This study provides a novel characterization of the electrophysiological phenotype of nocebo hyperalgesia using EEG. Spectral and temporal dynamics of brain oscillations were studied at baseline, during resting-state pre- and post- measurements and during nocebo acquisition and evocation. The main findings of this study are (i) a negative correlation between LRTC of beta oscillations and the magnitude of nocebo hyperalgesia, (ii) a positive correlation between baseline LRTC and magnitude of nocebo hyperalgesia, (iii) alpha and beta power suppression during nocebo conditioning, and (iv) biomarker differences between the experience of high pain at baseline and the experience of nocebo-augmented pain.
In previous research, reduction in LRTC of oscillations has been reported in the alpha and beta bands in patients with cognitive disorders 23,24,38, while other studies link increased LRTC to reduced attention or cognitive performance 22,39−42. LRTC characterize neuronal systems that require rapid reorganization and responsiveness to changing processing demands 25. Previous research indicates that neuronal systems involved in sustained attention may be characterized by a less volatile state with decreased LRTC 25. LRTC changes in the present study may thus be related to reduced attention or cognitive performance. In other words, effective conditioning required sustained attention with a relatively low cognitive load to result in stronger learning and thus larger nocebo hyperalgesia was characterized by reduced complexity of neural dynamics.
While on one hand sustained attention (characterized by decreased LRTC) was related to larger nocebo responses, on the other hand strong resting-state LRTC at baseline predict more effective conditioning of nocebo responses. We found that, during rest, before the start of the experimental phases, strong LRTC predicted higher nocebo responses. This finding relates to the above-mentioned studies, that pointed towards an involvement of LRTC in cognitive ability 23,24,38. Stronger LRTC reflect more complex neural dynamics and therefore, it appears that people with more complex baseline brain activity may exhibit higher cognitive functioning (Montez et al., 2009) and are thus more susceptible to the acquisition of nocebo hyperalgesia through learning. Here, the implication of gamma band oscillations is in line with EEG research on (associative) learning, suggesting that memory encoding involves gamma oscillations 12,43,44 potentially in coordination with hippocampal function 45. This links gamma oscillations, which were shown to be involved in nocebo in this study, to a role of the hippocampus in learning and nocebo hyperalgesia 46,47. It is also noteworthy that emotional processes that may play a mediating role in nocebo hyperalgesia, such as fear 48, may engage patterns of gamma coupling in the amygdala 49, a structure that has also been implicated in nocebo hyperalgesia 47,50,51. Our finding of increased complexity of gamma-band oscillations in those more susceptible to nocebo hyperalgesia may thus provide electrophysiological evidence of specific underlying cognitive-emotional processes, such as associative learning ability as well as fear processing.
Alpha band oscillatory power has been shown to underlie the perceptual processing of incoming stimuli, including sensory perception 52. Our study was methodologically different from the two previous studies on electrophysiological nocebo correlates 15,18 and our results do not show consistent support of previous findings relating alpha oscillations to nocebo hyperalgesia. While our findings indicate an involvement of alpha band oscillations during acquisition, we did not find pre- to post-acquisition changes in alpha oscillations. Methodologically, it possible that the time elapsed between the first and second resting state recordings was too long, resulting in a failure to capture electrophysiological changes in alpha oscillations related to nocebo processing.
Nevertheless, we found that nocebo trials during the acquisition phase were characterized by decreased power in the alpha band, as compared to control trials. Our finding may reflect the formation of pain expectations and an inhibitory function of alpha oscillations in pain perception. Moreover, alpha-band oscillations where involved when comparing the experience of baseline high-pain stimulations to the experience of increased pain under nocebo hyperalgesic conditions, in the evocation phase. We found that there was a significant increase in alpha-band power during nocebo responses, compared to baseline pain of a matched, high intensity pain stimulus. In line with the literature, these findings may reflect the role of alpha-band oscillations in expectations 8,9, and the cognitive regulation of pain 10,11.
We then aimed to differentiate the temporal electrophysiological profile of experiencing high pain at baseline from that of experiencing high pain as a result of induced nocebo hyperalgesia. We found that the complexity of neuronal oscillations was lower during nocebo-augmented pain compared to baseline pain of a matched, high intensity pain stimulus. Lower oscillatory complexity during nocebo-augmented pain may be in line with our finding that lower LRTC during acquisition were associated with higher nocebo magnitudes. This could mean that the evocation of nocebo hyperalgesia, due to a state of sustained attention, may be characterized by decreased LRTC 22,39−42. Nocebo-augmented pain seems to rely on cognitive processes such as learning, memory recall, and pain modulation. Decreased LRTC may thus indicate increased attentional load or cognitive performance during nocebo-augmented pain responses. More specifically, the decreased LRTC of gamma oscillations during nocebo evocation, as compared to the baseline high pain, may alternatively or additionally indicate a learning process. It has previously been shown that while learning new information may lead to increased gamma power or synchronization 43,45,53, power of gamma oscillations may show a decrease after learning 54. It is thus possible that in nocebo evocation, when learning is discontinued, gamma oscillations exhibit a decrease in power that reflects a previous active learning state. These results may thus highlight pronounced cognitive and learning-related differences between the neurophysiology of experiencing high pain and experiencing nocebo-evoked increased pain. Nevertheless, the LRTC findings in this study also highlight the intricacy of such complex biomarkers of temporal brain function and how they may characterize diverse cognitive functions and loads in different ways.
A number of limitations may have impacted the results of this study. First, aggregating trials of specific conditions into 10-second segments may have smeared out effects that could have been better captured using an event-related paradigm, in which the exact onset of each pain stimulus or response could be used to epoch the data into segments locked to each trial. Furthermore, the generalizability of our findings may be limited by the recruitment of a healthy, young participant sample. Findings of this study may not be consistent with results derived from pain patients or individuals who have experienced severe or chronic pain in the past, as their electrophysiological phenotype may differ from that of healthy people 55.
With a number of novel aspects of the neurophysiological phenotype of nocebo hyperalgesia emerging through the present study, future directions are also coming into view. It is imperative for future research to focus on the generalizability and translation of experimental results into clinical practice. This study highlighted novel EEG biomarkers that are related to the experimental nocebo context. EEG is a practical and relatively cost-effective method that may provide a valuable means for the identification of nocebo-augmented pain as well as nocebo contexts. For these diagnostic potentials to be realized, a next step is for future studies to replicate our findings in clinical contexts and populations.
In sum, the present study yielded a number of novel findings regarding the electrophysiology that may underlie or mediate nocebo hyperalgesia. We identified both spectral and temporal parameters that are related to nocebo-augmented pain, with the latter presenting as the most important correlate of nocebo hyperalgesia in this study. The role of learning and attention at the electrophysiological level was highlighted through the involvement of LRTC as well as the extensive involvement of gamma oscillations under hyperalgesic conditions. These results are an important step towards identifying physiological biomarkers of nocebo hyperalgesia, a phenomenon that, to date, does not have any formal diagnostic criteria. The identification of biomarkers of nocebo hyperalgesia may thus prove imperative in the strive to identify and treat these effects.