Musculoskeletal Research Paper The neurologic pain signature responds to nonsteroidal anti-inflammatory treatment vs placebo in knee osteoarthritis Downloaded from http://journals.lww.com/painrpts by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 03/07/2022 Marina López-Solàa,*, Jesus Pujolb,c, Jordi Monfortd, Joan Deusb,e, Laura Blanco-Hinojob,c, Ben J. Harrisonf, Tor D. Wagerg Abstract Introduction: Many drug trials for chronic pain fail because of high placebo response rates in primary endpoints. Neurophysiological measures can help identify pain-linked pathophysiology and treatment mechanisms. They can also help guide early stop/go decisions, particularly if they respond to verum treatment but not placebo. The neurologic pain signature (NPS), an fMRI-based measure that tracks evoked pain in 40 published samples and is insensitive to placebo in healthy adults, provides a potentially useful neurophysiological measure linked to nociceptive pain. Objectives: This study aims to validate the NPS in knee osteoarthritis (OA) patients and test the effects of naproxen on this signature. Methods: In 2 studies (50 patients, 64.6 years, 75% females), we (1) test the NPS and other control signatures related to negative emotion in knee OA pain patients; (2) test the effect of placebo treatments; and (3) test the effect of naproxen, a routinely prescribed nonsteroidal anti-inflammatory drug in OA. Results: The NPS was activated during knee pain in OA (d 5 1.51, P , 0.001) and did not respond to placebo (d 5 0.12, P 5 0.23). A single dose of naproxen reduced NPS responses (vs placebo, NPS d 5 0.34, P 5 0.03 and pronociceptive NPS component d 5 0.38, P 5 0.02). Naproxen effects were specific for the NPS and did not appear in other control signatures. Conclusion: This study provides preliminary evidence that fMRI-based measures, validated for nociceptive pain, respond to acute OA pain, do not appear sensitive to placebo, and are mild-to-moderately sensitive to naproxen. Keywords: Chronic pain, Osteoarthritis, fMRI, Neurologic pain signature, NSAID, Placebo 1. Introduction treatments can be developed that specifically target them.35 Treatment-related improvement in subjective reports of pain is Pain reports are also highly variable within and across usually the endpoint in clinical trials for drug development in individuals37,72,78 and highly sensitive to social and contextual chronic pain.27 Although pain reduction is the hallmark of factors, which are usually independent of pathophysiol- successful treatment, symptoms (including pain) are often poor ogy36,44,55,56,64,81 but related to placebo responses.81 These indicators of pathophysiology. When sufficient knowledge factors add substantial noise to clinical trials,29,92 resulting in about pathophysiological mechanisms of disease exists, increasing numbers of failed trials.79 There is a need to develop, Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. a Department of Medicine, School of Medicine and Health Sciences, Serra Hunter Faculty Program, University of Barcelona, Barcelona, Spain, b MRI Research Unit, Department of Radiology, Hospital del Mar, Barcelona, Spain, c Centro Investigación Biomédica en Red de Salud Mental, CIBERSAM, Barcelona, Spain, d Rheumatology Department, Hospital del Mar, Barcelona, Spain, e Department of Clinical and Health Psychology, Autonomous University of Barcelona, Barcelona, Spain, f Department of Psychiatry, Melbourne Neuropsychiatry Centre, The University of Melbourne & Melbourne Health, Melbourne, Australia, g Department of Psychological and Brain Sciences, Dartmouth College, Dartmouth, MA, USA *Corresponding author. Address: Unit of Psychological Medicine, Department of Medicine, School of Medicine and Health Sciences, University of Barcelona, Casanova 143, 5a Planta, Barcelona 08036, Spain. Tel.: 00 34 93 403 9299. E-mail address: mlopezsola@ub.edu (M. López-Solà). Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.painrpts.com). Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of The International Association for the Study of Pain. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. PR9 7 (2022) e986 http://dx.doi.org/10.1097/PR9.0000000000000986 7 (2022) e986 www.painreportsonline.com 1 2 · M. López-Solà et al. 7 (2022) e986 PAIN Reports® test, and validate measures of pain-related pathophysiology in distress but they are not correlated with nociceptive pain humans1,6,12,13,22,23,25,30,40–42,50,51,53,60,61,76 that may be used (including its evaluative components.4,18 We expected these as complementary measures of interest in clinical trials. Such signatures to show no response to noxious stimulation and no measures are not intended to replace pain reports23,77 but naproxen effects. rather serve as physiological markers useful to track different This study involves a reanalysis of data from 2 previously outcomes.25,57,74 For example, physiological markers may be published randomized clinical trial fMRI studies29,59 with the novel useful to confirm expected pharmacological effects on the aims and approach of validating NPS responsiveness in 2 physiological processes they are intended to target and such separate cohorts of knee osteoarthritis (OA) patients and results can then be used to make early stop/go decisions in assessing treatment responses to placebo and active pharma- clinical trials.25 cological treatment (Fig. 1). We expected significant, robust NPS Functional magnetic resonance imaging (fMRI) could be a responses to evoked knee pain in OA patients and NPS useful tool for understanding the neurophysiological processes reductions after treatment with naproxen, a nonsteroidal anti- that accompany chronic pain and developing biomarkers for inflammatory drug targeting inflammation through cyclooxyge- nociceptive, cognitive/emotional, and social aspects of nase inhibition.16,19,21,32,43 Based on a recent meta-analysis with pain.7,11,29,31,34,40,49,50,59,61,62,67,68 Prior knowledge about the healthy adults,92 we expected the NPS to be unaffected by functional specialization of brain circuits and their alteration in placebo treatment. We also anticipated a significant response on pain patients complement pain report by adding a neurophys- the SIIPS1 during pain in OA patients, in the same direction as in iological dimension. However, standard fMRI maps of regional healthy adults, but not necessarily effects of naproxen. Finally, we brain activity are neither sensitive nor specific for any particular expected the 2 emotion-related (non–pain-related) measures, the experiential category, including pain.50,82,85,89 Finding increased/ PINES and the Distress Signature, to show neither responses to decreased activity in any one region/circuit is insufficient painful stimulation nor naproxen or placebo effects, given foundation to infer changes in pain-related processes. To previous findings showing they do not respond to painful overcome this limitation, new approaches using pattern recog- stimulation.18,24,88 nition algorithms can identify distributed patterns optimized for sensitivity and specificity to pain and other 2. Methods outcomes85,15,20,39,42,50,51,57,75,80,88,91. Here, we tested a brain measure, the neurologic pain signature (NPS),82 which was Figure 1 summarizes study design and methodological previously validated to track pain across multiple types of evoked approach. We reanalyzed data from 2 previously published noxious stimuli14,38,45–47,50,54,82,88,92 and shows no response to clinical trial studies29,59 registered in the European Clinical Trials several classes of nonpainful aversive events in hu- Database EudraCT (study 1: EudraCT Number 2008-004501- mans.18,38,46,54,82,86,87 The NPS is a distributed pattern that 33, and study 2: EudraCT Number 2009-017468) and also in spans multiple brain regions involved in nociception and pain. It one case (study 2) in ClinicalTrials.gov (Identifier: provides the weights used to calculate a weighted average that NCT01226615) to test an entirely new hypothesis not contem- constitutes a brain-based predicted pain score. The NPS was plated by the original clinical trials, for which study hypotheses developed to predict subjective pain in response to different and primary and secondary outcome measures have been intensities of noxious input and it is tailored to capture the published elsewhere29,59 (see supplementary methods for all association between increasing levels of nociceptive input to the specific method details for each study, available at http://links. brain and increasing pain ratings. Considering that the analgesic lww.com/PR9/A148). Study 1 (Fig. 1) included novel analyses effects of naproxen occur, at least in part, via reducing testing the NPS response in OA patients during knee pain and nociceptive input to the brain as a result of its peripheral anti- the effects of both a conventional anti-inflammatory agent and inflammatory action,21,32,43 we a priori hypothesized that nociceptive-unspecific placebo treatment in a single cohort of naproxen would significantly reduce NPS expression. We also 23 knee OA chronic pain patients (3 study visits in a randomized tested the effects of placebo and naproxen on several “control” order, within-subjects design; mean age 64 6 7.1 years, 80% signatures beyond the NPS, for which we did not anticipate female, all white Caucasian). Functional magnetic resonance naproxen-related reductions. One such signature was the imaging results from a previously published double-blind, Stimulus Intensity–Independent Pain Signature 1 (SIIPS188), a randomized, placebo-controlled clinical trial were used for this brain pattern more related to cognitive-evaluative aspects of pain purpose.29 Study 1 was based on a within-person crossover that predicts pain after controlling for (1) noxious stimulus intensity design, in which each patient participated in 3 separate and (2) NPS expression. We conceptualized the SIIPS1 as a sessions in different days in a randomized order, including a control signature because we did not have previous evidence to placebo session, a naproxen session, and a no treatment hypothesize that naproxen would directly affect brain responses session. The study specifically tested the effects of a single oral associated with the cognitive/evaluative aspects of pain after administration of naproxen on brain responses to painful controlling for nociceptive-specific aspects; however, it is also pressure stimulation in patients with knee OA, at a dose plausible that the SIIPS1 could show effects of naproxen. Finally, previously shown to reduce spinal sensitization.16,58 We we tested 2 non–pain-related control signatures predicting hypothesized that naproxen would reduce activity in the NPS different types of negative emotional experiences (but not when contrasted with placebo because of its well-established pain).24,50 The non–pain-related signatures, (1) the Picture- analgesic effect and clear anti-inflammatory mechanism of Induced Negative Emotion Signature (PINES)18 and (2) the action. Distress Signature,4 are whole-brain weighted patterns that were In study 2 (Fig. 1, N 5 27, one patient cohort, within-subjects developed and validated to predict (1) ratings of negative emotion design with 2 study visits [baseline, no treatment, and placebo] in response to aversive pictures (the PINES18) and (2) ratings of after 120 days of receiving a placebo; 65.6 6 6.2 years, 70.4% empathic distress while listening to others explaining difficult life female, all white Caucasian), data from the placebo arm from an experiences (the Distress Signature4). These signatures capture additional neuroimaging clinical trial was used to replicate the increasing levels of arousal and saliency during different kinds of findings on NPS responses in OA patients during knee pain and 7 (2022) e986 www.painreportsonline.com 3 Figure 1. Study designs and summary of methodological approach. (A) Study designs for study 1 and study 2. The 23 patients in study 1 experienced 3 different visits in separate days: One with no treatment, one after a placebo pill, and a third after a naproxen pill (see Methods and supplementary materials for full description, http://links.lww.com/PR9/A148). The 27 patients in study 2 underwent a first visit (basal, no treatment) and a second visit (final, placebo) after receiving placebo treatment (pills) for 120 days. One contrast image representing [pain activation period (1) . rest (21)] was obtained for each patient and each condition. (B) Summary of the methodology. Each individual contrast image was multiplied by each of the preselected, validated pain-related (NPSp, NPS, and SIIPS1) and control (emotion related, PINES and Distress) signatures (multivariate brain weighted maps that had been previously identified to maximally predict different aspects of pain or emotion in previous studies, see referenced articles). This yields one pattern response score per person per condition. NPSp, pronociceptive neurologic pain signature; NPS, neurologic pain signature; SIIPPS1, Stimulus Intensity–Independent Pain Signature 1. test the effects of extended placebo treatment. The placebo arm rate the subjective pain perceived during the entire fMRI sequence from the neuroimaging clinical trial in study 2 involved 2 visits: visit using a verbally administered numerical rating scale (NRS) ranging 1, before placebo administration; and visit 2, after 120 days of from 0 (“no pain”) to 10 (“extreme pain”).63 double-blind placebo administration (for each patient random- ized to the placebo arm). Study 1 and study 2 were conducted at the Hospital del Mar, Barcelona.29,59 The specific clinical 2.2. Functional magnetic resonance imaging preprocessing characteristics of the patient samples and experimental proce- and single-subject, first-level neuroimaging analysis dures for each study are thoroughly described in the supple- Because of strict word count limit, this section is fully mentary materials and have been previously published in detail. described in the supplementary materials (available at http:// We include the Statistical Analysis Plan for study 1 and study 2 links.lww.com/PR9/A148). In brief, fMRI time series for each and the respective prior publications as supplementary materials study were preprocessed and analyzed using Statistical (available at http://links.lww.com/PR9/A148). Parametric Mapping (SPM8) software, Welcome Department Here, we provide a summary of the common experimental of Imaging Neuroscience, running on Matlab 7.1. Note that the details across both studies. A detailed explanation of common processing and first-level model code is unchanged in SM12, experimental details can be found in the supplementary materials and we confirmed that NPS responses did not differ as a (available at http://links.lww.com/PR9/A148). function of whether contrast images were generated using SPM8 or SPM12. Images were realigned to the first volume of the time series, co- 2.1. Functional magnetic resonance imaging task and registered and normalized to the Montreal Neurological Institute- painful stimuli space provided by SPM (voxel size 5 3 3 3 3 3 mm3) and The same experimental paradigm was used in the scanner for both smoothed with a full width at half maximum Gaussian kernel of studies. The task consisted of a 6-minute sequence alternating 11 8 mm. We provide a detailed description of our motion analyses baseline “rest” periods of 20 seconds (plus a final baseline “rest” and lack of correlation between motion parameters and NPS period of 30 seconds) and 11 painful stimulation periods of 10 expression in the supplementary materials (available at http:// seconds (see the detailed information in the supplementary links.lww.com/PR9/A148). In brief, we verified that the included materials, available at http://links.lww.com/PR9/A148). Immedi- patients had head displacements of less than 2 mm translation ately after the end of the MRI sequence, each subject was asked to and 2˚ rotation, and for both studies, we computed mean 4 · M. López-Solà et al. 7 (2022) e986 PAIN Reports® framewise head displacement for each patient and condition condition, NPSp study 1: t 5 5.93, Cohen d 5 1.24, P , 0.001; following previously published methods.66 NPSp study 2: t 5 4.49, d 5 0.86, P , 0.001; NPS study 1: t 5 Consistent with previous studies,5,29,48,59,65,82 single-subject 8.88, d 5 1.85, P , 0.001; NPS study 2: t 5 6.06, d 5 1.17, P , GLM first-level analyses in SPM included a regressor modelling 0.001; SIIPS1 study 1: t 5 4.47, d 5 0.93, P , 0.001; SIIPS1 pain epochs with a duration of 16 seconds, which is somewhat study 2: t 5 3.04, d 5 0.59, P 5 0.005). The NPSp, NPS, and longer than the 10-second stimulus duration. This is advanta- SIIPS1 were reliably activated in response to knee pain in OA geous for pain because previous studies have found that painful across both studies (mean effect size for NPSp: d 5 1.05, mean stimulation elicits fMRI activity for an extended period, and effect size for NPS: d 5 1.51, mean effect size for SIIPS1: d 5 models with an extended epoch provide better fits to the 0.76; all P’s , 0.001). Neither one dose of placebo (study 1) nor data.5,29,48,59,65,82 This analysis also reproduces the same 120 days of placebo (study 2) were associated with reductions in single-subject, first-level analysis approach presented in the any of the 3 pain-specific signatures: NPSp (Table 1 and Fig. 2; clinical trial studies, which further allows for comparability study 1: t 5 20.26, P 5 0.54; study 2: t 5 0.33, P 5 0.74), NPS between the studies.29,59 (study 1: t 5 21.62, P 5 0.13; study 2: t 5 0.98, P 5 0.33), or SIIPS1 (study 1: t 5 20.33, P 5 0.74; study 2: t 5 20.36, P 5 0.72) responses. 2.3. Brain signatures In study 1, patients were exposed to a single dose of naproxen Information regarding the procedure to compute signature or placebo in a double-blind fashion. The 3 pain-related pattern expression is described in the supplementary materials signatures, ie, NPSp, NPS and SIIPS1, were strongly activated (available at http://links.lww.com/PR9/A148). The NPS includes during pain for the naproxen condition (NPSp: t 5 5.18, P , voxel weights in an a priori defined mask of brain regions that were 0.001; NPS: t 5 7.65, d 5 1.59, P , 0.001, and SIIPS1: t 5 4.26, significantly related to the term “pain” in the Neurosynth meta- P , 0.001). As shown in Figure 2, a single dose of naproxen analytic database (http://neurosynth.org/); see Ref. 87 for a significantly reduced NPSp and NPS responses compared with detailed description. Data outside this mask did not contribute to placebo (NPSp: t 5 22.13, d 5 0.38, P 5 0.02; NPS: t 5 21.90, the pattern expression value. For this analysis, we used a d 5 0.34, P 5 0.03) with a small-to-medium effect size. As previously defined NPS component, the “pronociceptive NPS” anticipated, naproxen did not have an effect on the SIIPS1 pattern (NPSp), which comprised regions likely to be related to (SIIPS1: t 5 0.21, P 5 0.83). nociceptive pain (associated with pain-evoked activation in the Although we had planned a priori contrasts of interest as in NPS).24,52 In this subset of regions, which comprises most of the previous work26,70,82–84 and given the relatively small patient regions in the NPS, activity increases with increasing intensity of samples in each study, we also ran, for completeness, a linear the noxious stimulus. These regions include the major targets of mixed effects repeated-measures analysis including treatment ascending nociceptive afferents, including the thalamus, sec- (categorical factor including the within-subject randomized ondary somatosensory regions (SI/SII), posterior, mid, and conditions no treatment, placebo, and naproxen in study 1) as anterior insula and adjacent opercula, midbrain, dorsal anterior the predictive factor, and NPSp (model 1), NPS (model 2), and cingulate cortex, inferior frontal gyrus, and amygdala (Fig. 1). The SIIPS1 (model 3) responses as the dependent variables in SIIPS1, PINES, and Distress Signature are whole-brain weighted separate models. We found that treatment was a significant patterns identified using machine learning techniques. The predictor (F 5 2.79, P 5 0.03) of NPS responses and did not SIIPS1 was optimized to predict pain ratings in response to reach significance when predicting NPSp responses (F 5 1.96, P acute painful stimulation after controlling for stimulus intensity and 5 0.07). Pairwise comparisons naproxen , placebo were NPS expression. The PINES was optimized to predict the significant in both models (NPS P 5 0.03 and NPSp P 5 0.03). intensity of negative emotion ratings in response to aversive We did not find a significant effect of treatment on SIIPS1 (F 5 images and was shown to be unresponsive to physical pain. The 0.53, P 5 0.95). We did not find a drug , no treatment effect on Distress Signature was optimized to predict moment-by-moment experienced distress while individuals listened to true biographies describing human suffering.4 All these signatures were validated Table 1 in independent test samples that were not included in signature Summary of signature responses for each study, condition and training analyses. Information regarding the linear mixed effects signature. models and planned contrasts run in this study is detailed in the Signature Condition Study 1, mean (SD) Study 2, mean (SD) supplementary materials (available at http://links.lww.com/PR9/ NPS No treatment 2.11 (1.14) 20.43 (17.51) A148). Because we had strongly directional a priori hypotheses Placebo 2.47 (1.12) 16.74 (17.86) about standard planned comparisons (drug , placebo) for the Naproxen 2.05 (1.29) — NPS and NPSp signatures, statistical tests were performed on a NPSp No treatment 1.78 (1.43) 23.83 (26.58) one-tail basis.70,83,84 Placebo 1.97 (1.32) 21.92 (26.48) Naproxen 1.46 (1.35) — SIIPS1 No treatment 102.17 (109.57) 864.1 (1473.30) 3. Results Placebo 107.52 (123.96) 1024.5 (1838.50) Naproxen 101.99 (114.72) — 3.1. Pain signatures respond to evoked knee pain in osteoarthritis patients and are insensitive to placebo PINES No treatment 20.008 (0.082) 21.01 (2.01) Placebo 0.006 (0.089) 20.31 (1.42) 3.1.1. Pronociceptive neurologic pain signature and Naproxen 20.002 (0.097) — neurologic pain signature specifically respond to naproxen Distress No treatment 0.031 (0.26) 20.532 (5.49) We observed robust NPSp, NPS, and SIIPS1 responses during Placebo 20.085 (0.26) 20.434 (5.39) Naproxen 20.037 (0.24) — painful pressure stimulation applied to the medial articular Group mean and SD measures for each signature and condition are shown. interline of the patients’ most affected knee in 2 separate NPSp, pronociceptive neurologic pain signature; NPS, neurologic pain signature; PINES, Picture-Induced randomized clinical trials (Table 1 and Fig. 2, “no treatment” Negative Emotion Signature; SIIPPS, Stimulus Intensity–Independent Pain Signature. 7 (2022) e986 www.painreportsonline.com 5 Figure 2. Pain-related multivariate signatures (previously published and validated) and signature response (dot-product pattern expression) for each signature, for each study and condition. The top row shows a graphic representation of the signature brain weighted maps for reference (and the original publications): the NPSp, the NPS, and the SIIPS1. Individual dots represent signature responses (dot-product pattern expression) for each OA patient in each study (2 separate cohorts, 23 patients in study 1 and 27 patients in study 2). Bars around the mean represent within-person SE bars (Loftus and Masson, 1994). ***P , 0.001, **P , 0.01, *P , 0.05. NPSp, pronociceptive neurologic pain signature; NPS, neurologic pain signature; SIIPPS, Stimulus Intensity–Independent Pain Signature 1. any of the pain-related signatures in this study (all P’s . 0.1). This preplanned contrasts of interest and the linear mixed effects may have been because of a lack of sufficient statistical power to repeated-measures analysis with the emotion-related control detect the difference given a relatively small and variable patient signatures. sample. We also checked the effects of age and gender for both Neither a single dose (study 1) nor 120 days of placebo (study studies and found no effects of age or gender for any of the 2) were associated with reductions in any emotion-related control analyses (all P’s . 0.2). signature: PINES (study 1: t 5 20.69, P 5 0.49; study 2: t 5 2 1.70, P 5 0.12), Distress Signature (study 1: t 5 1.62, P 5 0.12; study 2: t 5 20.08, P 5 0.94) responses (Table 1 and Fig. 3). 3.2. Emotion-related signatures do not respond to evoked In study 1, patients were exposed to a single dose of naproxen knee pain in osteoarthritis patients and are insensitive to or placebo in a double-blind fashion. None of the emotion-related placebo and naproxen control signatures were activated during the pain naproxen As anticipated, neither the PINES nor the Distress Signature condition (PINES: t 5 20.09, P 5 0.93; distress: t 5 20.73, P 5 were significantly positively expressed during pain for any study 0.47; Table 1 and Fig. 3). As shown in Figure 3, a single dose of or group (all P’s . 0.1, with the exception of the PINES, which naproxen did not have an effect on the PINES nor on the Distress was negatively expressed—deactivated—during pain for the no Signature compared with placebo (PINES: t 5 0.29, P 5 0.78; treatment condition in study 2; t 5 22.6, d 5 20.5, P 5 0.01; distress: t 5 20.69, P 5 0.49). Table 1 and Fig. 3). This finding shows specificity of the PINES We also ran a linear mixed effects repeated-measures and Distress Signature, ie, these emotion signatures do not analysis including treatment (categorical factor including the respond to pain in OA. Between-group effects are not meaning- within-subject randomized conditions like no treatment, pla- ful when the signatures are not significantly expressed because cebo, and naproxen in study 1) as the predictive factor and they do not track the psychological experience they were PINES (model 1) and Distress Signature (model 2) responses as developed to track. However, for completion, we run the the dependent variables in separate models. We found that 6 · M. López-Solà et al. 7 (2022) e986 PAIN Reports® treatment was not a significant predictor neither of PINES (F 5 0.14, SEM 5 0.31, P 5 0.49; study 2: t 5 0.67, d 5 0.12, SEM 5 0.20, P 5 0.82) nor of Distress Signature responses (F 5 1.4, P 0.44, P 5 0.51). Single-dose naproxen significantly attenuated 5 0.26). pain ratings (vs single-dose placebo, t 5 2.13, d 5 0.45, SEM 5 0.28, P 5 0.02). For completeness, we also ran a linear mixed effects repeated-measures analysis including treatment (cate- 3.3. Voxel-wise whole brain comparisons gorical factor including the within-subject randomized conditions like no treatment, placebo, and naproxen in study 1) as the For completeness, we performed a voxel-wise whole brain predictive factor and pain ratings as the dependent variable. We analysis for each of our 2 planned contrasts of interest (no found that treatment was a significant predictor (F 5 4.05, P 5 treatment vs placebo for study 1 and for study 2, and placebo vs 0.01) of pain ratings, with significant pairwise comparisons for naproxen for study 1). These analyses tested for significant naproxen , no treatment (t 5 3.04, d 5 0.64, SEM 5 0.27, P 5 effects in brain regions not included in the NPS or NPSp. We 0.004) and naproxen , placebo (t 5 2.13, d 5 0.45, SEM 5 0.28, found no significant differences at P , 0.05; false discovery rate P 5 0.02). corrected for multiple comparisons (corrected within either whole brain or gray matter only). 3.5. Correlation between neurologic pain signature and 3.4. Effects of placebo interventions and naproxen on pain ratings pain ratings There were no significant between-person (individual differences) Table 2 shows pain ratings for each study and condition (mean correlations between NPS or NPSp responses and subjective and SD). Neither of the placebo interventions presented in this pain ratings (for neither study group, all P’s . 0.2). Table 2 shows study modified subjective pain perception (study 1: t 5 0.71, d 5 the summary of the effects of condition on pain ratings. Figure 3. Pain-related multivariate signatures (previously published and validated) and signature response (dot-product pattern expression) for each signature, for each study and condition. The top row shows a graphic representation of the signature brain weighted maps for reference (and the original publications): the PINES (Picture-Induced Negative Emotion Signature) and Distress Signature. Individual dots represent signature responses (dot-product pattern expression) for each OA patient in each study (2 separate cohorts, 23 patients in study 1 and 27 patients in study 2). Bars around the mean represent within-person SE bars (Loftus and Masson, 1994). *P , 0.05. OA, osteoarthritis. 7 (2022) e986 www.painreportsonline.com 7 Table 2 Pain ratings for each study and condition and statistical comparison of pain ratings across conditions for each study. Study 1 Study 2 No treatment Mean (SD) 5 7.00 (1.00) Mean (SD) 5 7.19 (1.07) Placebo Mean (SD) 5 6.78 (1.20) Mean (SD) 5 6.89 (2.39) Naproxen Mean (SD) 5 6.17 (1.11) — Placebo , no treatment t(22) 5 0.71, d 5 0.15, P 5 0.49 t(26) 5 0.67, d 5 0.13, P 5 0.51 Naproxen , placebo t(22) 5 2.13, d 5 0.45, P 5 0.02 — 4. Discussion neither in the NPS nor directly in the SIIPS1. For example, the NPS does not include (or only partially includes) contributions from the The NPS and its specific pronociceptive component, the NPSp, lateral and medial prefrontal cortex, ventral striatum, and some were activated in response to knee pain in OA across studies and brainstem regions. These regions modulate pain responses and did not respond to 2 different kinds of placebo interventions. have been associated with transitions from acute to chronic pain Naproxen, a commonly prescribed anti-inflammatory drug for states2,3,8,10,17,31,52,73,87,88 and represent other potential neuro- chronic OA pain, reduced NPS and NPSp responses beyond physiological treatment targets. Thus, multivariate markers like placebo, in agreement with reductions in pain ratings. We found the NPS or the NPSp can be useful, particularly for limited sample no effect of placebo or naproxen on the SIIPS1, which specifically sizes (eg, early-stage clinical trials and most patient studies tracks pain after controlling for stimulus intensity. We also without major financial backing). checked responses in 2 emotion-related brain markers that have The NPS was developed in young healthy adults during acute shown high sensitivity and specificity for negative emotions in thermal pain in the forearm82; in this study, it is tested in older different contexts unrelated to pain. We found no significant pain- chronic pain patients with pain in the affected knee and during related response of these markers and no significant effects of painful knee pressure. Previous literature provides robust evidence placebo or naproxen, hence providing further proof of specificity for acute knee pain–evoked activation in OA patients in regions to the NPS and NPSp findings. This study provides initial proof of overlapping with the NPS marker, including somatosensory concept that fMRI-based measures validated for nociceptive pain cortices, insula, basal ganglia, thalamus, midbrain, anterior can be sensitive to evoked knee OA pain and to active treatment. cingulate cortex, and amygdala.9,33,59,62,71,90 As anticipated, the Larger samples are required to confirm and extend our results. NPS showed good generalizability to this clinical population, to a The results add utility value to the use of neurophysiological brain different pain modality and when applied to a clinically affected site markers always in combination with main outcome measures of in 2 different OA patient samples. We observed a difference in pain and disability in clinical trials. Multivariate markers like the absolute NPS scale between study 1 and study 2. Multiple factors NPS or the NPSp can be useful particularly in the context of influencing the absolute scale of the NPS response usually differ limited sample sizes (eg, early-stage clinical trials and most across studies, including MRI field strength, different experimental patient studies without major financial backing). Multivariate brain designs, voxel size, and first-level contrast (beta) image weights.54 markers provide a set of interpretable summary measures across Study 1 and study 2 differed in MRI field strength, voxel sizes, and hundreds of thousands of brain voxels, avoiding the need to first-level contrast image weights, which explains absolute scale correct for multiple comparisons when validating marker perfor- differences. Currently, BOLD fMRI responses are not considered mance. Developing and validating new brain markers capitalizing “quantitative” in the sense that one cannot compare absolute on pain modulation mechanisms not captured by the NPS or quantities across studies. The NPS can be used to quantify effect SIIPS1 could help identify neurophysiological effects of treat- sizes for relative comparisons within a study (eg, drug vs placebo), ments that are unrelated to nociceptive-specific factors. The but establishing absolute quantitative values across studies study has some limitations. First, the results for study 1 are based remains a challenge. We did not attempt to equate the absolute on 1.5-Tesla MRI scanner; second, sample sizes for both studies, scale of the NPS response across studies because the reported particularly study 1 (N 5 23 patients), are small, and all patients’ within-study comparisons are unaffected by scale issues. ethnicity was white Caucasian, which accentuates the need for Confirming our initial hypothesis, we found a reduction of future replication in larger, more diverse patient samples. NPS and NPSp responses by naproxen vs placebo. The The lack of placebo effect on NPS, NPSp, or SIIPS1 responses reduction became numerically stronger—numerically larger in suggests that placebo is not targeting the neurophysiological magnitude—when the NPS response was tested specifically on process captured by the NPS—ie, nociceptive processing at the nociceptive regions (NPSp),24,50 which agrees with the ob- brain level—or the SIIPS1—cognitive-evaluative brain processes served reduction in subjective pain reports after naproxen. NPS predicting pain after controlling for stimulus intensity and reductions after naproxen—an anti-inflammatory drug with NPS—even under a regime involving 120 days of placebo previously identified nociceptive effects at the peripheral and administration. However, placebo treatment did not significantly central nervous system levels16,19,21,32,43,58—argues in favor of affect pain in these samples, so it is possible that a more the NPS and NPSp as good summary measures of drug effects “powerful placebo” would have shown an effect on the NPS or (vs placebo) on nociceptive processing in the human brain. NPS SIIPS1. Previous studies that did show placebo effects on pain and NPSp reductions during naproxen vs placebo align with also showed null or very small effects on the NPS,82,92 suggesting findings from a neuroimaging placebo-controlled trial testing that even effective placebo manipulations may have much smaller naproxen effects on brain activity.71 The study showed that effects on the NPS than they do on pain reports. naproxen reduced brain activation over placebo in bilateral There are multiple other brain, spinal, and peripheral mecha- primary somatosensory cortex, thalamus, and amygdala: all nisms that contribute to modulating pain that are not represented regions included in the NPS/NPSp. In the same line, in the 8 · M. López-Solà et al. 7 (2022) e986 PAIN Reports® previously published clinical trial results from study 1,29 our Appendix A. Supplemental digital content group found preliminary (uncorrected) effects of naproxen in the Supplemental digital content associated with this article can be second somatosensory cortex, bilateral insula, basal ganglia, found online at http://links.lww.com/PR9/A148. ACC, and amygdala. Although these studies provide detailed insight about the brain regions that were modulated by Article history: naproxen over placebo, they lacked sufficient statistical power Received 23 June 2021 to survive correction for multiple comparisons. Received in revised form 2 December 2021 The current results regarding naproxen effects on the NPS and Accepted 11 December 2021 NPSp require replication in larger samples and using different naproxen doses, particularly because the effects were specific to the comparison naproxen , placebo and were small to medium in References effect size. The rationale for using naproxen to test its effects on the [1] Ablin JN, Buskila D, Clauw DJ. Biomarkers in fibromyalgia. Curr Pain NPS over placebo was based on its well-established antinociceptive Headache Rep 2009;13:343–9. action,16,19,21,32,43 which was deemed optimal to test the hypothesis [2] Apkarian AV, Baliki MN, Farmer MA. Predicting transition to chronic pain. Curr Opin Neurol 2013;26:360–7. that a drug with known antinociceptive effects should significantly [3] Apkarian AV, Hashmi JA, Baliki MN. Pain and the brain: specificity and reduce NPS beyond the nociceptive-unspecific effects of placebo. plasticity of the brain in clinical chronic pain. PAIN 2011;152:S49–64. That said, the study does not provide data for comparison with [4] Ashar YK, Andrews-Hanna JR, Dimidjian S, Wager TD. Empathic care healthy controls, other forms of knee-free chronic pain patients or, and distress: predictive brain markers and dissociable brain systems. Neuron 2017;94:1263–73.e4. importantly, other forms of treatment. Future studies should [5] Atlas LY, Bolger N, Lindquist MA, Wager TD. Brain mediators of predictive compare the effects of naproxen with other commonly used cue effects on perceived pain. J Neurosci 2010;30:12964–77. pharmacological, psychological, and physical therapies for OA. [6] Bagarinao E, Johnson KA, Martucci KT, Ichesco E, Farmer MA, Labus J, Neurophysiological biomarkers in combination with conventional Ness TJ, Harris R, Deutsch G, Apkarian AV, Mayer EA, Clauw DJ, Mackey outcome measures in clinical trials for pain may show potential for S. Preliminary structural MRI based brain classification of chronic pelvic pain: a MAPP network study. PAIN 2014;155:2502–9. helping our understanding of the effects of different treatments on [7] Baliki MN, Baria AT, Apkarian aV. The cortical rhythms of chronic back previously characterized and validated neurophysiological compo- pain. J Neurosci 2011;31:13981–90. nents of pain. Future studies may successfully develop new markers [8] Baliki MN, Geha PY, Fields HL, Apkarian AV. Predicting value of pain and of spontaneous pain, which may complement information summa- analgesia: nucleus accumbens response to noxious stimuli changes in the presence of chronic pain. Neuron 2010;66:149–60. rized by the NPS/NPSp by relying on partially nonoverlapping brain [9] Baliki MN, Geha PY, Jabakhanji R, Harden N, Schnitzer TJ, Apkarian AV. circuits.7,9,28,62 By testing new brain markers that show sensitivity A preliminary fMRI study of analgesic treatment in chronic back pain and and specificity for different types of human pain experiences, acute knee osteoarthritis. Mol Pain 2008;4:47. and chronic, evoked and spontaneous,69,77,85 involving different [10] Baliki MN, Petre B, Torbey S, Herrmann KM, Huang L, Schnitzer TJ, pain modalities and in different body locations, we can start Fields HL, Apkarian AV. Corticostriatal functional connectivity predicts transition to chronic back pain. Nat Neurosci 2012;15:1117–19. generating more clinically translatable imaging models with the [11] Borsook D, Becerra L, Fava M. Use of functional imaging across clinical potential to optimize current and future treatments. phases in CNS drug development. Transl Psychiatry 2013;3:e282. [12] Borsook D, Becerra L, Hargreaves R. Biomarkers for chronic pain and analgesia. Part 1: the need, reality, challenges, and solutions. Discov Med Disclosures 2011;11:197–207. [13] Borsook D, Becerra L, Hargreaves R. Biomarkers for chronic pain and The authors have no conflicts of interest to declare. analgesia. Part 2: how, where, and what to look for using functional imaging. Discov Med 2011;11:209–19. [14] Bräscher A-K, Becker S, Hoeppli M-E, Schweinhardt P. Different brain circuitries mediating controllable and uncontrollable pain. J Neurosci Acknowledgements 2016;36:5013–25. Dr Lopez-Sola was in part supported by the Serra Hunter faculty [15] Brown JE, Chatterjee N, Younger J, Mackey S. Towards a physiology- based measure of pain: patterns of human brain activity distinguish program at the University of Barcelona. This work was in part painful from non-painful thermal stimulation. PLoS One 2011;6:e24124. supported by grants R01AR074795 (M.L.S.), R01DA035484 [16] Burian M, Geisslinger G. COX-dependent mechanisms involved in the (T.D.W.), R01DA035484 (T.D.W.), 3R01MH076136 (T.D.W.), antinociceptive action of NSAIDs at central and peripheral sites. R01EB026549 (T.D.W.), PSI2014-5352 (J.D.), and PSI2017- Pharmacol Ther 2005;107:139–54. 83777 (J.D.) and by the Agency of University Research Funding [17] Bushnell MC, Ceko M, Low LA. Cognitive and emotional control of pain and its disruption in chronic pain. Nat Rev Neurosci 2013;14:502–11. Management of the Catalonia Government (Research Group [18] Chang LJ, Gianaros PJ, Manuck SB, Krishnan A, Wager TD. A sensitive SGR2017-1198). The original studies were funded in part by the and specific neural signature for picture-induced negative affect. PLoS Ministry of Science and Innovation of Spain (Grant SAF2010- Biol 2015;13:e1002180. 19434), the Neuroscience Centre of Excellence for Drug [19] Cheng DS, Visco CJ. Pharmaceutical therapy for osteoarthritis. PM R 2012;4:S82–8. Discovery (CEDD)—GlaxoSmithKline (X71111961 and [20] Cheng JC, Bosma RL, Hemington KS, Kucyi A, Lindquist MA, Davis KD. A3360529) and Bioibérica, S.A. (CS/IV-RMF-01). Slow-5 dynamic functional connectivity reflects the capacity to sustain The clinical trials included in this study were approved by the local cognitive performance during pain. Neuroimage 2017;157:61–8. Ethics Committee (Clinical Research Ethical Committee—Institut [21] Cicala C, Ianaro A, Fiorucci S, Calignano A, Bucci M, Gerli R, Santucci L, Municipal d’Assistència Sanitària, Barcelona) and conducted in Wallace JL, Cirino G. NO-naproxen modulates inflammation, nociception and downregulates T cell response in rat Freund’s adjuvant arthritis. Br J compliance with the Code of Ethics of the World Medical Pharmacol 2000;130:1399–405. Association (Declaration of Helsinki). [22] Dadabhoy D, Crofford LJ, Spaeth M, Russell IJ, Clauw DJ. Biology and Author contributions: M. López-Solà and T. D. Wager conceived therapy of fibromyalgia. Evidence-based biomarkers for fibromyalgia the study. M. López-Solà and T. D. Wager wrote the paper. syndrome. Arthritis Res Ther 2008;10:211. [23] Davis KD, Flor H, Greely HT, Iannetti GD, Mackey S, Ploner M, Pustilnik A, M. López-Solà analyzed the data. J. Pujol, J. Monfort, J. Deus, Tracey I, Treede R-D, Wager TD. Brain imaging tests for chronic pain: L. Blanco-Hinojo, and B. J. Harrison ran the studies and medical, legal and ethical issues and recommendations. Nat Rev Neurol contributed to writing the paper. 2017;13:624–38. 7 (2022) e986 www.painreportsonline.com 9 [24] Duff EP, Moultrie F, van der Vaart M, Goksan S, Abos A, Fitzgibbon SP, clinical pain using multimodal neuroimaging and autonomic metrics. PAIN Baxter L, Wager TD, Slater R. Inferring pain experience in infants using 2019;160:550–60. quantitative whole-brain functional MRI signatures: a cross-sectional, [43] Levy D, Zhang X-C, Jakubowski M, Burstein R. Sensitization of meningeal observational study. Lancet Digit Health 2020;2:e458–67. nociceptors: inhibition by naproxen. Eur J Neurosci 2008;27:917–22. [25] Duff EP, Vennart W, Wise RG, Howard MA, Harris RE, Lee M, [44] Linde K, Fässler M, Meissner K. Placebo interventions, placebo effects Wartolowska K, Wanigasekera V, Wilson FJ, Whitlock M, Others. and clinical practice. Philos Trans R Soc Lond B Biol Sci 2011;366: Learning to identify CNS drug action and efficacy using multistudy fMRI 1905–12. data. Sci Transl Med 2015;7:274ra16. [45] López-Solà M, Geuter S, Koban L, Coan JA, Wager TD. Brain [26] Evidence-Based Medicine Working Group. Evidence-based medicine. A mechanisms of social touch-induced analgesia in females. PAIN 2019; new approach to teaching the practice of medicine. JAMA 1992;268: 160:2072–85. 2420–5. [46] López-Solà M, Koban L, Krishnan A, Wager TD. When pain really matters: [27] Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM. Clinical A vicarious-pain brain marker tracks empathy for pain in the romantic importance of changes in chronic pain intensity measured on an 11-point partner. Neuropsychologia 2020;145:106427. numerical pain rating scale. PAIN 2001;94:149–58. [47] López-Solà M, Koban L, Wager TD. Transforming pain with prosocial [28] Geha PY, Baliki MN, Chialvo DR, Harden RN, Paice JA, Apkarian AV. meaning: a functional magnetic resonance imaging study. Psychosom Brain activity for spontaneous pain of postherpetic neuralgia and its Med 2018;80:814–25. modulation by lidocaine patch therapy. PAIN 2007;128:88–100. [48] López-Solà M, Pujol J, Hernández-Ribas R, Harrison BJ, Ortiz H, [29] Giménez M, Pujol J, Ali Z, López-Solà M, Contreras-Rodrı́guez O, Deus J, Soriano-Mas C, Deus J, Menchón JM, Vallejo J, Cardoner N. Dynamic Ortiz H, Soriano-Mas C, Llorente-Onaindia J, Monfort J. Naproxen effects assessment of the right lateral frontal cortex response to painful on brain response to painful pressure stimulation in patients with knee stimulation. Neuroimage 2010;50:1177–87. osteoarthritis: a double-blind, randomized, placebo-controlled, single- [49] López-Solà M, Pujol J, Wager TD, Garcia-Fontanals A, Blanco-Hinojo L, dose study. J Rheumatol 2014;41:2240–8. Garcia-Blanco S, Poca-Dias V, Harrison BJ, Contreras-Rodrı́guez O, [30] Harris RE, Sundgren PC, Craig AD, Kirshenbaum E, Sen A, Napadow V, Monfort J, Garcia-Fructuoso F, Deus J. Altered functional magnetic Clauw DJ. Elevated insular glutamate in fibromyalgia is associated with resonance imaging responses to nonpainful sensory stimulation in experimental pain. Arthritis Rheum Official J Am Coll Rheumatol 2009;60: fibromyalgia patients. Arthritis Rheumatol (Hoboken, NJ) 2014;66: 3146–52. 3200–9. [31] Hashmi JA, Baliki MN, Huang L, Baria AT, Torbey S, Hermann KM, [50] López-Solà M, Woo C-W, Pujol J, Deus J, Harrison BJ, Monfort J, Wager Schnitzer TJ, Apkarian AV. Shape shifting pain: chronification of back TD. Towards a neurophysiological signature for fibromyalgia. PAIN 2017; pain shifts brain representation from nociceptive to emotional circuits. 158:34–47. Brain 2013;136:2751–68. [51] Mackey S, Greely HT, Martucci KT. Neuroimaging-based pain [32] Hautaniemi T, Petrenko N, Skorinkin A, Giniatullin R. The inhibitory action biomarkers: definitions, clinical and research applications, and of the antimigraine nonsteroidal anti-inflammatory drug naproxen on evaluation frameworks to achieve personalized pain medicine. Pain Rep P2X3 receptor-mediated responses in rat trigeminal neurons. 2019;4:e762. Neuroscience 2012;209:32–8. [52] Martucci KT, MacNiven KH, Borg N, Knutson B, Mackey SC. Apparent [33] Hiramatsu T, Nakanishi K, Yoshimura S, Yoshino A, Adachi N, Okamoto effects of opioid use on neural responses to reward in chronic pain. Sci Y, Yamawaki S, Ochi M. The dorsolateral prefrontal network is involved in Rep 2019;9:9633. pain perception in knee osteoarthritis patients. Neurosci Lett 2014;581: [53] Martucci KT, Shirer WR, Bagarinao E, Johnson KA, Farmer MA, Labus 109–14. JS, Apkarian AV, Deutsch G, Harris RE, Mayer EA, Clauw DJ, Greicius [34] Ichesco E, Schmidt-Wilcke T, Bhavsar R, Clauw DJ, Peltier SJ, Kim J, MD, Mackey SC. The posterior medial cortex in urologic chronic pelvic Napadow V, Hampson JP, Kairys AE, Williams DA, Harris RE. Altered pain syndrome: detachment from default mode network-a resting- resting state connectivity of the insular cortex in individuals with state study from the MAPP Research Network. PAIN 2015;156: fibromyalgia. J Pain 2014;15:815–26.e1. 1755–64. [35] Jensen TS, Baron R. Translation of symptoms and signs into [54] Ma Y, Wang C, Luo S, Li B, Wager TD, Zhang W, Rao Y, Han S. Serotonin mechanisms in neuropathic pain. PAIN 2003;102:1–8. transporter polymorphism alters citalopram effects on human pain [36] Kaptchuk TJ, Kelley JM, Conboy LA, Davis RB, Kerr CE, Jacobson EE, responses to physical pain. Neuroimage 2016;135:186–96. Kirsch I, Schyner RN, Nam BH, Nguyen LT, Park M, Rivers AL, McManus [55] Meissner K, Bingel U, Colloca L, Wager TD, Watson A, Flaten MA. The C, Kokkotou E, Drossman DA, Goldman P, Lembo AJ. Components of placebo effect: advances from different methodological approaches. placebo effect: randomised controlled trial in patients with irritable bowel J Neurosci 2011;31:16117–24. syndrome. BMJ 2008;336:999–1003. [56] Meissner K, Distel H, Mitzdorf U. Evidence for placebo effects on physical [37] de Koning EJ, Timmermans EJ, van Schoor NM, Stubbs B, van den but not on biochemical outcome parameters: a review of clinical trials. Kommer TN, Dennison EM, Limongi F, Castell MV, Edwards MH, Queipo BMC Med 2007;5:3. R, Cooper C, Siviero P, van der Pas S, Pedersen NL, Sánchez-Martı́nez [57] van der Miesen MM, Lindquist MA, Wager TD. Neuroimaging-based M, Deeg DJH, Denkinger MD; EPOSA Group. Within-person pain biomarkers for pain: state of the field and current directions. Pain Rep variability and mental health in older adults with osteoarthritis: an 2019;4:e751. analysis across 6 European cohorts. J Pain 2018;19:690–8. [58] Miranda JA, Stanley P, Gore K, Turner J, Dias R, Rees H. A preclinical [38] Krishnan A, Woo C-W, Chang LJ, Ruzic L, Gu X, López-Solà M, Jackson physiological assay to test modulation of knee joint pain in the spinal cord: PL, Pujol J, Fan J, Wager TD. Somatic and vicarious pain are represented effects of oxycodone and naproxen. PLoS One 2014;9:e106108. by dissociable multivariate brain patterns. Elife 2016;5:e15166. doi: [59] Monfort J, Pujol J, Contreras-Rodrı́guez O, Llorente-Onaindia J, López- 10.7554/eLife.15166. Solà M, Blanco-Hinojo L, Vergés J, Herrero M, Sánchez L, Ortiz H, [39] Kucyi A, Davis KD. The dynamic pain connectome. Trends Neurosci Montañés F, Deus J, Benito P. Effects of chondroitin sulfate on brain 2015;38:86–95. response to painful stimulation in knee osteoarthritis patients. A [40] Kutch JJ, Ichesco E, Hampson JP, Labus JS, Farmer MA, Martucci KT, randomized, double-blind, placebo-controlled functional magnetic Ness TJ, Deutsch G, Apkarian AV, Mackey SC, Klumpp DJ, Schaeffer AJ, resonance imaging study. Med Clin 2017;148:539–47. Rodriguez LV, Kreder KJ, Buchwald D, Andriole GL, Lai HH, Mullins C, [60] Napadow V, Harris RE. What has functional connectivity and chemical Kusek JW, Landis JR, Mayer EA, Clemens JQ, Clauw DJ, Harris RE; neuroimaging in fibromyalgia taught us about the mechanisms and MAPP Research Network. Brain signature and functional impact of management of “centralized” pain?. Arthritis Res Ther 2014;16:425. centralized pain: a multidisciplinary approach to the study of chronic [61] Napadow V, Kim J, Clauw DJ, Harris RE. Decreased intrinsic brain pelvic pain (MAPP) network study. Pain 2017;158:1979–91. connectivity is associated with reduced clinical pain in fibromyalgia. [41] Kutch JJ, Labus JS, Harris RE, Martucci KT, Farmer MA, Fenske S, Arthritis Rheum 2012;64:2398–403. Fling C, Ichesco E, Peltier S, Petre B, Guo W, Hou X, Stephens AJ, [62] Parks EL, Geha PY, Baliki MN, Katz J, Schnitzer TJ, Apkarian AV. Brain Mullins C, Clauw DJ, Mackey SC, Apkarian AV, Landis JR, Mayer EA; activity for chronic knee osteoarthritis: dissociating evoked pain from MAPP Research Network. Resting-state functional connectivity spontaneous pain. Eur J Pain 2011;15:843.e1–14. predicts longitudinal pain symptom change in urologic chronic [63] Price DD, Bush FM, Long S, Harkins SW. A comparison of pain pelvic pain syndrome: a MAPP network study. PAIN 2017;158: measurement characteristics of mechanical visual analogue and simple 1069–82. numerical rating scales. PAIN 1994;56:217–26. [42] Lee J, Mawla I, Kim J, Loggia ML, Ortiz A, Jung C, Chan S-T, Gerber J, [64] Price DD, Finniss DG, Benedetti F. A comprehensive review of the Schmithorst VJ, Edwards RR, Wasan AD, Berna C, Kong J, Kaptchuk TJ, placebo effect: recent advances and current thought. Annu Rev Psychol Gollub RL, Rosen BR, Napadow V. Machine learning-based prediction of 2008;59:565–90. 10 · M. López-Solà et al. 7 (2022) e986 PAIN Reports® [65] Pujol J, López-Solà M, Ortiz H, Vilanova JC, Harrison BJ, Yücel M, as biomarkers in chronic pain clinical trials: IMMPACT considerations. Soriano-Mas C, Cardoner N, Deus J. Mapping brain response to pain in J Pain 2017;18:757–77. fibromyalgia patients using temporal analysis of FMRI. PLoS One 2009;4: [77] Tracey I, Woolf CJ, Andrews NA. Composite pain biomarker signatures e5224. for objective assessment and effective treatment. Neuron 2019;101: [66] Pujol J, Macià D, Blanco-hinojo L, Martı́nez-vilavella G, Sunyer J, De R, 783–800. Caixàs A, Martı́n-santos R, Deus J, Harrison BJ. NeuroImage Does [78] Tupper SM, Rosenberg AM, Pahwa P, Stinson JN. Pain intensity motion-related brain functional connectivity reflect both artifacts and variability and its relationship with quality of life in youths with juvenile genuine neural activity ?. Neuroimage 2014;101:87–95. idiopathic arthritis. Arthritis Care Res 2013;65:563–70. [67] Pujol J, Macià D, Garcia-Fontanals A, Blanco-hinojo L, López-Solà M, [79] Tuttle AH, Tohyama S, Ramsay T, Kimmelman J, Schweinhardt P, Garcia-Blanco S, Poca-Dias V, Harrison BJ, Contreras-Rodrı́guez O, Bennett GJ, Mogil JS. Increasing placebo responses over time in US Monfort J, Garcia-Fructuoso F, Deus J. The contribution of sensory clinical trials of neuropathic pain. PAIN 2015;156:2616–26. system functional connectivity reduction to clinical pain in fibromyalgia. [80] Tu Y, Ortiz A, Gollub RL, Cao J, Gerber J, Lang C, Park J, Wilson G, Shen PAIN 2014.155:1492–1503. doi: 10.1016/j.pain.2014.04.028. W, Chan S-T, Wasan AD, Edwards RR, Napadow V, Kaptchuk TJ, Rosen [68] Pujol J, Martı́nez-Vilavella G, Llorente-Onaindia J, Harrison BJ, López- B, Kong J. Multivariate resting-state functional connectivity predicts Solà M, López-Ruiz M, Blanco-Hinojo L, Benito P, Deus J, Monfort J. responses to real and sham acupuncture treatment in chronic low back Brain imaging of pain sensitization in patients with knee osteoarthritis. pain. Neuroimage Clin 2019;23:101885. PAIN 2017;158:1831–8. [81] Wager TD, Atlas LY. The neuroscience of placebo effects: connecting [69] Reckziegel D, Vachon-Presseau E, Petre B, Schnitzer TJ, Baliki MN, context, learning and health. Nat Rev Neurosci 2015;16:403–18. Apkarian AV. Deconstructing biomarkers for chronic pain: context- and [82] Wager TD, Atlas LY, Lindquist Ma, Roy M, Woo C-W, Kross E. An fMRI- hypothesis-dependent biomarker types in relation to chronic pain. PAIN based neurologic signature of physical pain. N Engl J Med 2013;368: 2019;160(Suppl 1):S37S48. 1388–97. [70] Sacks H, Chalmers TC, Smith H Jr. Randomized versus historical [83] Wager TD, Fields H. Placebo analgesia. Textbook of pain 2013:362373. controls for clinical trials. Am J Med 1982;72:233–40. [84] Wager TD, Rilling JK, Smith EE, Sokolik A, Casey KL, Davidson RJ, [71] Sanders D, Krause K, O’Muircheartaigh J, Thacker MA, Huggins JP, Kosslyn SM, Rose RM, Cohen JD. Placebo-induced changes in FMRI in Vennart W, Massat NJ, Choy E, Williams SCR, Howard MA. the anticipation and experience of pain. Science 2004;303:1162–7. Pharmacologic modulation of hand pain in osteoarthritis: a double-blind [85] Woo C-W, Chang LJ, Lindquist MA, Wager TD. Building better placebo-controlled functional magnetic resonance imaging study using biomarkers: brain models in translational neuroimaging. Nat Neurosci naproxen. Arthritis Rheumatol 2015;67:741–51. 2017;20:365–77. [72] Schneider S, Junghaenel DU, Keefe FJ, Schwartz JE, Stone AA, [86] Woo C-W, Koban L, Kross E, Lindquist MA, Banich MT, Ruzic L, Broderick JE. Individual differences in the day-to-day variability of pain, Andrews-Hanna JR, Wager TD. Separate neural representations for fatigue, and well-being in patients with rheumatic disease: associations physical pain and social rejection. Nat Commun 2014;5:5380. with psychological variables. PAIN 2012;153:813–22. [87] Woo C-W, Roy M, Buhle JT, Wager TD. Distinct brain systems mediate [73] Seminowicz DA, Remeniuk B, Krimmel SR, Smith MT, Barrett FS, Wulff the effects of nociceptive input and self-regulation on pain. Plos Biol AB, Furman AJ, Geuter S, Lindquist MA, Irwin MR, Finan PH. Pain-related 2015;13:e1002036. nucleus accumbens function: modulation by reward and sleep [88] Woo C-W, Schmidt L, Krishnan A, Jepma M, Roy M, Lindquist MA, Atlas disruption. PAIN 2019;160:1196–207. LY, Wager TD. Quantifying cerebral contributions to pain beyond [74] Silver Spring (MD): Food and Drug Administration (US). Co-published nociception. Nat Commun 2017;8:14211. by National Institutes of Health (US), Bethesda (MD). FDA-NIH [89] Yarkoni T, Poldrack RA, Nichols TE, Van Essen DC, Wager TD. Large- Biomarker Working Group. BEST (Biomarkers, EndpointS, and scale automated synthesis of human functional neuroimaging data. Nat other Tools). 2016. Available at: https://www.ncbi.nlm.nih.gov/ Methods 2011;8:665–70. books/NBK326791/. [90] Yue Y, Collaku A. Correlation of pain reduction with fMRI BOLD response [75] Smith A, López-Solà M, McMahon K, Pedler A, Sterling M. Multivariate in osteoarthritis patients treated with paracetamol: randomized, double- pattern analysis utilizing structural or functional MRI-In individuals with blind, crossover clinical efficacy study. Pain Med 2018;19:355–67. musculoskeletal pain and healthy controls: a systematic review. Semin [91] Zhang B, Jung M, Tu Y, Gollub R, Lang C, Ortiz A, Park J, Wilson G, Arthritis Rheum 2017;47:418–31. Gerber J, Mawla I, Chan S-T, Wasan A, Edwards R, Lee J, Napadow V, [76] Smith SM, Dworkin RH, Turk DC, Baron R, Polydefkis M, Tracey I, Kaptchuk T, Rosen B, Kong J. Identifying brain regions associated with Borsook D, Edwards RR, Harris RE, Wager TD, Arendt-Nielsen L, Burke the neuropathology of chronic low back pain: a resting-state amplitude of LB, Carr DB, Chappell A, Farrar JT, Freeman R, Gilron I, Goli V, Haeussler low-frequency fluctuation study. Br J Anaesth 2019;123:e303–11. J, Jensen T, Katz NP, Kent J, Kopecky EA, Lee DA, Maixner W, Markman [92] Zunhammer M, Bingel U, Wager TD. Placebo effects on the JD, McArthur JC, McDermott MP, Parvathenani L, Raja SN, Rappaport neurologic pain signature: a meta-analysis of individual participant BA, Rice ASC, Rowbotham MC, Tobias JK, Wasan AD, Witter J. The functional magnetic resonance imaging data. JAMA Neurol 2018;75: potential role of sensory testing, skin biopsy, and functional brain imaging 1321–30.