp H
    3. Veldhuijzen DS
    4. Huizinga TWJ
    5. Evers AWM

    (2019) How to prevent, minimize, or extinguish nocebo effects in pain: A narrative review on mechanisms, predictors, and interventions

    Pain Reports 4:e699.

    https://doi.org/10.1097/PR9.0000000000000699

    • PubMed
    • Google Scholar
47. 1. McNeil DW
    2. Rainwater AJ

    (1998) Development of the fear of pain questionnaire--III

    Journal of Behavioral Medicine 21:389–410.

    https://doi.org/10.1023/a:1018782831217

    • PubMed
    • Google Scholar
48. 1. Mestdagh M
    2. Pe M
    3. Pestman W
    4. Verdonck S
    5. Kuppens P
    6. Tuerlinckx F

    (2018) Sidelining the mean: The relative variability index as a generic mean-corrected variability measure for bounded variables

    Psychological Methods 23:690–707.

    https://doi.org/10.1037/met0000153

    • PubMed
    • Google Scholar
49. 1. Nickel MM
    2. Tiemann L
    3. Hohn VD
    4. May ES
    5. Gil Ávila C
    6. Eippert F
    7. Ploner M

    (2022) Temporal-spectral signaling of sensory information and expectations in the cerebral processing of pain

    PNAS 119:e2116616119.

    https://doi.org/10.1073/pnas.2116616119

    • PubMed
    • Google Scholar
50. 1. Oostenveld R
    2. Fries P
    3. Maris E
    4. Schoffelen JM

    (2011) FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data

    Computational Intelligence and Neuroscience 2011:156869.

    https://doi.org/10.1155/2011/156869

    • PubMed
    • Google Scholar
51. 1. Palermo S
    2. Benedetti F
    3. Costa T
    4. Amanzio M

    (2015) Pain anticipation: an activation likelihood estimation meta-analysis of brain imaging studies

    Human Brain Mapping 36:1648–1661.

    https://doi.org/10.1002/hbm.22727

    • PubMed
    • Google Scholar
52. 1. Ploghaus A
    2. Tracey I
    3. Gati JS
    4. Clare S
    5. Menon RS
    6. Matthews PM
    7. Rawlins JN

    (1999) Dissociating pain from its anticipation in the human brain

    Science 284:1979–1981.

    https://doi.org/10.1126/science.284.5422.1979

    • PubMed
    • Google Scholar
53. 1. Ploner M
    2. Sorg C
    3. Gross J

    (2017) Brain Rhythms of Pain

    Trends in Cognitive Sciences 21:100–110.

    https://doi.org/10.1016/j.tics.2016.12.001

    • Google Scholar
54. 1. Prince JS
    2. Charest I
    3. Kurzawski JW
    4. Pyles JA
    5. Tarr MJ
    6. Kay KN

    (2022) Improving the accuracy of single-trial fMRI response estimates using GLMsingle

    eLife 11:e77599.

    https://doi.org/10.7554/eLife.77599

    • PubMed
    • Google Scholar
55. 1. Rossettini G
    2. Campaci F
    3. Bialosky J
    4. Huysmans E
    5. Vase L
    6. Carlino E

    (2023) The biology of placebo and nocebo effects on experimental and chronic pain: state of the art

    Journal of Clinical Medicine 12:4113.

    https://doi.org/10.3390/jcm12124113

    • PubMed
    • Google Scholar
56. 1. Schmid J
    2. Bingel U
    3. Ritter C
    4. Benson S
    5. Schedlowski M
    6. Gramsch C
    7. Forsting M
    8. Elsenbruch S

    (2015) Neural underpinnings of nocebo hyperalgesia in visceral pain: A fMRI study in healthy volunteers

    NeuroImage 120:114–122.

    https://doi.org/10.1016/j.neuroimage.2015.06.060

    • PubMed
    • Google Scholar
57. 1. Scott DJ
    2. Stohler CS
    3. Egnatuk CM
    4. Wang H
    5. Koeppe RA
    6. Zubieta JK

    (2008) Placebo and nocebo effects are defined by opposite opioid and dopaminergic responses

    Archives of General Psychiatry 65:220–231.

    https://doi.org/10.1001/archgenpsychiatry.2007.34

    • PubMed
    • Google Scholar
58. 1. Shi Y
    2. Cui S
    3. Zeng Y
    4. Huang S
    5. Cai G
    6. Yang J
    7. Wu W

    (2021) Brain network to placebo and nocebo responses in acute experimental lower back pain: a multivariate granger causality analysis of fmri data

    Frontiers in Behavioral Neuroscience 15:696577.

    https://doi.org/10.3389/fnbeh.2021.696577

    • PubMed
    • Google Scholar
59. 1. Shih YW
    2. Tsai HY
    3. Lin FS
    4. Lin YH
    5. Chiang CY
    6. Lu ZL
    7. Tseng MT

    (2019) Effects of positive and negative expectations on human pain perception engage separate but interrelated and dependently regulated cerebral mechanisms

    The Journal of Neuroscience 39:1261–1274.

    https://doi.org/10.1523/JNEUROSCI.2154-18.2018

    • PubMed
    • Google Scholar
60. 1. Skvortsova A
    2. Veldhuijzen DS
    3. van Middendorp H
    4. Colloca L
    5. Evers AWM

    (2020) Effects of oxytocin on placebo and nocebo effects in a pain conditioning paradigm: a randomized controlled trial

    The Journal of Pain 21:430–439.

    https://doi.org/10.1016/j.jpain.2019.08.010

    • PubMed
    • Google Scholar
61. 1. Strube A
    2. Rose M
    3. Fazeli S
    4. Büchel C

    (2021) The temporal and spectral characteristics of expectations and prediction errors in pain and thermoception

    eLife 10:e62809.

    https://doi.org/10.7554/eLife.62809

    • PubMed
    • Google Scholar
62. 1. Taesler P
    2. Rose M

    (2016) Prestimulus theta oscillations and connectivity modulate pain perception

    The Journal of Neuroscience 36:5026–5033.

    https://doi.org/10.1523/JNEUROSCI.3325-15.2016

    • PubMed
    • Google Scholar
63. 1. Tu Y
    2. Zhang Z
    3. Tan A
    4. Peng W
    5. Hung YS
    6. Moayedi M
    7. Iannetti GD
    8. Hu L

    (2016) Alpha and gamma oscillation amplitudes synergistically predict the perception of forthcoming nociceptive stimuli

    Human Brain Mapping 37:501–514.

    https://doi.org/10.1002/hbm.23048

    • PubMed
    • Google Scholar
64. 1. Tu Y
    2. Park J
    3. Ahlfors SP
    4. Khan S
    5. Egorova N
    6. Lang C
    7. Cao J
    8. Kong J

    (2019) A neural mechanism of direct and observational conditioning for placebo and nocebo responses

    NeuroImage 184:954–963.

    https://doi.org/10.1016/j.neuroimage.2018.10.020

    • PubMed
    • Google Scholar
65. 1. Tu Y
    2. Wilson G
    3. Camprodon J
    4. Dougherty DD
    5. Vangel M
    6. Benedetti F
    7. Kaptchuk TJ
    8. Gollub RL
    9. Kong J

    (2021) Manipulating placebo analgesia and nocebo hyperalgesia by changing brain excitability

    PNAS 118:e2101273118.

    https://doi.org/10.1073/pnas.2101273118

    • Google Scholar
66. 1. Tu Y
    2. Zhang L
    3. Kong J

    (2022) Placebo and nocebo effects: from observation to harnessing and clinical application

    Translational Psychiatry 12:524.

    https://doi.org/10.1038/s41398-022-02293-2

    • PubMed
    • Google Scholar
67. 1. Vilberg KL
    2. Rugg MD

    (2012) The neural correlates of recollection: transient versus sustained FMRI effects

    The Journal of Neuroscience 32:15679–15687.

    https://doi.org/10.1523/JNEUROSCI.3065-12.2012

    • PubMed
    • Google Scholar
68. 1. Wager TD
    2. Rilling JK
    3. Smith EE
    4. Sokolik A
    5. Casey KL
    6. Davidson RJ
    7. Kosslyn SM
    8. Rose RM
    9. Cohen JD

    (2004) Placebo-induced changes in FMRI in the anticipation and experience of pain

    Science 303:1162–1167.

    https://doi.org/10.1126/science.1093065

    • PubMed
    • Google Scholar
69. 1. Wager TD
    2. Atlas LY
    3. Leotti LA
    4. Rilling JK

    (2011) Predicting individual differences in placebo analgesia: contributions of brain activity during anticipation and pain experience

    The Journal of Neuroscience 31:439–452.

    https://doi.org/10.1523/JNEUROSCI.3420-10.2011

    • Google Scholar
70. 1. Wager TD
    2. Atlas LY

    (2015) The neuroscience of placebo effects: connecting context, learning and health

    Nature Reviews. Neuroscience 16:403–418.

    https://doi.org/10.1038/nrn3976

    • PubMed
    • Google Scholar
71. 1. Watson A
    2. El-Deredy W
    3. Iannetti GD
    4. Lloyd D
    5. Tracey I
    6. Vogt BA
    7. Nadeau V
    8. Jones AKP

    (2009) Placebo conditioning and placebo analgesia modulate a common brain network during pain anticipation and perception

    Pain 145:24–30.

    https://doi.org/10.1016/j.pain.2009.04.003

    • PubMed
    • Google Scholar
72. 1. Wiech K
    2. Lin C
    3. Brodersen KH
    4. Bingel U
    5. Ploner M
    6. Tracey I

    (2010) Anterior insula integrates information about salience into perceptual decisions about pain

    The Journal of Neuroscience 30:16324–16331.

    https://doi.org/10.1523/JNEUROSCI.2087-10.2010

    • PubMed
    • Google Scholar
73. 1. Woo CW
    2. Schmidt L
    3. Krishnan A
    4. Jepma M
    5. Roy M
    6. Lindquist MA
    7. Atlas LY
    8. Wager TD

    (2017) Quantifying cerebral contributions to pain beyond nociception

    Nature Communications 8:14211.

    https://doi.org/10.1038/ncomms14211

    • PubMed
    • Google Scholar
74. 1. Zunhammer M
    2. Spisák T
    3. Wager TD
    4. Bingel U
    5. Placebo Imaging Consortium

    (2021) Meta-analysis of neural systems underlying placebo analgesia from individual participant fMRI data

    Nature Communications 12:1391.

    https://doi.org/10.1038/s41467-021-21179-3

    • PubMed
    • Google Scholar

Author details

1. Christoph Arne Wittkamp

   Department of Systems Neuroscience, University Medical Center Hamburg Eppendorf, Hamburg, Germany

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

   Contributed equally with

   Maren-Isabel Wolf

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0009-0000-8609-9053

2. Maren-Isabel Wolf

   Department of Systems Neuroscience, University Medical Center Hamburg Eppendorf, Hamburg, Germany

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

   Contributed equally with

   Christoph Arne Wittkamp

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3201-0134

3. Michael Rose

   Department of Systems Neuroscience, University Medical Center Hamburg Eppendorf, Hamburg, Germany

   Contribution

   Conceptualization, Resources, Software, Supervision, Funding acquisition, Validation, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   rose@uke.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9789-7066

• 491

  views

• 14

  downloads

• 1

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.