Enhancing Natural Reward Responsiveness Among Opioid Users Predicts Chronic Pain Relief: EEG Analyses From a Trial of Mindfulness-Oriented Recovery Enhancement

Abstract

Objective

Although opioid-treated chronic pain patients evidence blunted responsiveness to natural rewards, focusing on naturally rewarding stimuli can produce analgesia in these patients. A prior randomized controlled trial (RCT) demonstrated that a social work intervention—Mindfulness-Oriented Recovery Enhancement (MORE)—enhanced natural reward processing as indicated by event-related brain potentials (ERPs). The aim of the present study was to perform a secondary data analysis on ERPs collected in this RCT to explore whether improvements in electrocortical response to natural reward predicted pain relief.

Method

The sample for this secondary analysis included opioid-treated chronic pain patients with complete ERP data (N = 29). Participants were randomized to 8 weeks of MORE or a support group control condition, and ERPs to images representing naturally rewarding stimuli were measured before and after 8 weeks of treatment. We explored associations between changes in brain reward response, chronic pain symptoms, and pain coping.

Results

Increases in ERP reward response were significantly associated with decreased pain severity from pretreatment to posttreatment (β = −.48, p = .007) and improvements in pain catastrophizing (β = −.36, p = .05) and diverting attention as a means of pain coping (β = .38, p = .043). Increased ERP reward response predicted decreased pain interference by 3-month follow-up (β = −.37, p = .048).

Conclusions

Chronic pain may be alleviated by enhancing brain response to natural rewards.

Chronic pain is prevalent in modern Western society. As many as 100 million Americans suffer from chronic pain at any given time—more than all those with heart disease, diabetes, and cancer combined (Institute of Medicine, 2011). Chronic pain—defined as pain that lasts more than 90 days beyond the time expected for normal healing of damaged tissue (Dowell, Haegerich, & Chou, 2016)—is associated with a range of adverse impacts (Institute of Medicine, 2011), including reduced physical functioning, increased work absenteeism, and diminished quality of life due to withdrawal from pleasurable activities and experiences. Following the development and marketing of new opioid drugs in the 1990s, physicians began to more liberally prescribe opioids to treat chronic pain (National Academies of Sciences, Engineering, and Medicine, 2017). Consequently, opioid prescriptions had climbed over two decades to 208 million by 2011, a dramatic increase that was paralleled by an epidemic of opioid misuse and addiction (National Academies of Sciences, Engineering, and Medicine, 2017). Meta-analysis suggests that approximately 25% of chronic pain patients who are prescribed opioids engage in opioid-misusing behaviors such as unauthorized opioid dose escalation or use of opioids to self-medicate dysphoric mood states (Vowles et al., 2015).

Although acute opioid administration can produce euphoria, protracted or high-dose opioid use disrupts functioning in corticostriatal brain networks that mediate reward processing (Cahill, Walwyn, Taylor, Pradhan, & Evans, 2016; Kim, Ham, Hong, Moon, & Im, 2016), decreasing physiological responsivity to naturally rewarding objects and events (Garland, Bryan, Nakamura, Froeliger, & Howard, 2017; Garland, Froeliger, & Howard, 2015a) and leading to reduced ability to experience pleasure in everyday life (i.e., hedonic capacity). Evidence for hedonic deficits and dysregulation of natural reward processing may be revealed by cue-elicited event-related brain potentials (ERPs), which reflect the brain’s electrical activity (measured by electrodes placed on the scalp) in response to discrete perceptual, cognitive, or emotional events (Luck, 2014). ERP measurements have high temporal resolution and are therefore able to capture the rapid execution of psychological functions (Cacioppo, Tassinary, & Berntson, 2007), such as fleeting attention to natural rewards. Among opioid dependent individuals, studies demonstrate reduced ERPs to images depicting natural rewards (Lubman, Allen, Peters, & Deakin, 2007, 2008)—a predictor of future drug use (Lubman et al., 2009).

The anhedonic effects of prolonged opioid exposure (Shurman, Koob, & Gutstein, 2010) may compound the negative mood and blunting of reward function associated with chronic pain itself (Elvemo, Landrø, Borchgrevink, & Håberg, 2015; Marbach, Richlin, & Lipton, 1983). Neuroscientific evidence indicates that chronic pain leads to hedonic dysregulation via allostatic changes in stress (e.g., amygdala) and reward systems (e.g., ventral striatum, orbitofrontal cortex) in the brain (Borsook et al., 2016), marked by heightened sensitivity to aversive experience (e.g., pain) and insensitivity to natural rewards. Such brain changes may be accompanied by a form of “cognitive myopia” in which pain (like other aversive experiences and negative emotional states) can narrow and tune attention to focus on harmful and threatening stimuli (Schoth, Nunes, & Liossi, 2012)—a negative attentional bias that may fuel pain catastrophizing (Quartana, Campbell, & Edwards, 2009) and prevent one from noticing and appreciating pleasant events and experiences (Garland, Fredrickson, et al., 2010). Ultimately, the independent and interactive effects of chronic pain and extended opioid therapy on brain reward systems (Elman & Borsook, 2016) may promote opioid dose escalation as a means of preserving a dwindling sense of well-being, resulting in further blunting of reward function and hypersensitization to stress and pain—a downward spiral of hedonic dysregulation that promotes loss of control over opioid use (Garland, Froeliger, Zeidan, Partin, & Howard, 2013). Such loss of control may have serious consequences, including increased risk of overdose and escalation to heroin addiction when access to prescription opioids is limited (Kolodny et al., 2015).

In contrast, experimental pain research shows that inducing pleasure and positive emotions in the laboratory via pleasant visual, auditory, or olfactory stimuli can exert a significant analgesic effect (for a review, see Finan & Garland, 2015). Further, reward receipt also decreases pain perception via activation of medial prefrontal cortical structures (i.e., orbitofrontal cortex; Becker, Gandhi, Pomares, Wager, & Schweinhardt, 2017). Although the mechanisms of reward-based and positive-affect analgesia are far from understood, from a cognitive perspective focusing on naturally rewarding experiences might divert attention from pain and thereby decrease its severity and adverse consequences (Finan & Garland, 2015). In light of this research on positive-affect analgesia, it stands to reason that therapeutic interventions designed to increase pleasure and positive emotions by augmenting natural reward processing may reduce suffering and disability produced by chronic pain. Plausibly, therapies that provide training in attending to and savoring naturally rewarding objects and experiences may boost natural reward processing. Mindfulness training, which has been shown to increase attentional control (Tang, Hölzel, & Posner, 2015) and positive emotions (Fredrickson et al., 2017) may enhance natural reward processing. Although not the explicit aim of most mindfulness-based interventions, mindfulness training might nonetheless increase pleasure from perceptual and sensorimotor experiences in a fashion similar to sensate-focus techniques (Masters & Johnson, 1970) and promote positive emotion regulation by amplifying selective attentional processes (Wadlinger & Isaacowitz, 2010). Recently, it was hypothesized that the practice of using mindfulness to savor the sensory features of pleasant objects, people, and events—and the positive emotions that flow from experiencing them—might remediate hedonic dysregulation and reward processing deficits in the brain (Garland, 2016).

To test this mindful savoring hypothesis, we conducted a randomized controlled trial (RCT) of Mindfulness-Oriented Recovery Enhancement (MORE), a novel social work intervention that integrates training in savoring natural rewards with mindfulness training as a means of addressing addiction, stress, and pain (Garland, 2013). Results from this RCT of chronic pain patients receiving extended opioid pharmacotherapy (N = 115) indicated that compared to a support group (SG) control condition, MORE significantly decreased pain severity and functional impairment, as well as opioid craving and opioid misuse (Garland, Manusov, et al., 2014). Mechanistic analyses from this trial revealed that MORE led to enhanced cardiac-autonomic responses to natural reward cues (Garland, Froeliger, & Howard, 2014), suggesting that MORE might increase physiological responsivity to natural reward.

To test this hypothesis further, we used electroencephalography (EEG) to examine the effects of MORE on neurophysiological responses during natural reward processing. We focused on the late positive potential (LPP) of the EEG. The LPP is an ERP consisting of a positive deflection of the EEG waveform (a large “spike” in brain activity) that tends to reach maximum amplitude at 400–1000 ms at central-posterior sites on the scalp (e.g., parietal site Pz) after onset of an emotional stimulus. Automatically paying attention to emotional information elicits the LPP (Olofsson, Nordin, Sequeira, & Polich, 2008), as does consciously trying to focus on an emotional stimulus (Hajcak, Dunning, & Foti, 2009a), making the LPP a plausible candidate biomarker of mindful savoring. Results from our ERP analysis found that compared to the SG, MORE was associated with significantly greater increases in LPP during viewing of natural reward cues (relative to neutral cues) that were associated with heightened positive emotional responses and decreased opioid craving (Garland, Froeliger, & Howard, 2015b).

Although this RCT provided evidence that MORE may enhance electrocortical indices of natural reward processing in a sample of chronic pain patients at risk for prescription opioid misuse, we did not explore effects of increasing brain reward responses on chronic pain symptoms. Hence, the present secondary data analysis used ERP data from this trial to determine if increases in the LPP to natural reward cues were associated with decreases in pain severity and pain-related functional interference. Further, we examined whether changes in LPP reward response were associated with improvements in pain coping strategies. We hypothesized that increases in the parietal LPP to natural reward cues would be correlated with decreases in pain severity, pain interference, and pain catastrophizing, as well as correlated with increases in the ability to cope by diverting attention away from pain.

Method

Participants

This secondary data analysis used ERP data from a subset of participants (12 men and 17 women, mean age = 47.1, SD = 15.2) with complete pre-to-post treatment EEG data (MORE, n = 11; SG, n = 18) originally reported in Garland et al. (2015b). Participants were recruited between 2011 and 2012 from primary care clinics, pain clinics, and neurology clinics in the southeastern U.S. with flyers and online classified ads. Recruitment materials invited opioid-treated chronic pain patients to participate in a study focused on improving ways to address pain symptoms and problems with prescription pain medication. Participants met study inclusion criteria if they reported chronic pain (participants had been in chronic pain for an average of 12.7 years) and had taken opioid analgesics daily or nearly every day for at least the past 90 days (Chou et al., 2009). At preassessment and postassessment, participants completed self-report measures of pain severity and pain coping, and then they participated in a lab protocol that involved passively viewing visual images of naturally rewarding stimuli and neutral stimuli while EEG was recorded. The protocol was approved by the Florida State University institutional review board, and all procedures complied with standards set forth in the Helsinki Declaration of 1975. Participants were assessed for comorbid psychiatric disorders with the Mini-International Neuropsychiatric Interview 6.0 (MINI; Sheehan et al., 1998); participants were excluded if they were suicidal or psychotic. The majority of participants (86.2%) met criteria for prescription opioid misuse as defined by a validated cut-point on the Current Opioid Misuse Measure (Butler et al., 2007). More than two thirds of participants met MINI diagnostic criteria for major depressive disorder (70.4%), and one third met diagnostic criteria for generalized anxiety disorder (33.3%). For more details about sociodemographic and clinical characteristics of the sample for the parent RCT from which these data are derived, see Garland, Manusov, et al. (2014). Participants were paid $200 for the study. After completing the pretreatment assessment, participants were randomized to an 8-week MORE group or SG. Randomization order was generated by computer using simple randomization in varying block sizes (6–8) to preserve random allocation unpredictability, and the allocation list was stored in a protected file inaccessible to project staff involved in conducting study assessments.

MORE Group

The manualized MORE intervention (Garland, 2013) involved eight weekly group sessions (for a description of all session topics, see Garland, Manusov, et al., 2014) that integrated mindfulness training, third-wave cognitive–behavioral therapy, and principles from positive psychology to address chronic pain and prescription opioid misuse. Group sessions were 2 hours long and were led by a master’s degree-level social worker with more than 10 years of experience delivering mindfulness-based interventions in clinical settings. This clinician was supervised by the developer of MORE (the first author, an experienced, licensed psychotherapist). The first author monitored treatment fidelity by reviewing video and audio recordings of the sessions to monitor therapist adherence and competence. MORE participants were asked to engage in daily 15-minute mindfulness practice sessions at home guided by audio recordings. Participants were instructed in the use of mindfulness to intentionally focus and maintain attention on the pleasant sensory aspects (i.e., visual, auditory, olfactory, gustatory, or tactile) of a positive experience or beautiful object (e.g., a sunset or a beautiful tree) while noticing, appreciating, and absorbing any positive emotions arising in response to the pleasant event. In Session 4, participants were taught to practice mindful savoring by mindfully focusing their attention on the pleasant colors, textures, and scents of a bouquet of fresh flowers and then then metacognitively reflecting on emotions of contentment and joy arising from the experience. Subsequent sessions discussed use of comparable savoring techniques with an array of sensory targets, and mindful savoring was prescribed as homework practice.

Support Group

In the active control condition of this study, a master’s degree-level clinical social worker facilitated an 8-week SG consisting of 2-hour sessions in which participants expressed emotions and discussed topics related to chronic pain and opioid use/misuse. This SG format was based on the evidence-based Matrix Model intensive outpatient treatment manual (Center for Substance Abuse Treatment, 2006). The first author monitored treatment fidelity by reviewing video and audio recordings of the sessions to monitor therapist adherence and competence. SG participants were asked to engage in 15 minutes of journaling on chronic pain-related themes at home each day.

Measures

At pretreatment and posttreatment, participants completed an affective picture-viewing task while their EEG was recorded, and they completed self-report measures of pain severity, pain interference, and pain coping. Pain severity and interference were also reported at a 3-month posttreatment follow-up; pain coping was not measured at follow-up. Data were collected by research assistants who were blinded to treatment assignment, and treatment assignment remained concealed throughout the study.

Pain severity

Pain severity was measured with the four-item pain severity sub-scale from the Brief Pain Inventory (BPI; α = .87), a well-validated, widely used measure of acute and chronic pain (Cleeland, 1991). Participants reported their worst pain during the past week, least pain during the past week, average pain, and current pain. Response options ranged from 0 (no pain) to 10 (pain as bad as I can imagine). An overall pain severity score was computed by taking the mean of the four items.

Pain interference

Pain-related functional interference was assessed with the BPI pain interference subscale (α =.88). On a scale of 0 (does not interfere) to 10 (completely interferes), participants rated the extent to which pain had interfered with each of seven domains of normal functioning in the past week: general activity, mood, walking ability, normal work, relations with other people, sleep, and enjoyment of life. An overall pain interference score was computed by taking the mean of the seven items.

Pain coping

Maladaptive pain coping—via pain catastrophizing and diverting attention from pain—were measured with subscales on the Coping Strategies Questionnaire (Rosenstiel & Keefe, 1983). The pain catastrophizing subscale has good internal consistency (α = 0.86) and is comprised of 6 items, including “I worry all the time about whether it will end” and “I feel like I can’t go on.” The diverting attention subscale has good internal consistency (α = 0.83) and is comprised of 6 items, such as “I think of things I enjoy doing” and “I try to think of something pleasant.” Participants were asked to report how much they generally engaged in this form of coping when they felt pain. Responses are rated on a scale ranging from 0 (never) to 6 (always), and a total score is computed.

Affective picture-viewing task

First, participants were seated approximately 0.5 m directly in front of a 17-in. computer monitor, and EEG electrodes were attached. Next, participants performed an affective picture-viewing task (Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000) programmed in Eprime 2.0 software, in which full-screen images from the International Affective Picture System (IAPS; Lang, Bradley, & Cuthbert, 1999) were displayed in random order for 6 s each following the offset of a fixation cross presented for 500 ms. Participants viewed 18 pleasant and 18 neutral IAPS images and were instructed to simply view the images as they were presented and allow themselves to respond naturally. The natural reward image set included photos of endearing children and animals, athletic triumphs, appealing foods, beautiful landscapes, people with happy facial expressions, and intimate/erotic couples. The neutral image set included photos of household items, furniture, and people with neutral facial expressions. Natural reward and neutral images differed on IAPS normative valence ratings (M =7.09 and 5.05, respectively) and arousal ratings (M = 5.03 and 2.89, respectively), with pleasant images rated as more positive and arousing than neutral images. Responses were recorded for the entire 6-s image presentation to measure effects on slower autonomic responses (not reported in the present manuscript); however, to assess effects of MORE on the LPP, EEG data from only the first second were analyzed. The intertrial interval was randomly jittered from 2–4 s.

EEG data collection and data processing

Continuous EEG activity was recorded during the affective picture-viewing task using a BIOPAC MP150 system. For the present ERP analysis, only data from parietal site Pz were analyzed. EEG reference electrodes were placed on the earlobes. Vertical and horizontal electrooculograms recorded eye blinks and eye movements with miniature electrodes placed approximately 1 cm above and below the participant’s right eye. All electrophysiological signals were digitized and analyzed on a computer with BIOPAC’s AcqKnowledge 4.3 software. The EEG was sampled at 1000 Hz, and signals were filtered online with a 35 Hz low-pass filter. Offline, the EEG for each trial was rereferenced to the mean of the ears, band-pass filtered between 0.01 and 30 Hz, and artifact-corrected for vertical and horizontal electrooculogram movements using Gratton, Coles, and Donchin’s (1983) classical approach. Trials were rejected for subsequent analyses if they were contaminated by excessive physiological artifacts. A semiautomated procedure identified and rejected physiological artifacts according to the following validated criteria: a voltage step > 50.0 μV between sample points, a voltage difference of > 300.0 μV within a trial, and a maximum voltage difference of < 0.50 μV within any 100-ms interval (Hajcak et al., 2009a; Hajcak, Dunning, & Foti, 2009b). EEG signals were smoothed digitally offline for visualization purposes, but analyses were conducted on the presmoothed data.

ERP analysis

For each participant, an average ERP waveform was generated separately for natural reward and neutral trials. To depict overall effects and identify the time window for establishing the largest LPP to be used in analyses of MORE effects on reward processing, grand average waveforms were generated by averaging individual participant waveforms across the MORE group (number of reward trials retained for analysis: M = 12.6 ± 2.7; number of neutral trials: M = 12.8 ± 2.9) and SG (number of reward trials retained for analysis: M = 12.4 ± 2.7; number of neutral trials: M = 13.1 ± 3.2). Statistical analyses for the parietal LPP component were conducted on averages from each participant. The LPP was quantified with the following procedure:

• First, a baseline equal to the average activity in the 150-ms window prior to image onset was subtracted from data points subsequent to image onset. Based on visual inspection and following conventions from previous research (Cuthbert et al., 2000), we defined the parietal LPP as the average voltage occurring across the entire period from 400–1000 ms after image onset.

• To isolate LPP maxima during this time period, we examined our ERP grand averages and followed conventions from previous research to define two time windows—400–700 ms and 700–1000 ms—for the maximum parietal LPP (Cuthbert et al., 2000; Moser, Hartwig, Moran, Jendrusina, & Kross, 2014).

Statistical analyses

We used repeated measures ANOVA (RM-ANOVA) to test effects of MORE versus SG on pre-to-post treatment changes in the LPP response to natural reward images (vs. the LPP response to neutral images) in each of these windows. Because age and gender are known to significantly influence the LPP to affective picture viewing (Olofsson et al., 2008), we controlled for these variables by entering them simultaneously as covariates in the RM-ANOVA model. RM-ANOVA included Treatment Group (MORE vs. SG) × Image Type (Reward vs. Neutral) × Assessment (Pretreatment vs. Posttreatment) × Time Window interaction (400–700 ms, 700–1000 ms) as factors, and age and gender as covariates.

Next, we used linear regression analyses controlling for treatment group assignment to examine individual differences in clinical correlates of natural reward-specific changes in LPP responses over time, including changes in pain severity, pain interference, and pain coping via diverting attention and catastrophizing. Because the interventions under study (MORE and SG) contain multiple therapeutic components, we controlled for treatment-group assignment to elucidate the specific therapeutic effect of changes in natural reward responsiveness on pain-related outcomes. Each linear regression model included change in natural reward LPP response from pretreatment to posttreatment as the independent variable, change in clinical correlate as the outcome, and treatment group assignment as a covariate. Finally, we computed Pearson correlations between changes in LPP response to natural reward cues and changes in each pain-related and pain-coping outcome separately for each treatment group.

Results

As reported in Garland et al. (2015b), one-way ANOVA revealed that at pretreatment, natural reward cues did not elicit significantly greater LPP activation than neutral cues, nor were there significant between-groups differences in pretreatment LPP activation to natural reward or neutral cues.

Primary ERP analyses reported in Garland et al. (2015b) found that compared to the SG, participation in MORE was associated with significantly greater increases in LPP activation across the entire 400–1000 ms period at parietal site Pz during viewing of natural reward images (relative to neutral images). In the current secondary analysis, we sought to isolate LPP maxima by examining the effect of MORE relative to the SG on two time windows for the parietal LPP (Cuthbert et al., 2000; Moser et al., 2014): 400–700 ms, and 700–1000 ms. To identify the time period where the LPP difference was greatest, we then analyzed treatment effects on the LPP across both time windows with repeated measures ANOVA. Emerging from this analysis was a significant Treatment Group (MORE vs. SG) × Image Type (Reward vs. Neutral) × Assessment (Pretreatment vs. Posttreatment) × Time Window interaction (400–700 ms, 700–1000 ms): F(1, 26) = 4.47, p = .04, η_partial² = .15. Inspection of the grand average waveforms (Figure 1) clarified this interaction by revealing that the maximal between-groups difference in change from pretreatment to posttreatment in LPP responsiveness to natural reward (vs. neutral images) was evident in the “later” LPP window of 700–1000 ms. As such, we extracted the individual mean LPP natural reward change score (post–pre) for each subject from this later window to use in individual difference analyses.

Figure 1.

Effects of Mindfulness-Oriented Recovery Enhancement (MORE) versus a Support Group (SG) control condition on the late positive potential (LPP) of the EEG in response to natural reward-related images presented during an affective picture-viewing task, depicted in two time windows (400–700 ms and 700–1000 ms) following image onset. For a depiction of the entire event-related potential (ERP) time series from −150 ms to 1000 ms, see Garland et al. (2015b). LPP voltage is depicted in microvolts (μV), and time is represented in milliseconds (ms).

Associations Between Change in Natural Reward Processing and Pain Severity

We found that increases in LPP activation during natural reward processing predicted decreases in pain severity from pretreatment to posttreatment (B = −40.8, SE = 13.9, β = −.48, p = .007). However, increases in LPP activation did not significantly predict decreases in pain severity from pretreatment to 3-month follow-up (B = −20.6, SE = 17.5, β = −.24, p = .25).

Associations Between Change in Natural Reward Processing and Pain Interference

We found that increases in LPP activation predicted decreases in pain-related functional interference from pretreatment to posttreatment (B = −51.7, SE = 19.3, β = −.46, p = .01). We also found that increases in LPP activation predicted decreases in pain-related functional interference from pretreatment to 3-month follow-up (B = −47.9, SE = 22.8, β = −.37, p = .048).

Associations Between Change in Natural Reward Processing and Pain Coping

With regard to pain coping, we found that increases in LPP activation during natural reward processing predicted decreases in pain catastrophizing from pretreatment to posttreatment (B = −136.1, SE = 66.1, β = −.36, p = .05). We also found that increases in LPP activation predicted increases in pain coping by diverting attention away from pain from pretreatment to posttreatment (B = 131.6, SE = 61.9, β = .38, p = .043).

Within-Group Correlations

Bivariate correlations were computed within each treatment group separately. Within the MORE group, the following associations between pre-to-post changes in clinical correlates and pre-to-post changes in LPP activation were observed: pain severity (r = −.57), pain interference (r = −.54), catastrophizing (r = −.14), and diverting attention (r =.27). Within the SG, the following correlations with change in LPP activation were observed: pain severity (r = −.49), pain interference (r = −.39), catastrophizing (r = −.25), and diverting attention (r = .50). Scatterplots for main outcomes, including pain severity and interference, are presented in Figures 2 and 3. Within-group Pearson correlations were substantially underpowered; a post hoc power analysis revealed that on the basis of the mean and within-groups correlation coefficient observed in the present study (r = .44), a within-group n of approximately 38 would be needed to obtain statistical power at the recommended .80 level (Cohen, 1988).

Figure 2.

Scatterplot depicting within-group associations between pre–post treatment increases in natural reward late positive potential (LPP) response (in microvolts) and pre–post treatment decreases in pain severity (on the Brief Pain Inventory). Positive values indicate increases in LPP and pain severity.

Figure 3.

Scatterplot depicting within-group associations between pre–post treatment increases in natural reward late positive potential (LPP) response (in microvolts) and pre–post treatment decreases in pain interference (on the Brief Pain Inventory). Positive values indicate increases in LPP and pain interference.

Discussion

Findings from this secondary data analysis revealed that treatment-related increases in electrocortical responses during natural reward processing were linked with decreases in chronic pain symptoms and improvements in pain coping. Specifically, MORE enhanced the parietal LPP from 700–1000 ms following natural reward cue onset to a significantly greater extent than an active control condition, and this increase in brain reward response significantly predicted decreased pain severity from pretreatment to posttreatment and decreased pain interference by 3-month follow-up. Moreover, increases in brain reward response were associated with decreased pain catastrophizing and increased capacity to cope with pain by shifting attention away from pain. Considered together, study findings suggest that focusing attention on naturally rewarding objects and events may improve pain coping and decrease impairment associated with chronic pain.

Basic science suggests that chronic pain and opioid misuse can blunt brain reward systems in response to naturally rewarding stimuli, resulting in a hedonic deficit that drives opioid dose escalation as a means of maintaining hedonic equilibrium. Conversely, experimental studies indicate that inducing positive affect via laboratory inductions can produce significant analgesic effects (Finan & Garland, 2015). Few studies of psychosocial interventions (if any) have tested whether increasing positive emotional experience can produce pain relief and whether such positive-affect analgesia is associated with increased physiological sensitivity to natural rewards. Findings from the present study suggest that a social work intervention focused on increasing positive emotions through mindful savoring can increase brain reward responses associated with pain relief.

To be clear, the association between increased LPP reward response and pain relief was not unique to the MORE condition. Individual difference analyses controlling for treatment condition indicated that, irrespective of the type of treatment received, chronic pain patients who evince the largest improvements in brain reward responses to cues representing natural rewards experience the largest reductions in pain severity and interference. This finding suggests that electrocortical responsiveness to natural reward may be a transtherapeutic mechanism, which, when stimulated by different strategies (e.g., mindfulness training or social support), may produce clinical benefits. However, in the original study from which these data are derived, MORE increased the LPP reward response (Garland et al., 2015b) and decreased pain severity and interference (Garland, Manusov, et al., 2014) to a significantly greater extent than a SG, indicating that specific training in mindful savoring may be particularly useful in enhancing natural reward responses and producing positive-affect analgesia.

In prior studies, LPP activation was strongly correlated with activity in medial prefrontal cortices as measured by functional magnetic resonance imaging (fMRI; e.g., Liu, Huang, McGinnis-Deweese, Keil, & Ding, 2012). Plausibly, increased LPP in response to natural rewards might be mediated by the effects of MORE on boosting medial prefrontal activation. Converging evidence for this claim was found in a recent fMRI study of MORE as a treatment for nicotine-addicted smokers (Froeliger et al., 2017). In this study, compared to a time-matched comparison group, MORE participants showed significant pretreatment-to-posttreatment increases in brain activity in the rostral anterior cingulate cortex (a medial prefrontal brain region) when savoring natural reward cues, and rostral anterior cingulate cortex activation was significantly associated with decreases in cigarette smoking. Additional evidence for the effect of MORE on physiological indices of reward processing was generated by a recent pilot RCT of behavioral interventions for obese cancer survivors, which found that MORE was associated with significant increases in zygomatic electromyography responses to natural reward cues that mediated MORE’s effect on reducing attentional fixation on appetizing, high-calorie foods (Thomas et al., 2018). Together, results from the current study and earlier research demonstrating that MORE-induced increases in autonomic and electrocortical responses to natural rewards are associated with decreased opioid craving suggest that enhancing natural reward responsiveness through mindfulness may be a means of restoring adaptive hedonic regulation of pain, pleasure, and addictive behavior (Garland, 2016).

Limitations

The primary limitation of the present study is that we do not know to what extent MORE increased reward processing by reducing opioid use, or vice versa. Because MORE decreased opioid misuse (Garland, Manusov, et al., 2014), it is plausible that the observed modulations of LPP function were due to decreases in opioid intake over the study. However, we were unable to test this hypothesis due to our inability to accurately quantify opioid dosing. Although we asked open-ended questions about opioid dosing, missing data and extreme variability in the quality of responses made it impossible to quantify this variable for use as a covariate in this study. To overcome this limitation, future studies should assess opioid dosing via a multi-pronged approach that includes pharmacy records, prescription drug monitoring databases, pill count, and self-report. Further, the study was limited by the small sample size, which reduced statistical power and—to prevent Type II error—precluded us from assessing the significance of within-group Pearson correlations or conducting more sophisticated, multivariate analyses to test changes in the LPP as a mediator of treatment effects on study outcomes. Finally, although we identified an association between increasing LPP response and pain relief in the present study, traditional mindfulness-based interventions tend to emphasize emotional nonreactivity over cultivating and savoring natural rewards. In that regard, mindfulness-based interventions other than MORE (e.g., Mindfulness-Based Stress Reduction; Kabat-Zinn, 1982) have been shown to reduce pain through a variety of therapeutic mechanisms independent from reward, including shifting from affective to sensory processing of pain sensations (Garland et al., 2012), pain acceptance (McCracken & Vowles, 2014), and increasing cortical representations of interoceptive awareness (Zeidan et al., 2011). Increasing reward response is not even the sole analgesic mechanism of MORE; indeed, increases in the capacity to reinterpret pain as innocuous sensory information mediated the pain-relieving effects of the intervention (Garland et al., 2014). Together, these findings suggest there may be multiple mediators of mindfulness-based analgesia.

Conclusion

Results from this secondary analysis indicate that MORE may enhance neurophysiological indices of natural reward responsiveness among opioid-treated chronic pain patients. Results also suggest that increasing brain responses to natural rewards is associated with improvements in clinical pain and functional impairment. Given this latter finding, interventions that increase exposure to positive and rewarding experiences and activities may be especially useful for enhancing resilience to chronic pain (Hassett & Finan, 2016). Beyond its clinical implications, the present investigation is one of a small-but-growing body of studies that use social work neuroscience methods for intervention research (e.g., Eack, Newhill, & Keshavan, 2016; Garland, Gaylord, Boettiger, & Howard, 2010; Garland, Froeliger, et al., 2014; Maynard, Boutwell, Vaughn, Naeger, & Dell, 2015). This research methodology may be especially important for mechanistic research; to assess the mechanisms of action for social interventions, social work researchers should consider using psychophysiological measures (e.g., fMRI, EEG, heart rate variability) coupled with tasks designed to probe cognitive and affective function. Although the efficacy of psychosocial treatments is often dismissed by reducing these approaches to nonspecific factors like therapeutic relationship, psychophysiological findings reveal that social work interventions can actually alter the functioning of the brain and body in ways consistent with theorized mechanisms of action to produce clinically significant change. In this respect, social work neuroscience is crucial for providing empirical support for the biopsychosocial framework that underlies social work practice (Garland & Howard, 2009).