Cannabis Dampens the Effects of Music in Brain Regions Sensitive to Reward and Emotion

Cannabis Dampens the Effects of Music in Brain Regions Sensitive to Reward and Emotion Background: Despite the current shift towards permissive cannabis policies, few studies have investigated the pleasurable effects users seek. Here, we investigate the effects of cannabis on listening to music, a rewarding activity that frequently occurs in the context of recreational cannabis use. We additionally tested how these effects are influenced by cannabidiol, which may offset cannabis-related harms. Methods: Across 3 sessions, 16 cannabis users inhaled cannabis with cannabidiol, cannabis without cannabidiol, and placebo. We compared their response to music relative to control excerpts of scrambled sound during functional Magnetic Resonance Imaging within regions identified in a meta-analysis of music-evoked reward and emotion. All results were False Discovery Rate corrected (P < .05). Results: Compared with placebo, cannabis without cannabidiol dampened response to music in bilateral auditory cortex (right: P = .005, left: P = .008), right hippocampus/parahippocampal gyrus (P = .025), right amygdala (P = .025), and right ventral striatum (P = .033). Across all sessions, the effects of music in this ventral striatal region correlated with pleasure ratings (P = .002) and increased functional connectivity with auditory cortex (right: P < .001, left: P < .001), supporting its involvement in music reward. Functional connectivity between right ventral striatum and auditory cortex was increased by cannabidiol (right: P = .003, left: P = .030), and cannabis with cannabidiol did not differ from placebo on any functional Magnetic Resonance Imaging measures. Both types of cannabis increased ratings of wanting to listen to music (P < .002) and enhanced sound perception (P < .001). Conclusions: Cannabis dampens the effects of music in brain regions sensitive to reward and emotion. These effects were offset by a key cannabis constituent, cannabidol. Received: May 4, 2017; Revised: August 16, 2017; Accepted: August 30, 2017 © The Author(s) 2017. Published by Oxford University Press on behalf of CINP. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons. org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is 21 properly cited. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 22 | International Journal of Neuropsychopharmacology, 2018 Significance Statement Here, we report that cannabis administration decreased response to music in several brain regions linked to reward and emotion. These included right ventral striatum, which showed increased functional connectivity with auditory cortex and correlated with pleasure ratings during musical listening, consistent with its critical role in reward processing. These effects were offset when cannabis contained cannabidiol, a key cannabinoid that has been found to reduce some harmful effects of cannabis. Keywords: cannabis, music, reward, pleasure, emotion Introduction The main psychoactive constituent of cannabis, THC (delta- Additionally, THC causes modest, regionally selective 9-tetrahydrocannabinol), produces subjective effects such as increased dopamine release in limbic striatum (Bossong et al., feeling “stoned” and can impair memory and elicit transient 2015). Such effects might enhance the rewarding experience of psychotic-like symptoms (Curran et  al., 2016). Certain types of music, which can also elicit dopamine release in ventral stri- cannabis also contain cannabidiol (CBD), which can have oppos- atum (Salimpoor et al., 2011) as well as enhancing activation ite effects of THC on a range of functional neuroimaging tasks and connectivity between mesolimbic brain regions (Blood (Bhattacharyya et  al., 2010Batalla et  ; al., 2014). Moreover, CBD and Zatorre 2001; Menon and Levitin 2005; Koelsch et al., 2006; has been found to offset harmful effects of THC (e.g., memory Salimpoor et al., 2011T ; rost et al., 2012). Functional connectivity impairment and psychotic-like symptoms) without influencing between ventral striatum and auditory cortex during listening subjective intoxication (Curran et al., 2016 Englund et  ; al., 2017). also predicts the rewarding experience of music (Salimpoor et Cannabis containing high THC and little if any CBD is becoming al., 2013; Zatorre and Salimpoor 2013; Martínez-Molina et al., increasingly prevalent (Hardwick and King 2008 ElSohl ; y et  al., 2016). 2016) and has been linked to greater mental health and addic- Here, we conducted the first controlled experimental study tion problems compared with less potent varieties of cannabis on the interactive effects of cannabis and music. Based on pre- (Di Forti et al., 2015; Freeman and Winstock 2015). vious findings that cannabis and music activate and increase Despite the changes currently occurring in cannabis legisla- connectivity between common regions in the reward net- tion worldwide, including legalization of use for medicine and work, whereas a CB1R antagonist dampened neural response pleasure (Room 2014), few studies have attempted to document to reward (Horder et  al., 2010), and observational data linking the effects that recreational users seek (Curran et al., 2016). The cannabis use and music, we hypothesized that cannabis would limited evidence of positive effects tends to have arisen inci- increase haemodynamic response to music in brain regions dentally in studies investigating cannabis-related harms. For sensitive to reward and emotion (Koelsch 2014) as well as sub- example, THC has been reported to increase phonological flu- jective ratings (wanting to listen to music, pleasure of listen- ency (Curran et al., 2002), a measure of divergent thinking, espe- ing). Given that CBD and THC can have opposing neural effects cially among people with low trait creativity (Schafer et al., 2012).(Bhattacharyya et  al., 2010Batalla et  ; al., 2014) and CBD can Cannabis has a strong historical link to music and is associated attenuate THC harms (Curran et al., 2016 Englund et  ; al., 2017), with several distinct styles, including jazz, reggae, and rock (Booth we predicted that these effects would be partially offset by CBD. 2004). Cannabis is reported to enhance appreciation of music (Tart 1970; Green et  al., 2003), and its use is consistently high among Methods people who attend music festivals and nightclubs (Lim et al., 2008; Van Havere et al., 2011; Palamar et al., 2015). This association may Design and Participants be partly attributable to shared effects on reward circuitry between drug and nondrug rewards (Berridge and Kringelbach 2015). Music A randomized, double-blind, crossover design compared can- recruits key regions in the reward network, including ventral stri- nabis with CBD (Cann+CBD), cannabis without CBD (Cann-CBD) atum, mediodorsal thalamus, anterior insula, orbitofrontal cortex,and matc hed placebo in 16 cannabis users. Experimental ses- amygdala, and hippocampus (Koelsch 2014). sions were separated by at least 1 week (>3 times the elimin- Many of these reward-related brain regions are character - ation half-life of THC) to minimize carryover effects (D’Souza ized by a high density of Cannabinoid Type-1 Receptors (CB1Rs) et al., 2004; Hindocha et al., 2015). In addition to the music task (Curran et al., 2016). THC is a partial agonist of CB1Rs and may described here, participants completed additional assessments influence response to music by interfering with endogenous that are reported elsewhere (Lawn et  al., 2016). Inclusion cri- CB1R ligands such as anandamide (Thieme et al., 2014), which teria were fluency in English, right-handedness, age between plays a causal role in consummatory response to reward (Mahler 18 and 70  years, and self-reported current cannabis use (≥4 et  al., 2007). A  human neuroimaging study found that THC (a times in the last year, ≤3 times/wk, ability to smoke a whole partial CB1R agonist) dampened the effects of monetary reward joint to oneself). We did not collect data on participants’ typ- feedback across a widespread network, including temporal and ical method of administering cannabis. However, previous orbitofrontal cortices, while leaving reward anticipation intact data from the UK suggest that the majority (~76%) of cannabis (van Hell et  al., 2012). By contrast, 7-day administration of a users typically smoke cannabis together with tobacco in joints, CB1R antagonist was found to diminish response to food reward and only a small minority (~4%) use a vaporizer as their most in ventral striatum and orbitofrontal cortex (Horder et al., 2010).common route (Hindocha et  al., 2016). Exclusion criteria were Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 23 self-reported frequent and/or severe adverse reactions to can- them with a 30-millisecond linear cross-fade between excerpts nabis, current use of illicit drugs other than cannabis more than (Menon and Levitin 2005). Scrambled excerpts retain the same twice per month, current alcohol use >4 d/wk, significant phys- distribution of pitch and loudness and the same spectral infor - ical health problems, color blindness, current treatment for a mation as normal music. However, they lack temporal structure psychiatric disorder, current/history of psychosis, and current/ and are rated as less pleasurable than normal excerpts (Menon history of psychosis in an immediate family member. This study and Levitin 2005). was approved by the UCL ethics committee, and all participants To deliver clear audio during scanning, clips were adapted provided written informed consent. to improve volume constancy during sections of low volume. Output volume was adapted for each participant in the scanner before the task commenced. Normal/scrambled excerpts were Procedure delivered using PsychoPy (Version 1.79.01) through MR com- Following telephone screening, eligible participants completed patible sensimetric earphones (http://www.sens.com/products/ a baseline session consisting of task training (outside of the MRI model-s14/) in a standard blocked fMRI design. The 12 normal/ scanner), video training for drug administration, drug history scrambled excerpts were presented in a pseudo-randomized (Freeman et al., 2012), and problematic cannabis use on Severity order across the 3 test sessions. Each 21-second excerpt was fol- of Dependence Scale (Gossop et al., 1995). The Beck Depression lowed by a 1-second interstimulus interval. Next, participants Inventory-II (Beck et  al., 1996) and Temporal Experiences of rated the pleasantness of each excerpt using a 2-finger response Pleasure (Gard et al., 2006) were also administered. Each experi- pad beneath their right hand (fixed time of 8 seconds). The mental session began with a urinary drug screen to verify recent numerical rating scale was anchored from 0 (not at all pleasant) use reported by Timeline Follow-back (Sobell and Sobell 1992). to 10 (very pleasant). This was followed by 12 seconds of pas- Next, 11-point (0–10) Numerical Rating Scales were adminis- sive fixation (rest). The total task time was 8 minutes 24 seconds, tered ~0 minutes before drug inhalation (Pre-Drug), ~5 minutes plus a 5-second end-buffer period. after first drug administration (Post-Drug), and ~90 minutes after first drug administration (Post-Scan). The Numerical fMRI Data Acquisition Rating Scales “Want to Listen to Music” and “Enhanced Sound Perception” were administered at all 3 of these time points; “Feel Imaging data were collected using a Siemens TIM Avanto 1.5T Drug Effect,” “Like Drug Effect,” and “Want More Drug” were scanner, using a 32-channel receive-only head coil, at the administered only after drug administration (Post-Drug and Birkbeck-UCL Centre for Neuroimaging, London. An automated Post-Scan). Heart rate and systolic and diastolic blood pressure shim procedure was applied to minimize possible magnetic were also recorded at the same 3 time points (Pre-Drug, Post- field homogeneities. Functional imaging used a multiband Drug, Post-Scan). (acceleration factor = 4) gr adient-echo T2*-weighted echo-planar imaging (EPI) sequence with 40 slices per volume (TR = 1000 ms; Drug Administration TE = 55 ms; in-plane matrix = 64 x 64; 3 mm isotropic voxels; flip angle= 75°; bandwidth= 1474 Hz/pixel; 509 volumes). The first 8 Cannabis was obtained from Bedrocan, The Netherlands and scans were treated as “dummy” scans and discarded to avoid used within 6  months of purchase. It was stored on site in T1-equilibrium effects. All scanning parameters were selected foil-sealed pouches at -20ºC and then at ambient temperature to optimize the quality of the BOLD signal while maintaining a prior to drug administration. Each dose was vaporized using a sufficient number of slices to acquire whole-brain data. To co- Volcano Medic Vaporizer (Storz and Bickel) at 210ºC in 2 sequen- register the fMRI data into standard space, we also acquired a tially administered balloons to minimize residual cannabinoids MPRAGE structural sequence (TR = 2730  ms; TE = 3.57  ms; mat- (Lawn et al., 2016). Participants inhaled at their own pace (each rix = 176 x 256 x 256; 1-mm isotropic voxels; flip angle = 7°; inhalation held for 8 seconds, enforced by the experimenter bandwidth= 190 Hz/pixel; parallel imaging acceleration fac- using a stopwatch) until the balloon was empty, which lasted ~5 tor= 2), and a B0 field map image (64 axial slices; TR = 1170 ms; minutes for both balloons. All participants complied with this TE1 = 10.0  ms; TE2 = 14.76  ms; in-plane matrix= 64 x 64; 3 x 3 x administration protocol. Bedrobinol (12% THC, <1% CBD), Bediol 2 mm voxels; flip angle = 90°; bandwidth= 260 Hz/pixel) to enable (6% THC, 7.5% CBD), and placebo cannabis were used to load distortion correction of the functional data. doses of 8 mg THC + 10 mg CBD (Cann+CBD), 8 mg THC (Cann- CBD), and placebo (Lawn et  al., 2016). Placebo cannabis had a comparable terpene profile to the 2 active forms of cannabis, fMRI Data Analysis ensuring it was matched for smell. The same physical quantity Preprocessing and data analysis were performed using of cannabis/placebo (133.4mg) was administered across each Statistical Parametric Mapping (SPM8; http://www.fil.ion.ucl. of the 3 sessions. This dose of THC has produced effects on ac.uk/spm/software/spm8/). Standard preprocessing procedures brain and behavior in studies with similar vaporizer protocols consisted of bias correction of EPI images to control for within- (Bossong et al., 2009; Hindocha et al., 2015; Mokrysz et al., 2016) volume signal intensity differences, realignment/unwarping and is roughly equivalent to one-quarter of a standard UK joint to correct for interscan movements, correction for differences (Freeman et al., 2014). in slice acquisition timing, and normalization of the images to an EPI template specific to our sequence and scanner that was Music Task (Menon and Levitin 2005) aligned to the T1 MNI template. Finally, the normalized func- Six 21-second excerpts of standard instrumental classical tional images were spatially smoothed with an isotropic 8-mm FWHM Gaussian kernel. music were taken from compact disc recordings, adapted from a previous study (Menon and Levitin 2005). Six scrambled ver - At the first level, the normal and scrambled epochs were each modelled as a 21-second boxcar convolved with the canon- sions were created by randomly drawing 250- to 350-millisec- ond variable-sized sections from each piece and concatenating ical hemodynamic response function combined with time and Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 24 | International Journal of Neuropsychopharmacology, 2018 dispersion derivatives to create the contrast music>scrambled, SDS, and drug history) were missing for one participant. Missing as previously used with this task (Menon and Levitin 2005). data from experimental sessions (0.69% of Numerical Rating The interstimulus interval, rating, and passive fixation (rest) Scale data, 0.69% of cardiovascular data) were imputed with the were also modelled. Each subject’s movement parameters were mean for that session and time point to retain each participant included as confounds. Low-frequency noise was removed with in the repeated-measures analysis. Repeated-measures ANOVA a high-pass filter (cut-off frequency 1/128 Hz). Parameter esti- models were used for all data collected on the 3 experimental mates pertaining to the height of the hemodynamic response sessions, including within-subject factors of drug (Cann+CBD, function for each regressor of interest were then calculated for Cann-CBD, placebo) and time (Pre-Drug, Post-Drug, Post-Scan) each voxel. or (Post-Drug, Post-Scan) and additional factors where appro- At the second level, the contrast music>scrambled was priate. Posthoc pairwise tests were Bonferroni-corrected locally entered into a within-subject ANOVA model with a single fac- within each ANOVA model. Additional repeated-measures tor of drug (Cann+CBD, Cann-CBD, placebo). Within this ANOVA ANOVA models were used to aid interpretation of interactions model, we used t contrasts to investigate music>scrambled and where appropriate. The Greenhouse-Geisser correction was the reverse contrast (scrambled>music) across all scans. Drug applied where assumptions of sphericity were violated, with effects on music>scrambled were conducted using t contrasts degrees of freedom rounded to the nearest integer. To reduce within this ANOVA model. To aid interpretation of signifi- type I and type II error rates, correlations with fMRI data were cant drug effects (which could reflect changes in response to collapsed across each of the sessions using mixed effects mod- music, scrambled, or both), separate parameter estimates were els, with a Bonferroni-adjusted α threshold. These accounted for extracted from these coordinates for the contrasts music>rest fixed effects of drug and session order, with a random intercept and scrambled>rest using the MarsBaR region of interest tool- of participant and maximum likelihood estimation. Equivalent box; these were analyzed using repeated-measures ANOVA mixed effects models were used to assess possible confound- in SPSS. ing by cardiovascular measures, cannabis use, and session order. Psychophysiological interaction (PPI) analysis was per - formed to assess task-related functional connectivity (O’Reilly Results et  al., 2012) using seed regions identified by drug effects. We extracted the representative time-course from voxels in the Participants seed region (6-mm radius sphere) using the first eigenvari- ate calculated from singular value composition. This time Seventeen participants completed the study. One participant course (physiological) was entered into a General Linear Model was excluded due to excessive head movement on one session together with the contrast music>scrambled (psychological) (exceeding thresholds for both translation [>6 mm] and rotation and their interaction (PPI). Motion parameters were included in [>6º]) and was replaced, leaving a final sample of 16. Demographic first-level models as nuisance regressors. The PPI regressor was and drug use data are shown in Table  1. The following num- analyzed using a within-subject ANOVA. We used contr t asts ber of participants completed each treatment order: Placebo, to investigate PPI effects across all scans and to compare drug Cann+CBD, Cann-CBD: n = 3; Placebo, Cann-CBD, Cann+CBD: effects. n = 2; Cann+CBD, Placebo, Cann-CBD: n= 3; Cann+CBD, Cann- A False Discovery Rate correction (< P .05) w as applied to all CBD, Placebo: n= 3; Cann-CBD, Placebo, Cann+CBD: n= 2; Cann- fMRI analyses. Regions of interest were defined from a previ- CBD, Cann+CBD, Placebo: n= 3. ous meta-analysis of music-evoked reward and emotion (see Figure  1 and supplementary Table  1 in Koelsch 2014) using the MarsBaR toolbox. Firstly, each of the structures identified in the Behavioral Results meta-analysis was converted into a single sphere. Coordinates Subjective Effects were converted to MNI using the Yale BioImage Suite (Lacadie Subjective effects are shown in Figure 2 . A main effect of drug et al., 2008). Sphere radius was estimated from the cluster size (F = 107.659, P < .001, η = 0.878) emerged for Feel Drug Effect, reported in the meta-analysis. Where clusters contained mul- 1,22 p reflecting increased scores following Cann+CBD (P < .001) and tiple structures, size was determined using the cluster mean. Cann-CBD (P < .001) compared with placebo, but no differences Subthreshold clusters were assigned a default size of 200 mm . between Cann+CBD and Cann-CBD (P = 1.000). There was also Each of these spheres was combined into a single mask that a main effect of time, indicating that scores decreased from was applied to second-level analysis. This mask (41 240  mm ) Post-Drug to Post-Scan (F = 19.057, P < .001, η = 0.560), but included bilateral hippocampal formation, bilateral amygdala, 1,15 p there was no evidence for an interaction between drug and bilateral auditory cortex, right ventral striatum, left caudate time (F = 0.796, P = .461, η = 0.050). Like Drug Effect showed nucleus, presupplementary motor area, frontomedian cortex, 2, 30 p a similar profile of results. There was a main effect of drug rostral cingulate zone, pre-genual and middle cingulate cortex, (F = 44.371, P < .001, η = 0.747), reflecting increased scores fol- medial and laterial orbitofrontal cortex, right anterior insula, 2,30 p lowing Cann+CBD (P < .001) and Cann-CBD (P < .001) compared mediodorsal thalamus, and superior parietal lobule. with placebo but no difference between Cann + CBD and Cann- CBD (P = 1.000). There was also a main effect of time, indicating Behavioral Data Analysis that scores decreased from Post-Drug to Post-Scan (F = 19.454, 1,15 SPSS version 21 was used to analyze all behavioral data and par - P < .001, η = 0.565). Again, there was no evidence for an inter - ameter estimates extracted posthoc using MarsBaR. Outliers action between drug and time (F = 0.589, P = .561, η = 0.038). 2,30 p (>3 times IQR) were winsorized within each session and time For Want More Drug, there was no evidence for any effects or point. Histograms were used to investigate normality, and interactions: drug by time (F = 2.462, P = .102, η = 0.141), drug 2,30 p square root or log transformations were applied where appro- (F = 1.329, P = .280, η = 0.081), or Time (F = 0.388, P = .543, 2,30 p 1,15 priate. Trait measures (BDI, Temporal Experiences of Pleasure, η = 0.025). Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 25 Figure  1. Subjective effects. Both types of cannabis increased ratings for (A) Feel Drug Effect and (B) Like Drug Effect but did not influence (C) Want More Drug. Cann + CBD, cannabis with cannabidiol (CBD); Cann-CBD canna , bis without CBD. ***P < .001. Figure 2. Cardiovascular effects. Both types of cannabis increased (A) heart rate and (B) systolic blood pressure. (C) Diastolic blood pressure increased from Pre- to Post- Drug following cannabis without cannabidiol (CBD), but not following cannabis with CBD; Cann + CBD , cannabis with CBD; Cann-CBD ,cannabis without CBD; *P < .05, ***P < .001; Difference between cannabis types. Figure 3. Subjective music ratings. (A) Both types of cannabis increased ratings of Want to Listen to Music. (B) Both types of cannabis increased scores for Enhanced Sound Perception and this increase was greater for cannabis with cannabidiol (CBD). (C) Neither type of cannabis influenced the pleasure of listening to music or scrambled sound clips. Cann + CBD, cannabis with CBD; Cann-CBD ,cannabis without CBD. ***P < .001; Difference between cannabis types. Cardiovascular Effects cannabis alone, there were no differences between the effects of Cardiovascular effects are shown in Figure  2. For heart rate Cann+CBD and Cann-CBD on heart rate across the 3 time points (BPM), a drug by time interaction emerged (F = 18.243, P < .001, (drug by time interaction: F = 0.123, P = .885, η = 0.008; main 2,28 2,30 p 2 2 η = 0.549) as well as main effects of both drug (F = 13.999, effect of drug: F = 0.090, P = .768, η = 0.006, main effect of time: p 2,30 1,15 p 2 2 2 P < .001, η = 0.483) and time (F = 45.977, P < .001, η = 0.754). F = 87.391, P < .001, η = 0.854). For systolic blood pressure, a p 2,30 p 2,30 p Heart rate increased from Pre-Drug to Post-Drug following main effect of drug was found (F = 6.297, P = .005, η = 0.296). 2,30 p Cann+CBD (P < .001) and Cann-CBD (P < .001) but not placebo This reflected increased blood pressure for both Cann+CBD (P = .456). It then decreased from Post-Drug to Post-Scan for (P = .030) and Cann-CBD (P = .006), compared with placebo, but both Cann+CBD (P < .001) and Cann-CBD (P < .001) but did not no differences between Cann+CBD and Cann-CBD (P = 1.000). change on placebo (P = 1.000). When comparing the 2 types of There was no evidence for an interaction between drug and Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 26 | International Journal of Neuropsychopharmacology, 2018 Table 1. Demographic and Drug Use Data and 0.00 (0.00) on Cann-CBD. Analysis of variance was there- fore restricted to 2 time points (Post-Drug, Post-Scan). A  main Mean/ effect of drug (F = 44.810, P < .001, η = 0.749) reflected increased 2,30 p frequency SD scores from placebo following Cann+CBD (P < .001) and Cann- CBD (P < .001) and higher scores following Cann+CBD compared Age 26.25 7.35 with Cann-CBD (P = .015). There was no evidence for an inter - Gender (male/female) 8/8 - action between drug and time (F = 2.056, P = .146, η = 0.121) or 2,30 p Days of cannabis use per month 8.06 5.48 a main effect of time (F = 1.248, P = .281, η = 0.077). Finally, we Years of cannabis use 8.94 7.02 1,15 p analyzed trial-by-trial pleasure ratings, recorded immediately Days since last cannabis use 19.25 45.28 after listening to classical music and scrambled sound excerpts Days to smoke 3.5 g cannabis 25.88 33.73 during MRI scanning. There was a main effect of excerpt Severity of dependence scale (cannabis) 1.13 1.26 (F = 133.860, P < .001, η = 0.899), indicating that music was rated Alcohol use (yes/no) 16/0 - 1,15 p Days of alcohol use per month 10.81 4.86 as more pleasant than scrambled sound. However, there was no Number of UK alcohol units (8 g) per 5.93 2.08 evidence for a main effect of drug (F = 1.205, P = .314, η = 0.074) 2,30 p session or a drug by excerpt interaction (F = 1.221, P = .309, η = 0.075). 2,30 p Current tobacco use (yes/no) 15/1 - Next, we calculated a pleasure rating score (music>scrambled) Days of tobacco use per month 11.30 10.27 equivalent to our fMRI contrast of interest to provide compar - Cigarettes per day 3.63 3.62 able metrics for brain and behavior. Mean (SD) pleasure rating Current MDMA use <twice a month (yes/ 6/10 - scores were 5.16 (2.27) for Cann+CBD, 4.78 (2.03) for Cann-CBD, no) and 5.53 (1.99) for placebo. Analysis of these scores provided no Current cocaine use <twice a month (yes/ 3/13 - evidence for an effect of drug (F = 1.221, P = .309, η = 0.075). 2,30 p no) Current ketamine use <twice a month 2/14 - fMRI Results (yes/no) Beck Depression Inventory-II 3.38 3.12 Main Effect of Task Temporal experiences of pleasure 42.06 4.85 All fMRI analyses were conducted among regions of inter - (anticipatory) est selected from a meta-analysis of previous studies (Koelsch Temporal experiences of pleasure 43.50 5.61 2014). Across all sessions, listening to music elicited activation (consummatory) in bilateral amygdala, bilateral striatum, left hippocampus, and left cingulate gyrus (see Table 2). For completion, we also exam- time (F = 0.953, P = .440, η = 0.060) or a main effect of time ined the reverse contrast (scrambled>music), which revealed 4,60 p activation in bilateral auditory cortex (see Table 2). (F = 2.641, P = .088, η = 0.150). For diastolic blood pressure, 2,30 p an interaction between drug and time was found (F = 3.217, 4,60 Drug Effects P = .019, η = 0.177) and a main effect of time (F = 7.702, P = .002, p 2,30 2 2 η = 0.339) but not drug (F = 2.975, P = .066, η = 0.165). Diastolic Response to music>scrambled was greater on placebo compared p 2,30 p with Cann-CBD in bilateral auditory cortex, right hippocam- blood pressure increased from Pre-Drug to Post-Drug for Cann- CBD (P < .001) but not Cann+CBD (P = .233) or placebo (P = 1.000). pus/parahippocampal gyrus, right ventral striatum, and right amygdala (see Table 2 and Figure 4). There was no evidence for It then increased from Post-Drug to Post-Scan following placebo (P = .030) but not Cann-CBD (P = 1.000) or Cann+CBD (P = 1.000). any differences when comparing Cann+CBD with placebo or Cann+CBD with Cann-CBD. Subjective Music Ratings To aid interpretation of these findings (which may have been driven by drug effects on music, scrambled sound, or both), we Subjective music ratings are shown in Figure  3. For Want to Listen to Music, we found a drug by time interaction (F = 5.256, extracted parameter estimates from each of the clusters iden- 4,60 tified in this drug effect (Table 2, bottom panel) for separate P = .001, η = 0.259) and main effects of drug (F = 5.664, P = .008, p 2,30 2 2 η = 0.274) and time (F = 6.300, P = .012, η = 0.296). Scores contrasts of music>rest and scrambled>rest. ANOVA revealed p 1,22 p an interaction between drug (placebo, Cann-CBD) and contrast increased from Pre-Drug to Post-Drug following Cann+CBD (P < .001) and Cann-CBD (P = .002) but not placebo (P = 1.000). (music>rest, scrambled>rest) (F = 37.851, P < .001, η = 0.716). 1,15 p This interaction indicated that relative to placebo, Cann-CBD Scores then decreased from Post-Drug to Post-Scan on Cann+CBD (P = .028), but these tests did not reach significance for Cann- decreased parameter estimates for music>rest (P = .009, mean difference -0.195, standard error 0.065). However, there was no CBD (P = .553) or placebo (P = .199). However, analysis of all three drug conditions suggested that the decrease in Want to Listen evidence for drug effects on scrambled>rest (P = .130, mean dif- ference 0.103, standard error 0.064). There were no other findings to Music from Post-Drug to Post-Scan was equivalent across the 3 sessions (drug by time interaction: F = 1.130, P = .321, involving drug (drug by contrast by region interaction: F = 0.687, 4,60 1,21 2 2 2 P = .604, η = 0.044; drug by region interaction: F = 0.919, P = .459, η = 0.070; main effect of drug: F = 9.158, P = .001, η = 0.379, p 2,30 p p 4,60 2 2 main effect of time: F = 7.164, P = .017, η = 0.323). Moreover, η = 0.058; main effect of drug: F = 0.585, P = .456, η = 0.038). p 1,15 p 1,15 p This suggests that Cann-CBD dampened response to music to a when comparing the 2 types of cannabis alone, there were no differences between the effects of Cann+CBD and Cann-CBD on similar extent across each of these regions (right auditory cor - tex, left auditory cortex, right hippocampus/parahippocampal Want to Listen to Music across the 3 time points (drug by time interaction: F = 0.804, P = .457, η = 0.051; main effect of drug: gyrus, right ventral striatum, and right amygdala) while having 2,30 p negligible effects on response to scrambled sound. F = 3.590, P = .078, η = 0.193, main effect of time: F = 8.251, 1,15 p 2,30 P = .001, η = 0.355). Brain-Behavior Correlations For Enhanced Sound Perception, Pre-Drug scores were removed from analysis due to floor effects on each session. Mean Next, we sought to examine correlations between brain (music>scrambled, extracted from the 5 clusters shown in (SD) values were 0.25 (0.45) on placebo, 0.00 (0.00) on Cann+CBD, Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 27 Table 2. MNI Coordinates for the Contrasts Music>Scrambled (Main Effect of Task, Top Panel) and Scrambled>Music (Main Effect of Task, Middle Panel) across All Sessions. The bottom panel shows brain regions in which participants’ response to music>scrambled was dampened follow- ing cannabis without CBD compared with placebo; +: additional peak within cluster. All P values are thresholded at P < .05 (FDR-corrected for multiple comparisons) x y Z mm P Main effect of task (music>scrambled) L Caudate -12 6 6 540 4.45 .006 L Amygdala -15 -3 -15 486 4.32 .006 L Hippocampus -18 -12 -18 + 3.59 .027 R Caudate/thalamus 9 3 6 594 3.91 .014 R Pallidum 15 -3 -6 54 3.33 .031 L Cingulate gyrus -6 -15 42 27 3.16 .035 R Amygdala 18 -3 -18 54 2.99 .040 Main effect of task (scrambled>music) R Planum temporale 60 -12 3 3834 6.61 <.001 R Planum temporale 54 -24 6 + 6.54 <.001 L Planum temporale -48 -33 9 2511 6.16 <.001 L Heschls gyrus -42 -24 3 + 5.26 <.001 Drug effect (placebo>cannabis without CBD) R Superior temporal gyrus 51 -27 6 2484 4.51 .005 R Planum temporale 60 -12 3 + 3.55 .016 R Planum temporale/heschls gyrus 42 -18 0 + 3.15 .026 L Planum temporale -42 -33 9 972 4.04 .008 R Hippocampus/parahippocampal gyrus 33 -18 -24 81 3.23 .025 R Amygdala 27 3 -27 27 3.19 .025 R Ventral striatum 15 15 -12 54 2.90 .033 Figure 4. Cannabis without cannabidiol (CBD) dampened brain response to music across several regions sensitive to music-evoked reward and emotion. (A) Bilateral auditory cortex activation clusters visualized on the cortical surface of a standard template (MNI152). (B) A ventral view of the same template showing right-hemi- sphere amygdala and hippocampal clusters. (C) Axial slice views of the same contrast showing amygdala, hippocampal, ventral striatal (top row), and auditory cortex (bottom row) activation clusters. All activation maps thresholded at < .05 (FDR corr P ected for multiple comparisons). anterior; L, A, left hemisphere; P posterior; R, , right hemisphere. the bottom panel of Table 2) and behavior (pleasure ratings all sessions to minimize type I and type II error using mixed for music>scrambled, Post-Drug Want to Listen to Music and effects models, resulting in a total of 15 correlations. One cor - Enhanced Sound Perception). Data were combined across relation reached statistical significance. This showed a positive Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 28 | International Journal of Neuropsychopharmacology, 2018 relationship between pleasure ratings and response to music region during music relative scrambled sound. Next, we exam- in right ventral striatum (F= 11.447, P = .002; Figure 5). The ined drug effects using t contrasts. Compared with Cann-CBD, 1,34 same relationship was found using a Pearson correlation ana- greater functional connectivity occurred on Cann+CBD between lysis across all scans (r = 0.463, P = .001). This correlation did not right ventral striatum and bilateral auditory cortex (Tabl ; e  3 contain any outlying values (all data points were <3 times the Figure 6). We also conducted a PPI analysis using an auditory cor - interquartile range). However, there was still evidence for a cor - tex seed (right superior temporal gyrus 51, -27, 6). However, this relation after excluding the 2 data points showing the highest PPI analysis did not identify any regions that showed increases and lowest right ventral striatum response to music (mixed in functional connectivity with the seed region. effects model: F = 4.438, P = .041; Pearson correlation analysis 1,41 r = 0.318, P = .032). Possible Confounding Cardiovascular Drug Effects Functional Connectivity We conducted correlations between all Post-Drug cardiovascu- Previous research has shown that the rewarding experience of lar measures (heart rate, systolic and diastolic blood pressure) music is predicted by increased functional connectivity between and the 9 clusters showing evidence of drug effects (see Tables right ventral striatum and auditory cortex (Salimpoor et al., 2013; 2 and 3) across all sessions (total 27 correlations). We found no Zatorre and Salimpoor 2013Martínez-Molina et  ; al., 2016). To test evidence for any association between cardiovascular and fMRI this, we conducted PPI analyses. These analyses were conducted data (all P > .05). posthoc, informed by our findings that Cann-CBD blunted partic- ipants’ response to music in right ventral striatum and auditory cortex. Within-subjects ANOVA revealed that across all sessions, Cannabis Use We also explored correlations between levels of cannabis of use the right ventral striatum region (15, 15, -12) identified in our ana- lysis showed a robust increase in functional connectivity with and our main findings. These were conducted between (1) years of cannabis use, (2) days of cannabis use per month and the 9 bilateral auditory cortex (and to a lesser extent, right caudate) during music relative to scrambled sound (Table ).  F 3 or comple- clusters showing evidence of drug effects (Tables 2 and ),3 Want to Listen to Music (Post-Drug), Enhanced Sound Perception (Post- tion, we conducted the reverse contrast. However, we found no evidence for any regions showing reduced connectivity with this Drug), and pleasure rating scores. Of the 24 correlations, we found Figure 5. Correlation between brain and behavior. (A) Axial slice of right ventral striatal region of interest, identified from voxelwise analysis. (B) Sagittal slice of the same region. (C) Across all scans, activation in right ventral striatum for the contrast music>scrambled correlated positively with pleasure ratings. Table 3. Functional Connectivity Analysis. MNI coordinates showing increased functional connectivity with right ventral striatum for music>scrambled across all sessions (main effect, top panel). Functional connectivity between right ventral striatum and auditory cortex increased on cannabis with CBD compared with cannabis without CBD (drug effect, bottom panel); +Additional peak within cluster. All P values are thresholded at < P .05 (False Discovery Rate-corrected for multiple comparisons) X y z mm Z P Main effect R Planum temporale 60 -12 6 5319 6.43 <.001 R Heschls gyrus/planum polare 48 -12 0 + 6.42 <.001 R Planum temporale 48 -27 9 + 6.26 <.001 L Heschls gyrus -42 -24 12 3429 6.31 <.001 L Planum temporale -42 -30 6 + 6.27 <.001 L Planum temporale -33 -33 15 + 5.48 <.001 R Caudate 9 15 9 54 2.42 .037 Drug effect (cannabis with CBD>cannabis without CBD) R Heschls gyrus 42 -18 9 1620 4.63 .003 L Hippocampus -30 -18 -21 81 3.64 .009 L Heschls gyrus -36 -27 9 54 3.08 .030 L Heschls gyrus -45 -24 15 27 3.05 .031 Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 29 Figure 6. Functional connectivity analysis. (A) Seed region in right ventral striatum. (B) This seed region showed increased task-related functional connectivity with bilateral auditory cortex following cannabis with cannabidiol (CBD) compared with cannabis without CBD (C). Axial slices depicting the same data in bilateral auditory cortex and additional left hippocampal cluster. All activation maps visualized on MNI152 and thresholded at < .05 (FDR corr P ected for multiple comparisons). L, left hemisphere; R, right hemisphere. no evidence for any associations (all > P .05) apart from a trend 2006; Salimpoor et al., 2011T ; rost et al., 2012) and, in common negative correlation between years of cannabis use and functional with THC, may increase striatal dopamine release (Salimpoor et connectivity between right ventral striatum and left hippocampus al., 2011; Bossong et al., 2015). Moreover, observational data sug- (F = 4.984, P = .030). However, this did not reach significance at a gests that cannabis is frequently used in the context of music 1,48 Bonferroni-corrected threshold ( =α 0.0021). Moreover, the effect of and may enhance its effects (Tart 1970Gr ; een et al., 2003; Lim et drug remained significant in this model (F = 7.455, P = .002). al., 2008; Van Havere et al., 2011; Palamar et al., 2015). 1,48 One possible explanation for our findings is that THC inter - Order Effects fered with the endocannabinoid system, which plays a critical Because the same music and scrambled sound excerpts were role in reward processing (Parsons and Hurd 2015). For example, presented across each of the 3 sessions, we investigated pos- acute THC may deplete the CB1R ligand anandamide (Thieme sible order effects. For all fMRI results showing drug effects, the et al., 2014), which increases consummatory response to reward effect of drug remained significant, and there was no evidence in the nucleus accumbens shell (Mahler et al., 2007). Disruption for an effect of session order (all > P .05). There was no evidence of the endocannabinoid system could explain why neural for effects of drug or session order for pleasure rating scores response to reward was previously dampened by 7-day admin- (all P > .05). Analysis of Want to Listen to Music (Post-Drug) and istration of a CB1R antagonist (Horder et al., 2010) as well as a Enhanced Sound Perception (Post-Drug) scores showed effects single dose of the partial CB1R agonist THC (van Hell et al., 2012). of drug (both P< .001) but not session (both P > .05). It should also be noted that our findings of dampened response to music occurred in the context of increased wanting to listen to music. These findings are broadly consistent with previous Discussion findings that THC may have dissociable effects on anticipatory To our knowledge, this is the first controlled experiment inves- (“wanting”) and consummatory (“liking”) components of reward tigating the interactive effects of cannabis and music. Cannabis (van Hell et  al., 2012; Jansma et  al., 2013), although our task dampened response to music in several regions implicated in lacked a neural index of reward anticipation. music-evoked reward and emotion (Koelsch 2014): bilateral audi- Cannabis with CBD did not differ from placebo on any fMRI tory cortex, right amygdala, right hippocampus/parahippocam- measures. Furthermore, it resulted in greater task-related pal gyrus, and right ventral striatum. Across all scans we found functional connectivity between ventral striatum and audi- a positive correlation between response to music in this ventral tory cortex compared with cannabis without CBD. These find- striatal region and the pleasure of listening to the same sound ings suggest that CBD was able to offset some effects of THC, clips, consistent with several studies implicating the ventral stri-consistent with previous research (Curran et al., 2016 Englund ; atum in musical pleasure (Blood and Zatorre 2001 Koelsc ; h et al., et al., 2017) and evidence that THC and CBD can have opposite 2006; Salimpoor et al., 2011T ; rost et al., 2012). The same ventral neural effects (Bhattacharyya et al., 2010 Batalla et  ; al., 2014). For striatal region showed increased task-related functional con- example, activation in right superior temporal gyrus (a region nectivity with bilateral auditory cortex, an effect that has pre-identified in our study) during word listening relative to rest was viously been shown to predict musical reward value (Salimpoor previously found to be decreased by THC but increased by CBD et al., 2013; Zatorre and Salimpoor 2013Martínez-Molina et al., ; (Winton-Brown et  al., 2011). Moreover, CBD may increase con- 2016). centrations of anandamide (Bisogno et  al., 2001Le ; weke et  al., These findings were contrary to our prediction that cannabis 2012). We found some evidence that CBD interacted with THC would increase the rewarding effects of music, which can acti- on additional measures. Taken together, CBD appeared to par - vate and increase connectivity within mesolimbic brain regions tially offset some negative effects of THC (increase in diastolic (Blood and Zatorre 2001Menon and Le ; vitin 2005; Koelsch et al., blood pressure, decreased response to music) while preserving Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 30 | International Journal of Neuropsychopharmacology, 2018 or potentiating desirable ones (enhanced sound perception, Acknowledgments functional connectivity with ventral striatum during musical These data were presented in preliminary form at the CINP listening). World Congress in Seoul, South Korea, the BAP Summer Meeting In terms of clinical implications, the effects of acute can- in Brighton, UK, and the ECNP Workshop for Junior Scientists nabis administration here are similar to previous findings in in Nice, France. We are grateful to Marty Sereno, Joseph Devlin, people with depression, who also show a blunted response to and the Birkbeck-UCL Centre for Neuroimaging team for their music in ventral striatum as well as medial orbitofrontal cortex assistance. (Osuch et al., 2009). In this respect, acute cannabis administra- tion may transiently mimic the diminished response to reward characteristic of some mental health disorders. The impact of Statement of Interest chronic cannabis administration remains unclear. However, a H.V.C. is a member of UK MRC boards and Drug Science. D.J.N. is 4-year prospective study found that increased cannabis use an advisor to the British National Formulary, MRC, GMC, was associated with subsequent reductions in ventral striatal Department of Health; President of European Brain Council; Past response to reward anticipation (Martz et  al., 2016). It there- President of British Neuroscience Association and European fore is possible that effects of cannabis on reward processing College of Neuropsychopharmacology, Chair of Drug Science may contribute to an increased risk of developing depression (UK); Member of International Centre for Science in Drug Policy; (Zhang et al., 2013; Lev-Ran et al., 2014) as well as other disor - advisor to Swedish government on drug, alcohol, and tobacco ders characterized by reward dysfunction such as addiction and research; editor of the Journal of Psychopharmacology; mem- psychosis (Radua et al., 2015; Luijten et al., 2017). Moreover, our ber of advisory boards of Lundbeck, MSD, Nalpharm, Orexigen, findings support the potential utility of CBD in reducing can- Shire, MSD; has received speaking honoraria (in addition to nabis harms while maintaining the positive effects users seek above) from BMS/Otsuka, GSK, Lilly, Janssen, Servier, AZ, and (Englund et al., 2017). Pfizer; is a member of the Lundbeck International Neuroscience Strengths of this study include its controlled experimen- Foundation; has received grants or clinical trial payments from tal design, comparison of cannabis with and without CBD (but P1vital, MRC, NHS, Lundbeck, RB; has share options in P1vital; matched for THC), a music task previously validated using fMRI has been an expert witness in a number of legal cases relat- (Menon and Levitin 2005), and regions of interest informed by ing to psychotropic drugs; and has edited/written 27 books, meta-analysis (Koelsch, 2014). Our sample size was equivalent some purchased by pharma companies. C.J.A.M. has consulted or larger than previous studies with comparable designs (van for Janssen and GlaxoSmithKline and received compensation. Hell et al., 2012; Jansma et al., 2013) and neuroimaging music M.B.W.  is employed by Imanova Ltd., a private company that studies of music (mean n= 14.5 across 22 studies (Koelsch performs contract research work for the pharmaceutical indus- 2014)). We used a fixed set of classical music excerpts, com- try. The other authors declare no potential conflicts of interest. mensurate with previous use of this task (Menon and Levitin, 2005) and many other studies (Koelsch, 2014). Advantages of this approach include the absence of lyrics (which would influ- References ence neural response due to speech) and ease of comparison Batalla A, Crippa J, Busatto G, Guimaraes F, Zuardi A, Valverde O, with existing data. Although participants rated classical sound Atakan Z, McGuire P, Bhattacharyya S, Martin-Santos R (2014) clips as highly pleasant (~7.5 of 10), results may have differed if Neuroimaging studies of acute effects of THC and CBD in preferred music was preselected by participants (Osuch et al., humans and animals: a systematic review. Curr Pharm Des 2009). Drug order was not completely balanced in this study, 20:2168–2185. and the same sound clips were presented on each of the 3 ses- Beck AT, Steer RA, Brown GK (1996) Beck depression inventory-II. sions. However, we found no evidence that session order influ- San Antonio, TX: The Psychological Corporation. enced our results. We screened for personal/family history of Berridge KC, Kringelbach ML (2015) Pleasure systems in the psychosis and current treatment for a psychiatric disorder, brain. Neuron 86:646–664. but not for lifetime history of other mental health problems. Bhattacharyya S, Morrison PD, Fusar-Poli P, Martin-Santos R, Additionally, our sample were cannabis users, which may have Borgwardt S, Winton-Brown T, Nosarti C, MO’Carroll C, Seal M, prolonged cannabinoid clearance between sessions. However, Allen P (2010) Opposite effects of Δ-9-tetrahydrocannabinol we found minimal evidence for associations between cannabis and cannabidiol on human brain function and psychopath- use and fMRI findings, and their response to cannabinoids may ology. Neuropsychopharmacology 35:764–774. be more representative of typical use than healthy volunteers Bisogno T, Hanuš L, De Petrocellis L, Tchilibon S, Ponde DE, who never use cannabis. Brandi I, Moriello AS, Davis JB, Mechoulam R, Di Marzo V In conclusion, cannabis dampened the effects of music in (2001) Molecular targets for cannabidiol and its synthetic bilateral auditory cortex, right hippocampus/parahippocam- analogues: effect on vanilloid VR1 receptors and on the cel- pal gyrus, right amygdala, and right ventral striatum. During lular uptake and enzymatic hydrolysis of anandamide. Br J musical listening, this ventral striatal region correlated with Pharmacol 134:845–852. pleasure ratings and showed increased functional connectivity Blood AJ, Zatorre RJ (2001) Intensely pleasurable responses with auditory cortex. By contrast, cannabis containing canna- to music correlate with activity in brain regions impli- bidiol did not influence the effects of music in brain regions sen- cated in reward and emotion. Proc Natl Acad Sci U S A sitive to reward and emotion. 98:11818–11823. Booth M (2004) Cannabis: a history. New York: St. Martin’s Press. Funding Bossong MG, van Berckel BN, Boellaard R, Zuurman L, Schuit RC, T.P.F. was funded by the UK Medical Research Council and a Senior Windhorst AD, van Gerven JM, Ramsey NF, Lammertsma AA, Kahn RS (2009) Δ9-tetrahydrocannabinol induces dopamine Academic Fellowship from the Society for the Study of Addiction. This study was funded by Drug Science/Channel 4 television. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 31 release in the human striatum. Neuropsychopharmacology cannabinoid CB1 antagonist rimonabant in healthy volun- 34:759–766. teers. Int J Neuropsychopharmacol 13:1103–1113. Bossong MG, Mehta MA, van Berckel BN, Howes OD, Kahn Jansma J, Van Hell H, Vanderschuren L, Bossong M, Jager G, Kahn RS, Stokes PR (2015) Further human evidence for striatal R, Ramsey N (2013) THC reduces the anticipatory nucleus dopamine release induced by administration of∆ 9-tetra- accumbens response to reward in subjects with a nicotine hydrocannabinol (THC): selectivity to limbic striatum. addiction. Transl Psychiatry 3:e234. Psychopharmacology (Berl) 232:2723–9. Koelsch S (2014) Brain correlates of music-evoked emotions. Nat Curran VH, Brignell C, Fletcher S, Middleton P, Henry J (2002) Rev Neurosci 15:170–180. Cognitive and subjective dose-response effects of acute oral Koelsch S, Fritz T, Müller K, Friederici AD (2006) Investigating emo- Δ9-tetrahydrocannabinol (THC) in infrequent cannabis users. tion with music: an fMRI study. Hum Brain Mapp 27:239–250. Psychopharmacology (Berl) 164:61–70. Lacadie CM, Fulbright RK, Rajeevan N, Constable RT, Papademetris Curran HV, Freeman TP, Mokrysz C, Lewis DA, Morgan CJ, Parsons X (2008) More accurate Talairach coordinates for neuroimag- LH (2016) Keep off the grass? Cannabis, cognition and addic- ing using non-linear registration. Neuroimage 42:717–725. tion. Nat Rev Neurosci 17:293–306. Lawn W, Freeman TP, Pope RA, Joye A, Harvey L, Hindocha C, Di Forti M, Marconi A, Carra E, Fraietta S, Trotta A, Bonomo Mokrysz C, Moss A, Wall MB, Bloomfield MA (2016) Acute and M, Bianconi F, Gardner-Sood P, O’Connor J, Russo M (2015) chronic effects of cannabinoids on effort-related decision- Proportion of patients in south London with first-episode making and reward learning: an evaluation of the canna- psychosis attributable to use of high potency cannabis: a bis ‘amotivational’hypotheses. Psychopharmacology (Berl) case-control study. Lancet Psychiatry 2:233–238. 233:3537–3552. D’Souza DC, Perry E, MacDougall L, Ammerman Y, Cooper T, Wu Lev-Ran S, Roerecke M, Le Foll B, George T, McKenzie K, Rehm J Y-t, Braley G, Gueorguieva R, Krystal JH (2004) The psych- (2014) The association between cannabis use and depression: otomimetic effects of intravenous delta-9-tetrahydrocan- a systematic review and meta-analysis of longitudinal stud- nabinol in healthy individuals: implications for psychosis. ies. Psychol Med 44:797–810. Neuropsychopharmacology 29:1558–1572. Leweke F, Piomelli D, Pahlisch F, Muhl D, Gerth C, Hoyer C, ElSohly MA, Mehmedic Z, Foster S, Gon C, Chandra S, Church JC Klosterkötter J, Hellmich M, Koethe D (2012) Cannabidiol (2016) Changes in cannabis potency over the last 2 decades enhances anandamide signaling and alleviates psychotic (1995–2014): analysis of current data in the United States. Biol symptoms of schizophrenia. Transl Psychiatry 2:e94. Psychiatry 79:613–619. Lim MS, Lim MS, Hellard ME, Lim MS, Hellard ME, Hocking JS, Lim Englund A, Freeman TP, Murray RM, McGuire P (2017) Can we MS, Hellard ME, Hocking JS, Aitken CK (2008) A cross-sectional make cannabis safer? Lancet Psychiatry 4:643–648. survey of young people attending a music festival: associa- Freeman TP, Morgan CJ, Vaughn-Jones J, Hussain N, Karimi K, tions between drug use and musical preference. Drug Alcohol Curran HV (2012) Cognitive and subjective effects of mephe- Rev 27:439–441. drone and factors influencing use of a ‘new legal high’. Luijten M, Schellekens AF, Kühn S, Machielse MW, Sescousse Addiction 107:792–800. G (2017) Disruption of reward processing in addiction: an Freeman TP, Morgan CJA, Hindocha C, Schafer GL, Das RK, Curran image-based meta-analysis of functional magnetic reson- HV (2014) Just say ‘know’: how do cannabinoid concentra- ance imaging studies. JAMA Psychiatry 74:387–398. tions influence users’ estimates of cannabis potency and the Mahler SV, Smith KS, Berridge KC (2007) Endocannabinoid amount they roll in joints? Addiction 109:1686–1694. hedonic hotspot for sensory pleasure: anandamide in Freeman TP, Winstock AR (2015) Examining the profile of high- nucleus accumbens shell enhances’ liking’of a sweet reward. potency cannabis and its association with severity of canna- Neuropsychopharmacology 32:2267. bis dependence. Psychol Med 45:3181–3189. Martínez-Molina N, Mas-Herrero E, Rodríguez-Fornells A, Gard DE, Gard MG, Kring AM, John OP (2006) Anticipatory and Zatorre RJ, Marco-Pallarés J (2016) Neural correlates of specific consummatory components of the experience of pleasure: a musical anhedonia. Proc Natl Acad Sci U S A 113:E7337–E7345. scale development study. J Res Personal 40:1086–1102. Martz ME, Trucco EM, Cope LM, Hardee JE, Jester JM, Zucker RA, Gossop M, Darke S, Griffiths P, Hando J, Powis B, Hall W, Strang Heitzeg MM (2016) Association of marijuana use with blunted J (1995) The Severity of Dependence Scale (SDS): psycho- nucleus accumbens response to reward anticipation. JAMA metric properties of the SDS in English and Australian sam- Psychiatry 73:838–844. ples of heroin, cocaine and amphetamine users. Addiction Menon V, Levitin DJ (2005) The rewards of music listening: 90:607–614. response and physiological connectivity of the mesolimbic Green B, Kavanagh D, Young R (2003) Being stoned: a review of system. Neuroimage 28:175–184. self-reported cannabis effects. Drug Alcohol Rev 22:453–460. Mokrysz C, Freeman T, Korkki S, Griffiths K, Curran H (2016) Are Hardwick S, King LA (2008) Home Office Cannabis Potency Study adolescents more vulnerable to the harmful effects of canna- 2008. Home Office: Home Office Scientific Development bis than adults? A placebo-controlled study in human males. Branch United Kingdom. Translational Psychiatry 6:e961. Hindocha C, Freeman TP, Ferris JA, Lynskey MT, Winstock AR O’Reilly JX, Woolrich MW, Behrens TE, Smith SM, Johansen-Berg H (2016) No smoke without tobacco: a global overview of can- (2012) Tools of the trade: psychophysiological interactions and nabis and tobacco routes of administration and their associ- functional connectivity. Soc Cog Affective Neurosci 7:604–609. ation with intention to quit. Front Psychiatry 7:104. Osuch EA, Bluhm RL, Williamson PC, Theberge J, Densmore M, Hindocha C, Freeman TP, Schafer G, Gardener C, Das RK, Morgan CJ, Neufeld RW (2009) Brain activation to favorite music in healthy Curran HV (2015) Acute effects of delta-9-tetrahydrocannabi- controls and depressed patients. Neuroreport 20:1204–1208. nol, cannabidiol and their combination on facial emotion rec- Palamar JJ, Griffin-Tomas M, Ompad DC (2015) Illicit drug use ognition: a randomised, double-blind, placebo-controlled study among rave attendees in a nationally representative sample in cannabis users. Eur Neuropsychopharmacol 25:325–334. of US high school seniors. Drug Alcohol Depend 152:24–31. Horder J, Harmer CJ, Cowen PJ, McCabe C (2010) Reduced neural Parsons LH, Hurd YL (2015) Endocannabinoid signaling in reward response to reward following 7  days treatment with the and addiction. Nat Rev Neurosci 16:579–94. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 32 | International Journal of Neuropsychopharmacology, 2018 Radua J, Schmidt A, Borgwardt S, Heinz A, Schlagenhauf F, plasma levels to examine potential influences of tetrahydro- McGuire P, Fusar-Poli P (2015) Ventral striatal activation dur - cannabinol application on the endocannabinoid system in ing reward processing in psychosis: a neurofunctional meta- humans. Drug Test Anal 6:17–23. analysis. JAMA Psychiatry 72:1243–1251. Trost W, Ethofer T, Zentner M, Vuilleumier P (2012) Mapping Room R (2014) Legalizing a market for cannabis for pleas- aesthetic musical emotions in the brain. Cereb Cortex ure: Colorado, Washington, Uruguay and beyond. Addiction 22:2769–2783. 109:345–351. Van Havere T, Vanderplasschen W, Lammertyn J, Broekaert E, Salimpoor VN, Benovoy M, Larcher K, Dagher A, Zatorre RJ (2011) Bellis M (2011) Drug use and nightlife: more than just dance Anatomically distinct dopamine release during anticipation and music. Subst Abuse Treat Prev Policy 6:18. experience of peak emotion to music. Nat Neurosci 14:257–262. van Hell HH, Jager G, Bossong MG, Brouwer A, Jansma JM, Zuurman Salimpoor VN, van den Bosch I, Kovacevic N, McIntosh AR, L, Van Gerven J, Kahn RS, Ramsey NF (2012) Involvement of Dagher A, Zatorre RJ (2013) Interactions between the nucleus the endocannabinoid system in reward processing in the accumbens and auditory cortices predict music reward value. human brain. Psychopharmacology 219:981–990. Science 340:216–219. Winton-Brown TT, Allen P, Bhattacharrya S, Borgwardt SJ, Fusar- Schafer G, Feilding A, Morgan CJ, Agathangelou M, Freeman Poli P, Crippa JA, Seal ML, Martin-Santos R, Ffytche D, Zuardi TP, Curran HV (2012) Investigating the interaction between AW (2011) Modulation of auditory and visual processing by schizotypy, divergent thinking and cannabis use. Conscious delta-9-tetrahydrocannabinol and cannabidiol: an FMRI Cogn 21:292–298. study. Neuropsychopharmacology 36:1340–1348. Sobell LC, Sobell MB (1992) Timeline follow-back measuring Zatorre RJ, Salimpoor VN (2013) From perception to pleasure: alcohol consumption. New York: Springer, pp41–72. music and its neural substrates. Proc Natl Acad Sci U S A Tart CT (1970) Marijuana intoxication: common experiences. 110:10430–10437. Nature 226:701–704. Zhang WN, Chang SH, Guo LY, Zhang KL, Wang J (2013) The neural Thieme U, Schelling G, Hauer D, Greif R, Dame T, Laubender correlates of reward-related processing in major depressive RP, Bernhard W, Thieme D, Campolongo P, Theiler L (2014) disorder: a meta-analysis of functional magnetic resonance Quantification of anandamide and 2-arachidonoylglycerol imaging studies. J Affect Disord 151:531–539. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Neuropsychopharmacology Oxford University Press

Loading next page...
 
/lp/ou_press/cannabis-dampens-the-effects-of-music-in-brain-regions-sensitive-to-tl6QYOO6Rv
Publisher
Collegium Internationale Neuro-Psychopharmacologicum
Copyright
© The Author 2017. Published by Oxford University Press on behalf of CINP.
ISSN
1461-1457
eISSN
1469-5111
D.O.I.
10.1093/ijnp/pyx082
Publisher site
See Article on Publisher Site

Abstract

Background: Despite the current shift towards permissive cannabis policies, few studies have investigated the pleasurable effects users seek. Here, we investigate the effects of cannabis on listening to music, a rewarding activity that frequently occurs in the context of recreational cannabis use. We additionally tested how these effects are influenced by cannabidiol, which may offset cannabis-related harms. Methods: Across 3 sessions, 16 cannabis users inhaled cannabis with cannabidiol, cannabis without cannabidiol, and placebo. We compared their response to music relative to control excerpts of scrambled sound during functional Magnetic Resonance Imaging within regions identified in a meta-analysis of music-evoked reward and emotion. All results were False Discovery Rate corrected (P < .05). Results: Compared with placebo, cannabis without cannabidiol dampened response to music in bilateral auditory cortex (right: P = .005, left: P = .008), right hippocampus/parahippocampal gyrus (P = .025), right amygdala (P = .025), and right ventral striatum (P = .033). Across all sessions, the effects of music in this ventral striatal region correlated with pleasure ratings (P = .002) and increased functional connectivity with auditory cortex (right: P < .001, left: P < .001), supporting its involvement in music reward. Functional connectivity between right ventral striatum and auditory cortex was increased by cannabidiol (right: P = .003, left: P = .030), and cannabis with cannabidiol did not differ from placebo on any functional Magnetic Resonance Imaging measures. Both types of cannabis increased ratings of wanting to listen to music (P < .002) and enhanced sound perception (P < .001). Conclusions: Cannabis dampens the effects of music in brain regions sensitive to reward and emotion. These effects were offset by a key cannabis constituent, cannabidol. Received: May 4, 2017; Revised: August 16, 2017; Accepted: August 30, 2017 © The Author(s) 2017. Published by Oxford University Press on behalf of CINP. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons. org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is 21 properly cited. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 22 | International Journal of Neuropsychopharmacology, 2018 Significance Statement Here, we report that cannabis administration decreased response to music in several brain regions linked to reward and emotion. These included right ventral striatum, which showed increased functional connectivity with auditory cortex and correlated with pleasure ratings during musical listening, consistent with its critical role in reward processing. These effects were offset when cannabis contained cannabidiol, a key cannabinoid that has been found to reduce some harmful effects of cannabis. Keywords: cannabis, music, reward, pleasure, emotion Introduction The main psychoactive constituent of cannabis, THC (delta- Additionally, THC causes modest, regionally selective 9-tetrahydrocannabinol), produces subjective effects such as increased dopamine release in limbic striatum (Bossong et al., feeling “stoned” and can impair memory and elicit transient 2015). Such effects might enhance the rewarding experience of psychotic-like symptoms (Curran et  al., 2016). Certain types of music, which can also elicit dopamine release in ventral stri- cannabis also contain cannabidiol (CBD), which can have oppos- atum (Salimpoor et al., 2011) as well as enhancing activation ite effects of THC on a range of functional neuroimaging tasks and connectivity between mesolimbic brain regions (Blood (Bhattacharyya et  al., 2010Batalla et  ; al., 2014). Moreover, CBD and Zatorre 2001; Menon and Levitin 2005; Koelsch et al., 2006; has been found to offset harmful effects of THC (e.g., memory Salimpoor et al., 2011T ; rost et al., 2012). Functional connectivity impairment and psychotic-like symptoms) without influencing between ventral striatum and auditory cortex during listening subjective intoxication (Curran et al., 2016 Englund et  ; al., 2017). also predicts the rewarding experience of music (Salimpoor et Cannabis containing high THC and little if any CBD is becoming al., 2013; Zatorre and Salimpoor 2013; Martínez-Molina et al., increasingly prevalent (Hardwick and King 2008 ElSohl ; y et  al., 2016). 2016) and has been linked to greater mental health and addic- Here, we conducted the first controlled experimental study tion problems compared with less potent varieties of cannabis on the interactive effects of cannabis and music. Based on pre- (Di Forti et al., 2015; Freeman and Winstock 2015). vious findings that cannabis and music activate and increase Despite the changes currently occurring in cannabis legisla- connectivity between common regions in the reward net- tion worldwide, including legalization of use for medicine and work, whereas a CB1R antagonist dampened neural response pleasure (Room 2014), few studies have attempted to document to reward (Horder et  al., 2010), and observational data linking the effects that recreational users seek (Curran et al., 2016). The cannabis use and music, we hypothesized that cannabis would limited evidence of positive effects tends to have arisen inci- increase haemodynamic response to music in brain regions dentally in studies investigating cannabis-related harms. For sensitive to reward and emotion (Koelsch 2014) as well as sub- example, THC has been reported to increase phonological flu- jective ratings (wanting to listen to music, pleasure of listen- ency (Curran et al., 2002), a measure of divergent thinking, espe- ing). Given that CBD and THC can have opposing neural effects cially among people with low trait creativity (Schafer et al., 2012).(Bhattacharyya et  al., 2010Batalla et  ; al., 2014) and CBD can Cannabis has a strong historical link to music and is associated attenuate THC harms (Curran et al., 2016 Englund et  ; al., 2017), with several distinct styles, including jazz, reggae, and rock (Booth we predicted that these effects would be partially offset by CBD. 2004). Cannabis is reported to enhance appreciation of music (Tart 1970; Green et  al., 2003), and its use is consistently high among Methods people who attend music festivals and nightclubs (Lim et al., 2008; Van Havere et al., 2011; Palamar et al., 2015). This association may Design and Participants be partly attributable to shared effects on reward circuitry between drug and nondrug rewards (Berridge and Kringelbach 2015). Music A randomized, double-blind, crossover design compared can- recruits key regions in the reward network, including ventral stri- nabis with CBD (Cann+CBD), cannabis without CBD (Cann-CBD) atum, mediodorsal thalamus, anterior insula, orbitofrontal cortex,and matc hed placebo in 16 cannabis users. Experimental ses- amygdala, and hippocampus (Koelsch 2014). sions were separated by at least 1 week (>3 times the elimin- Many of these reward-related brain regions are character - ation half-life of THC) to minimize carryover effects (D’Souza ized by a high density of Cannabinoid Type-1 Receptors (CB1Rs) et al., 2004; Hindocha et al., 2015). In addition to the music task (Curran et al., 2016). THC is a partial agonist of CB1Rs and may described here, participants completed additional assessments influence response to music by interfering with endogenous that are reported elsewhere (Lawn et  al., 2016). Inclusion cri- CB1R ligands such as anandamide (Thieme et al., 2014), which teria were fluency in English, right-handedness, age between plays a causal role in consummatory response to reward (Mahler 18 and 70  years, and self-reported current cannabis use (≥4 et  al., 2007). A  human neuroimaging study found that THC (a times in the last year, ≤3 times/wk, ability to smoke a whole partial CB1R agonist) dampened the effects of monetary reward joint to oneself). We did not collect data on participants’ typ- feedback across a widespread network, including temporal and ical method of administering cannabis. However, previous orbitofrontal cortices, while leaving reward anticipation intact data from the UK suggest that the majority (~76%) of cannabis (van Hell et  al., 2012). By contrast, 7-day administration of a users typically smoke cannabis together with tobacco in joints, CB1R antagonist was found to diminish response to food reward and only a small minority (~4%) use a vaporizer as their most in ventral striatum and orbitofrontal cortex (Horder et al., 2010).common route (Hindocha et  al., 2016). Exclusion criteria were Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 23 self-reported frequent and/or severe adverse reactions to can- them with a 30-millisecond linear cross-fade between excerpts nabis, current use of illicit drugs other than cannabis more than (Menon and Levitin 2005). Scrambled excerpts retain the same twice per month, current alcohol use >4 d/wk, significant phys- distribution of pitch and loudness and the same spectral infor - ical health problems, color blindness, current treatment for a mation as normal music. However, they lack temporal structure psychiatric disorder, current/history of psychosis, and current/ and are rated as less pleasurable than normal excerpts (Menon history of psychosis in an immediate family member. This study and Levitin 2005). was approved by the UCL ethics committee, and all participants To deliver clear audio during scanning, clips were adapted provided written informed consent. to improve volume constancy during sections of low volume. Output volume was adapted for each participant in the scanner before the task commenced. Normal/scrambled excerpts were Procedure delivered using PsychoPy (Version 1.79.01) through MR com- Following telephone screening, eligible participants completed patible sensimetric earphones (http://www.sens.com/products/ a baseline session consisting of task training (outside of the MRI model-s14/) in a standard blocked fMRI design. The 12 normal/ scanner), video training for drug administration, drug history scrambled excerpts were presented in a pseudo-randomized (Freeman et al., 2012), and problematic cannabis use on Severity order across the 3 test sessions. Each 21-second excerpt was fol- of Dependence Scale (Gossop et al., 1995). The Beck Depression lowed by a 1-second interstimulus interval. Next, participants Inventory-II (Beck et  al., 1996) and Temporal Experiences of rated the pleasantness of each excerpt using a 2-finger response Pleasure (Gard et al., 2006) were also administered. Each experi- pad beneath their right hand (fixed time of 8 seconds). The mental session began with a urinary drug screen to verify recent numerical rating scale was anchored from 0 (not at all pleasant) use reported by Timeline Follow-back (Sobell and Sobell 1992). to 10 (very pleasant). This was followed by 12 seconds of pas- Next, 11-point (0–10) Numerical Rating Scales were adminis- sive fixation (rest). The total task time was 8 minutes 24 seconds, tered ~0 minutes before drug inhalation (Pre-Drug), ~5 minutes plus a 5-second end-buffer period. after first drug administration (Post-Drug), and ~90 minutes after first drug administration (Post-Scan). The Numerical fMRI Data Acquisition Rating Scales “Want to Listen to Music” and “Enhanced Sound Perception” were administered at all 3 of these time points; “Feel Imaging data were collected using a Siemens TIM Avanto 1.5T Drug Effect,” “Like Drug Effect,” and “Want More Drug” were scanner, using a 32-channel receive-only head coil, at the administered only after drug administration (Post-Drug and Birkbeck-UCL Centre for Neuroimaging, London. An automated Post-Scan). Heart rate and systolic and diastolic blood pressure shim procedure was applied to minimize possible magnetic were also recorded at the same 3 time points (Pre-Drug, Post- field homogeneities. Functional imaging used a multiband Drug, Post-Scan). (acceleration factor = 4) gr adient-echo T2*-weighted echo-planar imaging (EPI) sequence with 40 slices per volume (TR = 1000 ms; Drug Administration TE = 55 ms; in-plane matrix = 64 x 64; 3 mm isotropic voxels; flip angle= 75°; bandwidth= 1474 Hz/pixel; 509 volumes). The first 8 Cannabis was obtained from Bedrocan, The Netherlands and scans were treated as “dummy” scans and discarded to avoid used within 6  months of purchase. It was stored on site in T1-equilibrium effects. All scanning parameters were selected foil-sealed pouches at -20ºC and then at ambient temperature to optimize the quality of the BOLD signal while maintaining a prior to drug administration. Each dose was vaporized using a sufficient number of slices to acquire whole-brain data. To co- Volcano Medic Vaporizer (Storz and Bickel) at 210ºC in 2 sequen- register the fMRI data into standard space, we also acquired a tially administered balloons to minimize residual cannabinoids MPRAGE structural sequence (TR = 2730  ms; TE = 3.57  ms; mat- (Lawn et al., 2016). Participants inhaled at their own pace (each rix = 176 x 256 x 256; 1-mm isotropic voxels; flip angle = 7°; inhalation held for 8 seconds, enforced by the experimenter bandwidth= 190 Hz/pixel; parallel imaging acceleration fac- using a stopwatch) until the balloon was empty, which lasted ~5 tor= 2), and a B0 field map image (64 axial slices; TR = 1170 ms; minutes for both balloons. All participants complied with this TE1 = 10.0  ms; TE2 = 14.76  ms; in-plane matrix= 64 x 64; 3 x 3 x administration protocol. Bedrobinol (12% THC, <1% CBD), Bediol 2 mm voxels; flip angle = 90°; bandwidth= 260 Hz/pixel) to enable (6% THC, 7.5% CBD), and placebo cannabis were used to load distortion correction of the functional data. doses of 8 mg THC + 10 mg CBD (Cann+CBD), 8 mg THC (Cann- CBD), and placebo (Lawn et  al., 2016). Placebo cannabis had a comparable terpene profile to the 2 active forms of cannabis, fMRI Data Analysis ensuring it was matched for smell. The same physical quantity Preprocessing and data analysis were performed using of cannabis/placebo (133.4mg) was administered across each Statistical Parametric Mapping (SPM8; http://www.fil.ion.ucl. of the 3 sessions. This dose of THC has produced effects on ac.uk/spm/software/spm8/). Standard preprocessing procedures brain and behavior in studies with similar vaporizer protocols consisted of bias correction of EPI images to control for within- (Bossong et al., 2009; Hindocha et al., 2015; Mokrysz et al., 2016) volume signal intensity differences, realignment/unwarping and is roughly equivalent to one-quarter of a standard UK joint to correct for interscan movements, correction for differences (Freeman et al., 2014). in slice acquisition timing, and normalization of the images to an EPI template specific to our sequence and scanner that was Music Task (Menon and Levitin 2005) aligned to the T1 MNI template. Finally, the normalized func- Six 21-second excerpts of standard instrumental classical tional images were spatially smoothed with an isotropic 8-mm FWHM Gaussian kernel. music were taken from compact disc recordings, adapted from a previous study (Menon and Levitin 2005). Six scrambled ver - At the first level, the normal and scrambled epochs were each modelled as a 21-second boxcar convolved with the canon- sions were created by randomly drawing 250- to 350-millisec- ond variable-sized sections from each piece and concatenating ical hemodynamic response function combined with time and Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 24 | International Journal of Neuropsychopharmacology, 2018 dispersion derivatives to create the contrast music>scrambled, SDS, and drug history) were missing for one participant. Missing as previously used with this task (Menon and Levitin 2005). data from experimental sessions (0.69% of Numerical Rating The interstimulus interval, rating, and passive fixation (rest) Scale data, 0.69% of cardiovascular data) were imputed with the were also modelled. Each subject’s movement parameters were mean for that session and time point to retain each participant included as confounds. Low-frequency noise was removed with in the repeated-measures analysis. Repeated-measures ANOVA a high-pass filter (cut-off frequency 1/128 Hz). Parameter esti- models were used for all data collected on the 3 experimental mates pertaining to the height of the hemodynamic response sessions, including within-subject factors of drug (Cann+CBD, function for each regressor of interest were then calculated for Cann-CBD, placebo) and time (Pre-Drug, Post-Drug, Post-Scan) each voxel. or (Post-Drug, Post-Scan) and additional factors where appro- At the second level, the contrast music>scrambled was priate. Posthoc pairwise tests were Bonferroni-corrected locally entered into a within-subject ANOVA model with a single fac- within each ANOVA model. Additional repeated-measures tor of drug (Cann+CBD, Cann-CBD, placebo). Within this ANOVA ANOVA models were used to aid interpretation of interactions model, we used t contrasts to investigate music>scrambled and where appropriate. The Greenhouse-Geisser correction was the reverse contrast (scrambled>music) across all scans. Drug applied where assumptions of sphericity were violated, with effects on music>scrambled were conducted using t contrasts degrees of freedom rounded to the nearest integer. To reduce within this ANOVA model. To aid interpretation of signifi- type I and type II error rates, correlations with fMRI data were cant drug effects (which could reflect changes in response to collapsed across each of the sessions using mixed effects mod- music, scrambled, or both), separate parameter estimates were els, with a Bonferroni-adjusted α threshold. These accounted for extracted from these coordinates for the contrasts music>rest fixed effects of drug and session order, with a random intercept and scrambled>rest using the MarsBaR region of interest tool- of participant and maximum likelihood estimation. Equivalent box; these were analyzed using repeated-measures ANOVA mixed effects models were used to assess possible confound- in SPSS. ing by cardiovascular measures, cannabis use, and session order. Psychophysiological interaction (PPI) analysis was per - formed to assess task-related functional connectivity (O’Reilly Results et  al., 2012) using seed regions identified by drug effects. We extracted the representative time-course from voxels in the Participants seed region (6-mm radius sphere) using the first eigenvari- ate calculated from singular value composition. This time Seventeen participants completed the study. One participant course (physiological) was entered into a General Linear Model was excluded due to excessive head movement on one session together with the contrast music>scrambled (psychological) (exceeding thresholds for both translation [>6 mm] and rotation and their interaction (PPI). Motion parameters were included in [>6º]) and was replaced, leaving a final sample of 16. Demographic first-level models as nuisance regressors. The PPI regressor was and drug use data are shown in Table  1. The following num- analyzed using a within-subject ANOVA. We used contr t asts ber of participants completed each treatment order: Placebo, to investigate PPI effects across all scans and to compare drug Cann+CBD, Cann-CBD: n = 3; Placebo, Cann-CBD, Cann+CBD: effects. n = 2; Cann+CBD, Placebo, Cann-CBD: n= 3; Cann+CBD, Cann- A False Discovery Rate correction (< P .05) w as applied to all CBD, Placebo: n= 3; Cann-CBD, Placebo, Cann+CBD: n= 2; Cann- fMRI analyses. Regions of interest were defined from a previ- CBD, Cann+CBD, Placebo: n= 3. ous meta-analysis of music-evoked reward and emotion (see Figure  1 and supplementary Table  1 in Koelsch 2014) using the MarsBaR toolbox. Firstly, each of the structures identified in the Behavioral Results meta-analysis was converted into a single sphere. Coordinates Subjective Effects were converted to MNI using the Yale BioImage Suite (Lacadie Subjective effects are shown in Figure 2 . A main effect of drug et al., 2008). Sphere radius was estimated from the cluster size (F = 107.659, P < .001, η = 0.878) emerged for Feel Drug Effect, reported in the meta-analysis. Where clusters contained mul- 1,22 p reflecting increased scores following Cann+CBD (P < .001) and tiple structures, size was determined using the cluster mean. Cann-CBD (P < .001) compared with placebo, but no differences Subthreshold clusters were assigned a default size of 200 mm . between Cann+CBD and Cann-CBD (P = 1.000). There was also Each of these spheres was combined into a single mask that a main effect of time, indicating that scores decreased from was applied to second-level analysis. This mask (41 240  mm ) Post-Drug to Post-Scan (F = 19.057, P < .001, η = 0.560), but included bilateral hippocampal formation, bilateral amygdala, 1,15 p there was no evidence for an interaction between drug and bilateral auditory cortex, right ventral striatum, left caudate time (F = 0.796, P = .461, η = 0.050). Like Drug Effect showed nucleus, presupplementary motor area, frontomedian cortex, 2, 30 p a similar profile of results. There was a main effect of drug rostral cingulate zone, pre-genual and middle cingulate cortex, (F = 44.371, P < .001, η = 0.747), reflecting increased scores fol- medial and laterial orbitofrontal cortex, right anterior insula, 2,30 p lowing Cann+CBD (P < .001) and Cann-CBD (P < .001) compared mediodorsal thalamus, and superior parietal lobule. with placebo but no difference between Cann + CBD and Cann- CBD (P = 1.000). There was also a main effect of time, indicating Behavioral Data Analysis that scores decreased from Post-Drug to Post-Scan (F = 19.454, 1,15 SPSS version 21 was used to analyze all behavioral data and par - P < .001, η = 0.565). Again, there was no evidence for an inter - ameter estimates extracted posthoc using MarsBaR. Outliers action between drug and time (F = 0.589, P = .561, η = 0.038). 2,30 p (>3 times IQR) were winsorized within each session and time For Want More Drug, there was no evidence for any effects or point. Histograms were used to investigate normality, and interactions: drug by time (F = 2.462, P = .102, η = 0.141), drug 2,30 p square root or log transformations were applied where appro- (F = 1.329, P = .280, η = 0.081), or Time (F = 0.388, P = .543, 2,30 p 1,15 priate. Trait measures (BDI, Temporal Experiences of Pleasure, η = 0.025). Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 25 Figure  1. Subjective effects. Both types of cannabis increased ratings for (A) Feel Drug Effect and (B) Like Drug Effect but did not influence (C) Want More Drug. Cann + CBD, cannabis with cannabidiol (CBD); Cann-CBD canna , bis without CBD. ***P < .001. Figure 2. Cardiovascular effects. Both types of cannabis increased (A) heart rate and (B) systolic blood pressure. (C) Diastolic blood pressure increased from Pre- to Post- Drug following cannabis without cannabidiol (CBD), but not following cannabis with CBD; Cann + CBD , cannabis with CBD; Cann-CBD ,cannabis without CBD; *P < .05, ***P < .001; Difference between cannabis types. Figure 3. Subjective music ratings. (A) Both types of cannabis increased ratings of Want to Listen to Music. (B) Both types of cannabis increased scores for Enhanced Sound Perception and this increase was greater for cannabis with cannabidiol (CBD). (C) Neither type of cannabis influenced the pleasure of listening to music or scrambled sound clips. Cann + CBD, cannabis with CBD; Cann-CBD ,cannabis without CBD. ***P < .001; Difference between cannabis types. Cardiovascular Effects cannabis alone, there were no differences between the effects of Cardiovascular effects are shown in Figure  2. For heart rate Cann+CBD and Cann-CBD on heart rate across the 3 time points (BPM), a drug by time interaction emerged (F = 18.243, P < .001, (drug by time interaction: F = 0.123, P = .885, η = 0.008; main 2,28 2,30 p 2 2 η = 0.549) as well as main effects of both drug (F = 13.999, effect of drug: F = 0.090, P = .768, η = 0.006, main effect of time: p 2,30 1,15 p 2 2 2 P < .001, η = 0.483) and time (F = 45.977, P < .001, η = 0.754). F = 87.391, P < .001, η = 0.854). For systolic blood pressure, a p 2,30 p 2,30 p Heart rate increased from Pre-Drug to Post-Drug following main effect of drug was found (F = 6.297, P = .005, η = 0.296). 2,30 p Cann+CBD (P < .001) and Cann-CBD (P < .001) but not placebo This reflected increased blood pressure for both Cann+CBD (P = .456). It then decreased from Post-Drug to Post-Scan for (P = .030) and Cann-CBD (P = .006), compared with placebo, but both Cann+CBD (P < .001) and Cann-CBD (P < .001) but did not no differences between Cann+CBD and Cann-CBD (P = 1.000). change on placebo (P = 1.000). When comparing the 2 types of There was no evidence for an interaction between drug and Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 26 | International Journal of Neuropsychopharmacology, 2018 Table 1. Demographic and Drug Use Data and 0.00 (0.00) on Cann-CBD. Analysis of variance was there- fore restricted to 2 time points (Post-Drug, Post-Scan). A  main Mean/ effect of drug (F = 44.810, P < .001, η = 0.749) reflected increased 2,30 p frequency SD scores from placebo following Cann+CBD (P < .001) and Cann- CBD (P < .001) and higher scores following Cann+CBD compared Age 26.25 7.35 with Cann-CBD (P = .015). There was no evidence for an inter - Gender (male/female) 8/8 - action between drug and time (F = 2.056, P = .146, η = 0.121) or 2,30 p Days of cannabis use per month 8.06 5.48 a main effect of time (F = 1.248, P = .281, η = 0.077). Finally, we Years of cannabis use 8.94 7.02 1,15 p analyzed trial-by-trial pleasure ratings, recorded immediately Days since last cannabis use 19.25 45.28 after listening to classical music and scrambled sound excerpts Days to smoke 3.5 g cannabis 25.88 33.73 during MRI scanning. There was a main effect of excerpt Severity of dependence scale (cannabis) 1.13 1.26 (F = 133.860, P < .001, η = 0.899), indicating that music was rated Alcohol use (yes/no) 16/0 - 1,15 p Days of alcohol use per month 10.81 4.86 as more pleasant than scrambled sound. However, there was no Number of UK alcohol units (8 g) per 5.93 2.08 evidence for a main effect of drug (F = 1.205, P = .314, η = 0.074) 2,30 p session or a drug by excerpt interaction (F = 1.221, P = .309, η = 0.075). 2,30 p Current tobacco use (yes/no) 15/1 - Next, we calculated a pleasure rating score (music>scrambled) Days of tobacco use per month 11.30 10.27 equivalent to our fMRI contrast of interest to provide compar - Cigarettes per day 3.63 3.62 able metrics for brain and behavior. Mean (SD) pleasure rating Current MDMA use <twice a month (yes/ 6/10 - scores were 5.16 (2.27) for Cann+CBD, 4.78 (2.03) for Cann-CBD, no) and 5.53 (1.99) for placebo. Analysis of these scores provided no Current cocaine use <twice a month (yes/ 3/13 - evidence for an effect of drug (F = 1.221, P = .309, η = 0.075). 2,30 p no) Current ketamine use <twice a month 2/14 - fMRI Results (yes/no) Beck Depression Inventory-II 3.38 3.12 Main Effect of Task Temporal experiences of pleasure 42.06 4.85 All fMRI analyses were conducted among regions of inter - (anticipatory) est selected from a meta-analysis of previous studies (Koelsch Temporal experiences of pleasure 43.50 5.61 2014). Across all sessions, listening to music elicited activation (consummatory) in bilateral amygdala, bilateral striatum, left hippocampus, and left cingulate gyrus (see Table 2). For completion, we also exam- time (F = 0.953, P = .440, η = 0.060) or a main effect of time ined the reverse contrast (scrambled>music), which revealed 4,60 p activation in bilateral auditory cortex (see Table 2). (F = 2.641, P = .088, η = 0.150). For diastolic blood pressure, 2,30 p an interaction between drug and time was found (F = 3.217, 4,60 Drug Effects P = .019, η = 0.177) and a main effect of time (F = 7.702, P = .002, p 2,30 2 2 η = 0.339) but not drug (F = 2.975, P = .066, η = 0.165). Diastolic Response to music>scrambled was greater on placebo compared p 2,30 p with Cann-CBD in bilateral auditory cortex, right hippocam- blood pressure increased from Pre-Drug to Post-Drug for Cann- CBD (P < .001) but not Cann+CBD (P = .233) or placebo (P = 1.000). pus/parahippocampal gyrus, right ventral striatum, and right amygdala (see Table 2 and Figure 4). There was no evidence for It then increased from Post-Drug to Post-Scan following placebo (P = .030) but not Cann-CBD (P = 1.000) or Cann+CBD (P = 1.000). any differences when comparing Cann+CBD with placebo or Cann+CBD with Cann-CBD. Subjective Music Ratings To aid interpretation of these findings (which may have been driven by drug effects on music, scrambled sound, or both), we Subjective music ratings are shown in Figure  3. For Want to Listen to Music, we found a drug by time interaction (F = 5.256, extracted parameter estimates from each of the clusters iden- 4,60 tified in this drug effect (Table 2, bottom panel) for separate P = .001, η = 0.259) and main effects of drug (F = 5.664, P = .008, p 2,30 2 2 η = 0.274) and time (F = 6.300, P = .012, η = 0.296). Scores contrasts of music>rest and scrambled>rest. ANOVA revealed p 1,22 p an interaction between drug (placebo, Cann-CBD) and contrast increased from Pre-Drug to Post-Drug following Cann+CBD (P < .001) and Cann-CBD (P = .002) but not placebo (P = 1.000). (music>rest, scrambled>rest) (F = 37.851, P < .001, η = 0.716). 1,15 p This interaction indicated that relative to placebo, Cann-CBD Scores then decreased from Post-Drug to Post-Scan on Cann+CBD (P = .028), but these tests did not reach significance for Cann- decreased parameter estimates for music>rest (P = .009, mean difference -0.195, standard error 0.065). However, there was no CBD (P = .553) or placebo (P = .199). However, analysis of all three drug conditions suggested that the decrease in Want to Listen evidence for drug effects on scrambled>rest (P = .130, mean dif- ference 0.103, standard error 0.064). There were no other findings to Music from Post-Drug to Post-Scan was equivalent across the 3 sessions (drug by time interaction: F = 1.130, P = .321, involving drug (drug by contrast by region interaction: F = 0.687, 4,60 1,21 2 2 2 P = .604, η = 0.044; drug by region interaction: F = 0.919, P = .459, η = 0.070; main effect of drug: F = 9.158, P = .001, η = 0.379, p 2,30 p p 4,60 2 2 main effect of time: F = 7.164, P = .017, η = 0.323). Moreover, η = 0.058; main effect of drug: F = 0.585, P = .456, η = 0.038). p 1,15 p 1,15 p This suggests that Cann-CBD dampened response to music to a when comparing the 2 types of cannabis alone, there were no differences between the effects of Cann+CBD and Cann-CBD on similar extent across each of these regions (right auditory cor - tex, left auditory cortex, right hippocampus/parahippocampal Want to Listen to Music across the 3 time points (drug by time interaction: F = 0.804, P = .457, η = 0.051; main effect of drug: gyrus, right ventral striatum, and right amygdala) while having 2,30 p negligible effects on response to scrambled sound. F = 3.590, P = .078, η = 0.193, main effect of time: F = 8.251, 1,15 p 2,30 P = .001, η = 0.355). Brain-Behavior Correlations For Enhanced Sound Perception, Pre-Drug scores were removed from analysis due to floor effects on each session. Mean Next, we sought to examine correlations between brain (music>scrambled, extracted from the 5 clusters shown in (SD) values were 0.25 (0.45) on placebo, 0.00 (0.00) on Cann+CBD, Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 27 Table 2. MNI Coordinates for the Contrasts Music>Scrambled (Main Effect of Task, Top Panel) and Scrambled>Music (Main Effect of Task, Middle Panel) across All Sessions. The bottom panel shows brain regions in which participants’ response to music>scrambled was dampened follow- ing cannabis without CBD compared with placebo; +: additional peak within cluster. All P values are thresholded at P < .05 (FDR-corrected for multiple comparisons) x y Z mm P Main effect of task (music>scrambled) L Caudate -12 6 6 540 4.45 .006 L Amygdala -15 -3 -15 486 4.32 .006 L Hippocampus -18 -12 -18 + 3.59 .027 R Caudate/thalamus 9 3 6 594 3.91 .014 R Pallidum 15 -3 -6 54 3.33 .031 L Cingulate gyrus -6 -15 42 27 3.16 .035 R Amygdala 18 -3 -18 54 2.99 .040 Main effect of task (scrambled>music) R Planum temporale 60 -12 3 3834 6.61 <.001 R Planum temporale 54 -24 6 + 6.54 <.001 L Planum temporale -48 -33 9 2511 6.16 <.001 L Heschls gyrus -42 -24 3 + 5.26 <.001 Drug effect (placebo>cannabis without CBD) R Superior temporal gyrus 51 -27 6 2484 4.51 .005 R Planum temporale 60 -12 3 + 3.55 .016 R Planum temporale/heschls gyrus 42 -18 0 + 3.15 .026 L Planum temporale -42 -33 9 972 4.04 .008 R Hippocampus/parahippocampal gyrus 33 -18 -24 81 3.23 .025 R Amygdala 27 3 -27 27 3.19 .025 R Ventral striatum 15 15 -12 54 2.90 .033 Figure 4. Cannabis without cannabidiol (CBD) dampened brain response to music across several regions sensitive to music-evoked reward and emotion. (A) Bilateral auditory cortex activation clusters visualized on the cortical surface of a standard template (MNI152). (B) A ventral view of the same template showing right-hemi- sphere amygdala and hippocampal clusters. (C) Axial slice views of the same contrast showing amygdala, hippocampal, ventral striatal (top row), and auditory cortex (bottom row) activation clusters. All activation maps thresholded at < .05 (FDR corr P ected for multiple comparisons). anterior; L, A, left hemisphere; P posterior; R, , right hemisphere. the bottom panel of Table 2) and behavior (pleasure ratings all sessions to minimize type I and type II error using mixed for music>scrambled, Post-Drug Want to Listen to Music and effects models, resulting in a total of 15 correlations. One cor - Enhanced Sound Perception). Data were combined across relation reached statistical significance. This showed a positive Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 28 | International Journal of Neuropsychopharmacology, 2018 relationship between pleasure ratings and response to music region during music relative scrambled sound. Next, we exam- in right ventral striatum (F= 11.447, P = .002; Figure 5). The ined drug effects using t contrasts. Compared with Cann-CBD, 1,34 same relationship was found using a Pearson correlation ana- greater functional connectivity occurred on Cann+CBD between lysis across all scans (r = 0.463, P = .001). This correlation did not right ventral striatum and bilateral auditory cortex (Tabl ; e  3 contain any outlying values (all data points were <3 times the Figure 6). We also conducted a PPI analysis using an auditory cor - interquartile range). However, there was still evidence for a cor - tex seed (right superior temporal gyrus 51, -27, 6). However, this relation after excluding the 2 data points showing the highest PPI analysis did not identify any regions that showed increases and lowest right ventral striatum response to music (mixed in functional connectivity with the seed region. effects model: F = 4.438, P = .041; Pearson correlation analysis 1,41 r = 0.318, P = .032). Possible Confounding Cardiovascular Drug Effects Functional Connectivity We conducted correlations between all Post-Drug cardiovascu- Previous research has shown that the rewarding experience of lar measures (heart rate, systolic and diastolic blood pressure) music is predicted by increased functional connectivity between and the 9 clusters showing evidence of drug effects (see Tables right ventral striatum and auditory cortex (Salimpoor et al., 2013; 2 and 3) across all sessions (total 27 correlations). We found no Zatorre and Salimpoor 2013Martínez-Molina et  ; al., 2016). To test evidence for any association between cardiovascular and fMRI this, we conducted PPI analyses. These analyses were conducted data (all P > .05). posthoc, informed by our findings that Cann-CBD blunted partic- ipants’ response to music in right ventral striatum and auditory cortex. Within-subjects ANOVA revealed that across all sessions, Cannabis Use We also explored correlations between levels of cannabis of use the right ventral striatum region (15, 15, -12) identified in our ana- lysis showed a robust increase in functional connectivity with and our main findings. These were conducted between (1) years of cannabis use, (2) days of cannabis use per month and the 9 bilateral auditory cortex (and to a lesser extent, right caudate) during music relative to scrambled sound (Table ).  F 3 or comple- clusters showing evidence of drug effects (Tables 2 and ),3 Want to Listen to Music (Post-Drug), Enhanced Sound Perception (Post- tion, we conducted the reverse contrast. However, we found no evidence for any regions showing reduced connectivity with this Drug), and pleasure rating scores. Of the 24 correlations, we found Figure 5. Correlation between brain and behavior. (A) Axial slice of right ventral striatal region of interest, identified from voxelwise analysis. (B) Sagittal slice of the same region. (C) Across all scans, activation in right ventral striatum for the contrast music>scrambled correlated positively with pleasure ratings. Table 3. Functional Connectivity Analysis. MNI coordinates showing increased functional connectivity with right ventral striatum for music>scrambled across all sessions (main effect, top panel). Functional connectivity between right ventral striatum and auditory cortex increased on cannabis with CBD compared with cannabis without CBD (drug effect, bottom panel); +Additional peak within cluster. All P values are thresholded at < P .05 (False Discovery Rate-corrected for multiple comparisons) X y z mm Z P Main effect R Planum temporale 60 -12 6 5319 6.43 <.001 R Heschls gyrus/planum polare 48 -12 0 + 6.42 <.001 R Planum temporale 48 -27 9 + 6.26 <.001 L Heschls gyrus -42 -24 12 3429 6.31 <.001 L Planum temporale -42 -30 6 + 6.27 <.001 L Planum temporale -33 -33 15 + 5.48 <.001 R Caudate 9 15 9 54 2.42 .037 Drug effect (cannabis with CBD>cannabis without CBD) R Heschls gyrus 42 -18 9 1620 4.63 .003 L Hippocampus -30 -18 -21 81 3.64 .009 L Heschls gyrus -36 -27 9 54 3.08 .030 L Heschls gyrus -45 -24 15 27 3.05 .031 Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 29 Figure 6. Functional connectivity analysis. (A) Seed region in right ventral striatum. (B) This seed region showed increased task-related functional connectivity with bilateral auditory cortex following cannabis with cannabidiol (CBD) compared with cannabis without CBD (C). Axial slices depicting the same data in bilateral auditory cortex and additional left hippocampal cluster. All activation maps visualized on MNI152 and thresholded at < .05 (FDR corr P ected for multiple comparisons). L, left hemisphere; R, right hemisphere. no evidence for any associations (all > P .05) apart from a trend 2006; Salimpoor et al., 2011T ; rost et al., 2012) and, in common negative correlation between years of cannabis use and functional with THC, may increase striatal dopamine release (Salimpoor et connectivity between right ventral striatum and left hippocampus al., 2011; Bossong et al., 2015). Moreover, observational data sug- (F = 4.984, P = .030). However, this did not reach significance at a gests that cannabis is frequently used in the context of music 1,48 Bonferroni-corrected threshold ( =α 0.0021). Moreover, the effect of and may enhance its effects (Tart 1970Gr ; een et al., 2003; Lim et drug remained significant in this model (F = 7.455, P = .002). al., 2008; Van Havere et al., 2011; Palamar et al., 2015). 1,48 One possible explanation for our findings is that THC inter - Order Effects fered with the endocannabinoid system, which plays a critical Because the same music and scrambled sound excerpts were role in reward processing (Parsons and Hurd 2015). For example, presented across each of the 3 sessions, we investigated pos- acute THC may deplete the CB1R ligand anandamide (Thieme sible order effects. For all fMRI results showing drug effects, the et al., 2014), which increases consummatory response to reward effect of drug remained significant, and there was no evidence in the nucleus accumbens shell (Mahler et al., 2007). Disruption for an effect of session order (all > P .05). There was no evidence of the endocannabinoid system could explain why neural for effects of drug or session order for pleasure rating scores response to reward was previously dampened by 7-day admin- (all P > .05). Analysis of Want to Listen to Music (Post-Drug) and istration of a CB1R antagonist (Horder et al., 2010) as well as a Enhanced Sound Perception (Post-Drug) scores showed effects single dose of the partial CB1R agonist THC (van Hell et al., 2012). of drug (both P< .001) but not session (both P > .05). It should also be noted that our findings of dampened response to music occurred in the context of increased wanting to listen to music. These findings are broadly consistent with previous Discussion findings that THC may have dissociable effects on anticipatory To our knowledge, this is the first controlled experiment inves- (“wanting”) and consummatory (“liking”) components of reward tigating the interactive effects of cannabis and music. Cannabis (van Hell et  al., 2012; Jansma et  al., 2013), although our task dampened response to music in several regions implicated in lacked a neural index of reward anticipation. music-evoked reward and emotion (Koelsch 2014): bilateral audi- Cannabis with CBD did not differ from placebo on any fMRI tory cortex, right amygdala, right hippocampus/parahippocam- measures. Furthermore, it resulted in greater task-related pal gyrus, and right ventral striatum. Across all scans we found functional connectivity between ventral striatum and audi- a positive correlation between response to music in this ventral tory cortex compared with cannabis without CBD. These find- striatal region and the pleasure of listening to the same sound ings suggest that CBD was able to offset some effects of THC, clips, consistent with several studies implicating the ventral stri-consistent with previous research (Curran et al., 2016 Englund ; atum in musical pleasure (Blood and Zatorre 2001 Koelsc ; h et al., et al., 2017) and evidence that THC and CBD can have opposite 2006; Salimpoor et al., 2011T ; rost et al., 2012). The same ventral neural effects (Bhattacharyya et al., 2010 Batalla et  ; al., 2014). For striatal region showed increased task-related functional con- example, activation in right superior temporal gyrus (a region nectivity with bilateral auditory cortex, an effect that has pre-identified in our study) during word listening relative to rest was viously been shown to predict musical reward value (Salimpoor previously found to be decreased by THC but increased by CBD et al., 2013; Zatorre and Salimpoor 2013Martínez-Molina et al., ; (Winton-Brown et  al., 2011). Moreover, CBD may increase con- 2016). centrations of anandamide (Bisogno et  al., 2001Le ; weke et  al., These findings were contrary to our prediction that cannabis 2012). We found some evidence that CBD interacted with THC would increase the rewarding effects of music, which can acti- on additional measures. Taken together, CBD appeared to par - vate and increase connectivity within mesolimbic brain regions tially offset some negative effects of THC (increase in diastolic (Blood and Zatorre 2001Menon and Le ; vitin 2005; Koelsch et al., blood pressure, decreased response to music) while preserving Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 30 | International Journal of Neuropsychopharmacology, 2018 or potentiating desirable ones (enhanced sound perception, Acknowledgments functional connectivity with ventral striatum during musical These data were presented in preliminary form at the CINP listening). World Congress in Seoul, South Korea, the BAP Summer Meeting In terms of clinical implications, the effects of acute can- in Brighton, UK, and the ECNP Workshop for Junior Scientists nabis administration here are similar to previous findings in in Nice, France. We are grateful to Marty Sereno, Joseph Devlin, people with depression, who also show a blunted response to and the Birkbeck-UCL Centre for Neuroimaging team for their music in ventral striatum as well as medial orbitofrontal cortex assistance. (Osuch et al., 2009). In this respect, acute cannabis administra- tion may transiently mimic the diminished response to reward characteristic of some mental health disorders. The impact of Statement of Interest chronic cannabis administration remains unclear. However, a H.V.C. is a member of UK MRC boards and Drug Science. D.J.N. is 4-year prospective study found that increased cannabis use an advisor to the British National Formulary, MRC, GMC, was associated with subsequent reductions in ventral striatal Department of Health; President of European Brain Council; Past response to reward anticipation (Martz et  al., 2016). It there- President of British Neuroscience Association and European fore is possible that effects of cannabis on reward processing College of Neuropsychopharmacology, Chair of Drug Science may contribute to an increased risk of developing depression (UK); Member of International Centre for Science in Drug Policy; (Zhang et al., 2013; Lev-Ran et al., 2014) as well as other disor - advisor to Swedish government on drug, alcohol, and tobacco ders characterized by reward dysfunction such as addiction and research; editor of the Journal of Psychopharmacology; mem- psychosis (Radua et al., 2015; Luijten et al., 2017). Moreover, our ber of advisory boards of Lundbeck, MSD, Nalpharm, Orexigen, findings support the potential utility of CBD in reducing can- Shire, MSD; has received speaking honoraria (in addition to nabis harms while maintaining the positive effects users seek above) from BMS/Otsuka, GSK, Lilly, Janssen, Servier, AZ, and (Englund et al., 2017). Pfizer; is a member of the Lundbeck International Neuroscience Strengths of this study include its controlled experimen- Foundation; has received grants or clinical trial payments from tal design, comparison of cannabis with and without CBD (but P1vital, MRC, NHS, Lundbeck, RB; has share options in P1vital; matched for THC), a music task previously validated using fMRI has been an expert witness in a number of legal cases relat- (Menon and Levitin 2005), and regions of interest informed by ing to psychotropic drugs; and has edited/written 27 books, meta-analysis (Koelsch, 2014). Our sample size was equivalent some purchased by pharma companies. C.J.A.M. has consulted or larger than previous studies with comparable designs (van for Janssen and GlaxoSmithKline and received compensation. Hell et al., 2012; Jansma et al., 2013) and neuroimaging music M.B.W.  is employed by Imanova Ltd., a private company that studies of music (mean n= 14.5 across 22 studies (Koelsch performs contract research work for the pharmaceutical indus- 2014)). We used a fixed set of classical music excerpts, com- try. The other authors declare no potential conflicts of interest. mensurate with previous use of this task (Menon and Levitin, 2005) and many other studies (Koelsch, 2014). Advantages of this approach include the absence of lyrics (which would influ- References ence neural response due to speech) and ease of comparison Batalla A, Crippa J, Busatto G, Guimaraes F, Zuardi A, Valverde O, with existing data. Although participants rated classical sound Atakan Z, McGuire P, Bhattacharyya S, Martin-Santos R (2014) clips as highly pleasant (~7.5 of 10), results may have differed if Neuroimaging studies of acute effects of THC and CBD in preferred music was preselected by participants (Osuch et al., humans and animals: a systematic review. Curr Pharm Des 2009). Drug order was not completely balanced in this study, 20:2168–2185. and the same sound clips were presented on each of the 3 ses- Beck AT, Steer RA, Brown GK (1996) Beck depression inventory-II. sions. However, we found no evidence that session order influ- San Antonio, TX: The Psychological Corporation. enced our results. We screened for personal/family history of Berridge KC, Kringelbach ML (2015) Pleasure systems in the psychosis and current treatment for a psychiatric disorder, brain. Neuron 86:646–664. but not for lifetime history of other mental health problems. Bhattacharyya S, Morrison PD, Fusar-Poli P, Martin-Santos R, Additionally, our sample were cannabis users, which may have Borgwardt S, Winton-Brown T, Nosarti C, MO’Carroll C, Seal M, prolonged cannabinoid clearance between sessions. However, Allen P (2010) Opposite effects of Δ-9-tetrahydrocannabinol we found minimal evidence for associations between cannabis and cannabidiol on human brain function and psychopath- use and fMRI findings, and their response to cannabinoids may ology. Neuropsychopharmacology 35:764–774. be more representative of typical use than healthy volunteers Bisogno T, Hanuš L, De Petrocellis L, Tchilibon S, Ponde DE, who never use cannabis. Brandi I, Moriello AS, Davis JB, Mechoulam R, Di Marzo V In conclusion, cannabis dampened the effects of music in (2001) Molecular targets for cannabidiol and its synthetic bilateral auditory cortex, right hippocampus/parahippocam- analogues: effect on vanilloid VR1 receptors and on the cel- pal gyrus, right amygdala, and right ventral striatum. During lular uptake and enzymatic hydrolysis of anandamide. Br J musical listening, this ventral striatal region correlated with Pharmacol 134:845–852. pleasure ratings and showed increased functional connectivity Blood AJ, Zatorre RJ (2001) Intensely pleasurable responses with auditory cortex. By contrast, cannabis containing canna- to music correlate with activity in brain regions impli- bidiol did not influence the effects of music in brain regions sen- cated in reward and emotion. Proc Natl Acad Sci U S A sitive to reward and emotion. 98:11818–11823. Booth M (2004) Cannabis: a history. New York: St. Martin’s Press. Funding Bossong MG, van Berckel BN, Boellaard R, Zuurman L, Schuit RC, T.P.F. was funded by the UK Medical Research Council and a Senior Windhorst AD, van Gerven JM, Ramsey NF, Lammertsma AA, Kahn RS (2009) Δ9-tetrahydrocannabinol induces dopamine Academic Fellowship from the Society for the Study of Addiction. This study was funded by Drug Science/Channel 4 television. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Freeman et al. | 31 release in the human striatum. Neuropsychopharmacology cannabinoid CB1 antagonist rimonabant in healthy volun- 34:759–766. teers. Int J Neuropsychopharmacol 13:1103–1113. Bossong MG, Mehta MA, van Berckel BN, Howes OD, Kahn Jansma J, Van Hell H, Vanderschuren L, Bossong M, Jager G, Kahn RS, Stokes PR (2015) Further human evidence for striatal R, Ramsey N (2013) THC reduces the anticipatory nucleus dopamine release induced by administration of∆ 9-tetra- accumbens response to reward in subjects with a nicotine hydrocannabinol (THC): selectivity to limbic striatum. addiction. Transl Psychiatry 3:e234. Psychopharmacology (Berl) 232:2723–9. Koelsch S (2014) Brain correlates of music-evoked emotions. Nat Curran VH, Brignell C, Fletcher S, Middleton P, Henry J (2002) Rev Neurosci 15:170–180. Cognitive and subjective dose-response effects of acute oral Koelsch S, Fritz T, Müller K, Friederici AD (2006) Investigating emo- Δ9-tetrahydrocannabinol (THC) in infrequent cannabis users. tion with music: an fMRI study. Hum Brain Mapp 27:239–250. Psychopharmacology (Berl) 164:61–70. Lacadie CM, Fulbright RK, Rajeevan N, Constable RT, Papademetris Curran HV, Freeman TP, Mokrysz C, Lewis DA, Morgan CJ, Parsons X (2008) More accurate Talairach coordinates for neuroimag- LH (2016) Keep off the grass? Cannabis, cognition and addic- ing using non-linear registration. Neuroimage 42:717–725. tion. Nat Rev Neurosci 17:293–306. Lawn W, Freeman TP, Pope RA, Joye A, Harvey L, Hindocha C, Di Forti M, Marconi A, Carra E, Fraietta S, Trotta A, Bonomo Mokrysz C, Moss A, Wall MB, Bloomfield MA (2016) Acute and M, Bianconi F, Gardner-Sood P, O’Connor J, Russo M (2015) chronic effects of cannabinoids on effort-related decision- Proportion of patients in south London with first-episode making and reward learning: an evaluation of the canna- psychosis attributable to use of high potency cannabis: a bis ‘amotivational’hypotheses. Psychopharmacology (Berl) case-control study. Lancet Psychiatry 2:233–238. 233:3537–3552. D’Souza DC, Perry E, MacDougall L, Ammerman Y, Cooper T, Wu Lev-Ran S, Roerecke M, Le Foll B, George T, McKenzie K, Rehm J Y-t, Braley G, Gueorguieva R, Krystal JH (2004) The psych- (2014) The association between cannabis use and depression: otomimetic effects of intravenous delta-9-tetrahydrocan- a systematic review and meta-analysis of longitudinal stud- nabinol in healthy individuals: implications for psychosis. ies. Psychol Med 44:797–810. Neuropsychopharmacology 29:1558–1572. Leweke F, Piomelli D, Pahlisch F, Muhl D, Gerth C, Hoyer C, ElSohly MA, Mehmedic Z, Foster S, Gon C, Chandra S, Church JC Klosterkötter J, Hellmich M, Koethe D (2012) Cannabidiol (2016) Changes in cannabis potency over the last 2 decades enhances anandamide signaling and alleviates psychotic (1995–2014): analysis of current data in the United States. Biol symptoms of schizophrenia. Transl Psychiatry 2:e94. Psychiatry 79:613–619. Lim MS, Lim MS, Hellard ME, Lim MS, Hellard ME, Hocking JS, Lim Englund A, Freeman TP, Murray RM, McGuire P (2017) Can we MS, Hellard ME, Hocking JS, Aitken CK (2008) A cross-sectional make cannabis safer? Lancet Psychiatry 4:643–648. survey of young people attending a music festival: associa- Freeman TP, Morgan CJ, Vaughn-Jones J, Hussain N, Karimi K, tions between drug use and musical preference. Drug Alcohol Curran HV (2012) Cognitive and subjective effects of mephe- Rev 27:439–441. drone and factors influencing use of a ‘new legal high’. Luijten M, Schellekens AF, Kühn S, Machielse MW, Sescousse Addiction 107:792–800. G (2017) Disruption of reward processing in addiction: an Freeman TP, Morgan CJA, Hindocha C, Schafer GL, Das RK, Curran image-based meta-analysis of functional magnetic reson- HV (2014) Just say ‘know’: how do cannabinoid concentra- ance imaging studies. JAMA Psychiatry 74:387–398. tions influence users’ estimates of cannabis potency and the Mahler SV, Smith KS, Berridge KC (2007) Endocannabinoid amount they roll in joints? Addiction 109:1686–1694. hedonic hotspot for sensory pleasure: anandamide in Freeman TP, Winstock AR (2015) Examining the profile of high- nucleus accumbens shell enhances’ liking’of a sweet reward. potency cannabis and its association with severity of canna- Neuropsychopharmacology 32:2267. bis dependence. Psychol Med 45:3181–3189. Martínez-Molina N, Mas-Herrero E, Rodríguez-Fornells A, Gard DE, Gard MG, Kring AM, John OP (2006) Anticipatory and Zatorre RJ, Marco-Pallarés J (2016) Neural correlates of specific consummatory components of the experience of pleasure: a musical anhedonia. Proc Natl Acad Sci U S A 113:E7337–E7345. scale development study. J Res Personal 40:1086–1102. Martz ME, Trucco EM, Cope LM, Hardee JE, Jester JM, Zucker RA, Gossop M, Darke S, Griffiths P, Hando J, Powis B, Hall W, Strang Heitzeg MM (2016) Association of marijuana use with blunted J (1995) The Severity of Dependence Scale (SDS): psycho- nucleus accumbens response to reward anticipation. JAMA metric properties of the SDS in English and Australian sam- Psychiatry 73:838–844. ples of heroin, cocaine and amphetamine users. Addiction Menon V, Levitin DJ (2005) The rewards of music listening: 90:607–614. response and physiological connectivity of the mesolimbic Green B, Kavanagh D, Young R (2003) Being stoned: a review of system. Neuroimage 28:175–184. self-reported cannabis effects. Drug Alcohol Rev 22:453–460. Mokrysz C, Freeman T, Korkki S, Griffiths K, Curran H (2016) Are Hardwick S, King LA (2008) Home Office Cannabis Potency Study adolescents more vulnerable to the harmful effects of canna- 2008. Home Office: Home Office Scientific Development bis than adults? A placebo-controlled study in human males. Branch United Kingdom. Translational Psychiatry 6:e961. Hindocha C, Freeman TP, Ferris JA, Lynskey MT, Winstock AR O’Reilly JX, Woolrich MW, Behrens TE, Smith SM, Johansen-Berg H (2016) No smoke without tobacco: a global overview of can- (2012) Tools of the trade: psychophysiological interactions and nabis and tobacco routes of administration and their associ- functional connectivity. Soc Cog Affective Neurosci 7:604–609. ation with intention to quit. Front Psychiatry 7:104. Osuch EA, Bluhm RL, Williamson PC, Theberge J, Densmore M, Hindocha C, Freeman TP, Schafer G, Gardener C, Das RK, Morgan CJ, Neufeld RW (2009) Brain activation to favorite music in healthy Curran HV (2015) Acute effects of delta-9-tetrahydrocannabi- controls and depressed patients. Neuroreport 20:1204–1208. nol, cannabidiol and their combination on facial emotion rec- Palamar JJ, Griffin-Tomas M, Ompad DC (2015) Illicit drug use ognition: a randomised, double-blind, placebo-controlled study among rave attendees in a nationally representative sample in cannabis users. Eur Neuropsychopharmacol 25:325–334. of US high school seniors. Drug Alcohol Depend 152:24–31. Horder J, Harmer CJ, Cowen PJ, McCabe C (2010) Reduced neural Parsons LH, Hurd YL (2015) Endocannabinoid signaling in reward response to reward following 7  days treatment with the and addiction. Nat Rev Neurosci 16:579–94. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018 32 | International Journal of Neuropsychopharmacology, 2018 Radua J, Schmidt A, Borgwardt S, Heinz A, Schlagenhauf F, plasma levels to examine potential influences of tetrahydro- McGuire P, Fusar-Poli P (2015) Ventral striatal activation dur - cannabinol application on the endocannabinoid system in ing reward processing in psychosis: a neurofunctional meta- humans. Drug Test Anal 6:17–23. analysis. JAMA Psychiatry 72:1243–1251. Trost W, Ethofer T, Zentner M, Vuilleumier P (2012) Mapping Room R (2014) Legalizing a market for cannabis for pleas- aesthetic musical emotions in the brain. Cereb Cortex ure: Colorado, Washington, Uruguay and beyond. Addiction 22:2769–2783. 109:345–351. Van Havere T, Vanderplasschen W, Lammertyn J, Broekaert E, Salimpoor VN, Benovoy M, Larcher K, Dagher A, Zatorre RJ (2011) Bellis M (2011) Drug use and nightlife: more than just dance Anatomically distinct dopamine release during anticipation and music. Subst Abuse Treat Prev Policy 6:18. experience of peak emotion to music. Nat Neurosci 14:257–262. van Hell HH, Jager G, Bossong MG, Brouwer A, Jansma JM, Zuurman Salimpoor VN, van den Bosch I, Kovacevic N, McIntosh AR, L, Van Gerven J, Kahn RS, Ramsey NF (2012) Involvement of Dagher A, Zatorre RJ (2013) Interactions between the nucleus the endocannabinoid system in reward processing in the accumbens and auditory cortices predict music reward value. human brain. Psychopharmacology 219:981–990. Science 340:216–219. Winton-Brown TT, Allen P, Bhattacharrya S, Borgwardt SJ, Fusar- Schafer G, Feilding A, Morgan CJ, Agathangelou M, Freeman Poli P, Crippa JA, Seal ML, Martin-Santos R, Ffytche D, Zuardi TP, Curran HV (2012) Investigating the interaction between AW (2011) Modulation of auditory and visual processing by schizotypy, divergent thinking and cannabis use. Conscious delta-9-tetrahydrocannabinol and cannabidiol: an FMRI Cogn 21:292–298. study. Neuropsychopharmacology 36:1340–1348. Sobell LC, Sobell MB (1992) Timeline follow-back measuring Zatorre RJ, Salimpoor VN (2013) From perception to pleasure: alcohol consumption. New York: Springer, pp41–72. music and its neural substrates. Proc Natl Acad Sci U S A Tart CT (1970) Marijuana intoxication: common experiences. 110:10430–10437. Nature 226:701–704. Zhang WN, Chang SH, Guo LY, Zhang KL, Wang J (2013) The neural Thieme U, Schelling G, Hauer D, Greif R, Dame T, Laubender correlates of reward-related processing in major depressive RP, Bernhard W, Thieme D, Campolongo P, Theiler L (2014) disorder: a meta-analysis of functional magnetic resonance Quantification of anandamide and 2-arachidonoylglycerol imaging studies. J Affect Disord 151:531–539. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/1/21/4102982 by Ed 'DeepDyve' Gillespie user on 16 March 2018

Journal

International Journal of NeuropsychopharmacologyOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off