Graduate School of Psychology
N-acetylcysteine and Smoking Cessation
Date : 31/08/2014
Student
Name : Rosa H. Mulder
Student ID number : 5664098
Address : Albert Cuypstraat 86 III-v Postal code and residence : 1072 CX Amsterdam Telephone number : 0640655705
Email address : rosa.mulder@hotmail.com
Supervisors
Supervisor : Prof. dr. Reinout Wiers Daily supervisor : Mieke Schulte
Specialisation : Brain and Cognition Second assessor : Dr. Kiki Nikolaou
Research center : University of Amsterdam / Academic Medical Center
N-acetylcysteine and Smoking Cessation ABSTRACT
While one in five deaths is due to the consequences of smoking, 96% of people who try to quit relapse within a year. Craving and impulsivity are important factors in relapse, and both have been related to an abnormal glutamate (Glu) homeostasis in the nucleus accumbens (Nacc) and dorsal anterior cingulate cortex (dACC). Furthermore, aberrant Glu signalling seems to decrease prefrontal control on striatal signalling. N-acetylcysteine (NAC) is a new candidate in the
treatment of addiction, as it appears to normalize Glu levels in the brain. In the current study, 48 smokers and 17 nonsmokers were compared on Glu levels in the brain, functional connectivity between the Nacc and dACC during resting state (rs-FC), brain activity during cue reactivity (CR), and impulsivity. Secondly, the effect of two weeks of 2400 mg NAC per day on these measures, as well as craving and smoking cessation, was evaluated in a two-week, double-blind randomized controlled trial, with 19 smokers in the NAC group and 20 smokers in the placebo group. Glu levels were measured with proton magnetic resonance spectroscopy, brain activity during CR and resting state with functional magnetic resonance imaging, impulsivity with the Delay Discounting Task (DDT) and Stop Signal Task (SST), and craving with the Questionnaire of Smoking Urges. Results showed that smokers had higher levels of Glu in the dACC than nonsmokers, a trend towards lower rs-FC, and more activity in the visual cortex during CR. No differences were found on the DDT or SST. NAC did not seem to decrease the differences that were found between smokers and nonsmokers, nor did it help in reducing craving or quitting smoking. Also, results of the DDT indicated that the NAC group became more impulsive while the placebo group became less impulsive over time. It is concluded that the current dosage of NAC does not decrease differences between smokers and nonsmokers, and does not aid in the cessation of smoking.
INTRODUCTION
People suffering from addiction often find themselves caught in a paradox: the destructive consequences of the disease are only facilitated by their own actions. With about 19% of people in the US smoking (King, et al 2011), tobacco is the most common addiction. Furthermore, one
in five deaths are due to the effects of smoking (Adhikari, et al 2009). Still, while people seem to
be aware of the negative effects of their addiction, they are often unable to quit. Ninety-six
The paradoxical nature of addiction is often explained within the framework of dual-process models. These models explain behaviour as the result of the balance between bottom-up motivational processes and top-down control processes (Strack and Deutsch 2004). The former are more impulsive and thought to rely upon structures such as the amygdala and nucleus accumbens (NAcc), whereas the latter are more dependent upon prefrontal areas (Bechara, et al
2006).Addiction is interpreted as an imbalance between the two, with the bottom-up urges having the upper hand over the controlling processes (Wiers, et al 2013). This seems to leave the
patient oversensitive to substance-related cues and more likely to react towards them, which manifests itself as an attentional bias and approach bias, respectively (Field and Cox 2008; Field, et al 2008). Functional magnetic resonance imaging (fMRI) shows that brain areas related to
emotion, attention, visual processing and motoric action are over-reactive to substance related cues (Engelmann, et al 2012) , and impulsivity tasks show that addicts not only become more
impulsive towards the substance, but also in general (Petry 2001).
Classically, the neurobiological explanation of escalated bottom-up urges points toward the reward associated, dopaminergic (DA) signalling. Specifically, dopaminergic input from the ventral tegmental area (VTA) to the nucleus accumbens (Nacc; part of the ventral striatum) is thought to set an incentive response in motion, through the basal ganglia (Ikemoto 1999). More recent results however, have emphasised the importance of glutamatergic (Glu) modulation of the Nacc. It has been shown that repeated cocaine administration decreases extracellular Glu in the NAcc after repeated cocaine injection in rats (Baker, et al 2003) and that Glu mediates
cue-induced relapse (Cornish and Kalivas 2000; McFarland, et al 2003). In monkeys, repeated cocaine
administration afffected Glu metabolism in the putamen, a striatal area involved in automatized movements (Yin and Knowlton 2006). More specifically, the decrease in extracellular Glu seems to reduce the firing rate of projections from the prefrontal cortex (PFC) to the NAcc,
presumably reducing the ability of the PFC to control bottom-up urges (Kalivas 2009). In humans, Glu levels in the brain can be measured with proton magnetic resonance spectroscopy. The signal obtained by this method is a reflection of both extra- and intracellular Glu, with a larger contribution of the latter. From studies that employ MRS, it appears that frontal glutamatergic regulation is aberrant for humans as well. Glu in the dorsal anterior cingulate cortex (dACC) was higher in cocaine addicts than in healthy controls (Schmaal, et al
2012b), predictive of relapse after smoking (Mashhoon, et al 2011) and related to impulsivity in
healthy controls (Schmaal, et al 2012a) .The dACC has been shown to be activated when resisting
Nacc glutamine (Gln), a precursor and postcursor of Glu, was related to craving in detoxified alcoholics.
N-acetylcysteine (NAC) might offer new possibilities. NAC is sold as an over-the-counter mucolytic agent in drug stores. In rodents, NAC increased extracellular Glu levels in the NAcc, normalized excitatory characteristics of PFC projections to NAcc (Kalivas 2009), and prevented cocaine primed drug seeking (Baker, et al 2003). Pilot studies in humans so far are promising;
amongst others, NAC normalized dACC Glu levels of cocaine addicts (Schmaal, et al 2012b),
reduced marijuana use and craving (Gray, et al 2010), reduced the rewarding effect of smoking
(Schmaal, et al 2011), and decreased number of cigarettes smoked (Knackstedt, et al 2009).
However, research on NAC and addiction in humans is still in the pilot phase. Many of the human studies are open-label studies, involve few participants, and none of the employed significance levels have been corrected for multiple comparisons. Therefore, more converging evidence is needed to determine the usefulness of NAC in smoking cessation, and in drug treatment in general. Also, more information is needed on the specific modes of action of NAC in the human brain. First, while rodent studies ascribe a pivotal role to Glu in the effect of NAC on addictive symptoms, so far only one (Schmaal, et al 2012b) human study showed an effect of
NAC on Glu. Furthermore, it is unclear what the actual effect of NAC would be on the
functionality of the dACC and NAcc. Research indicates that the dACC has a controlling effect on urges (Hester and Garavan 2004) and Kalivas (2009) mentioned that prefrontal signalling to the NAcc might be disturbed in addiction. In humans, it was found that functional connectivity between the Nacc and dACC during resting state was lower for addicts than healthy controls (Motzkin, et al 2014). This suggests that a NAC induced normalization of Glu levels could restore
this signalling and increase the ability of the dACC to control urges. Last, while effects of NAC on reactivity to substance related cues are found in animals, no study to date attended the effect of NAC on brain activity while watching such cues.
Therefore, in order to examine the neurobiological effects of NAC we looked at its effects in smokers on Glu level in the dACC and striatum, functional connectivity (FC) between the dACC and Nacc, and brain activity while watching smoking related cues. To assess the utility of NAC in smoking cessation we examined the effects of NAC on impulsivity, craving and smoking. To quantify to what extend NAC induced changes would be meaningful, we first compared smokers and nonsmokers on Glu, FC, cue reactivity in the brain, and impulsivity.
Based on aforementioned research, it was expected that Glu would be higher in smokers, and FC would be lower. For the cue reactivity, we explored the whole brain, but specifically expected higher activity in the visual and motor cortex of smokers, as we suspected that such
activations give rise attentional and approach bias. Secondly, we expected that NAC would decrease the differences between smokers in Glu, FC, cue reactivity and impulsivity, as well as reduce craving and help quit smoking.
MATERIALS AND METHODS Design
Smokers and nonsmokers were contrasted in a two-group comparison on data obtained in session 1. Directly after session 1, smokers entered the randomized controlled trial (RCT). The trial lasted two weeks, in which the smokers took 2400 mg/day of either NAC or placebo. After this period, a second session took place in order to assess the effects of NAC.
Participants
Seventeen nonsmokers and 48 smokers participated in the research. All were male, between the ages of 18 and 55 and eligible for MRI. Nonsmokers were included if they had never regularly smoked. Smokers had to have the intention to quit smoking during the trial and smoke at least 15 cigarettes per day for the last six months. Smokers were excluded if they had severe stomach problems, both smokers and nonsmokers were excluded if they showed signs of current or past psychosis.
Smokers were randomly assigned to either the NAC or the placebo group, resulting in two groups with 24 participants. Five participants of the NAC group and four participants of the placebo group did not complete the trial; leaving 19 participants in the NAC, and 20 participants in the placebo group for the NAC analyses.
Procedures
Nonsmokers and smokers attended a session (session 1) with questionnaires, two impulsivity tasks and MRI. During scanning, a T1-weighted anatomical image was obtained, as well as MRS data, fMRI data during resting state, and fMRI data during a cue reactivity task.
Subsequently, the NAC and placebo group were instructed to take 2 capsules twice daily, amounting to 2400 mg of compound. This dosage is similar to the one administered by Schmaal
et al (2012b). During the two-week trial period, participants attempted to refrain from smoking.
After one week, participants returned for an intermediate session (session 1.5) to fill in
questionnaires regarding personality traits and obtain a refill of capsules. One week later, session 2 took place, which was similar to session 1. Additionally, carbon monoxide (CO) concentration
of the breath was measured with a calibrated Micro + Smokerlyzer (Bedfont Scientific Ltd., Rochester, UK) to verify self-reported abstinence. Abstinence was confirmed when CO concentration did not exceed 10 parts per million (ppm). Participants who had taken < 2/3 of pills in the last week or 2 weeks total were excluded from further analyses. Informed consent was obtained for all participants and the study was approved by Ethics Committee of University of Amsterdam.
Clinical Assessments
Smokers and nonsmokers were screened for psychopathology with the Dutch version of the M.I.N.I.-Plus International Neuropsychiatric interview (van Vliet and de Beurs 2007)to assess psychopathology as put forward by DSM-IV criteria. Trait impulsivity was measured with the Baratt Impulsiveness Scale (BIS-11; Patton and Stanford 1995), and with self-reported behavioral avoidance and approach with the Behavioral Inhibition System/Behavioral Activation System (BIS/BAS; Carver and White 1994). The BAS scale comprises three subscales, namely Drive (BAS-drive), Fun Seeking (BAS-fun) and Reward Responsiveness (BAS-reward). The Dutch version of the American National Adult Reading Test (NART), the NLV (Nederlandse Leestest voor Volwassenen; Schmand, et al 1991) was used as an estimate of intelligence.
For the smokers, the Timeline Follow-Back (TLFB; Sobell and Sobell 1992) method was employed to quantify the number of cigarettes smoked before and during the experiment. Level of nicotine dependence was assessed with the Fagerström Test for Nicotine Dependence (FTND; Heatherton, et al 1991) and presence of indicators of alcohol misuse was assessed with
the Alcohol Use Disorder Identification Test (AUDIT; Saunders, et al 1993). The Dutch version
of the Readiness to Change Questionnaire (RCQ-D; Defuentes-Merillas, et al 2002) was used to
measure motivation to quit smoking, while the Obsessive Compulsive Smoking Scale (OCSS; Hitsman, et al 2010) was used to measure compulsive smoking. Craving was assessed with Dutch
version of the brief Questionnaire of Smoking Urges (QSU-Brief; Littel, et al 2011). This
questionnaire has a two factor structure, with the factors being ‘desire and intention to smoke’ (QSU-1 subscale) and ‘anticipation of reduction of negative affect and withdrawal symptoms’ (QSU-2 subscale).
MRS Acquisition and Analyses
Neuroimaging data were obtained with a Philips Achieva 3.0 Tesla head-only MRI scanner (Phillips Healthcare, Best, The Netherlands), using an 32-channel SENSE receiver head coil. Three-dimensional T1 weighted-images were acquired in the saggital plane with a gradient echo
Figure 1 Proton magnetic resonance spectroscopy voxel placement in the anterior cingulate cortex (a), right putamen (b), and right basal ganglia (c).
sequence (TR= 8.2 ms; TE = 3.7 ms; 220 slices; voxel size 1x1x1 mm; FOV 240 x 188 x 120). To determine local Glu levels, a bilateral dACC voxel, and a right putamen or a BG MRS voxel (size 35 x 20 x 15 mm and 35 x 20 x 20 mm or 30 x 20 x 20, respectively) were superimposed on the T1-weighed images (see Figure 1) and a localized MEGA-PRESS J-difference sequence
(Mescher, et al 1998) TR = 2000 ms; TE = 73 ms; recycle time = 2,5 s; 2048 complex points;
bandwidth = 2 kHz) was performed. This sequence allows one to selectively invert metabolite peaks that are overlapping with peaks of interest. Selective inversions were achieved with editing pulses applied at 1.91 ppm (N-acetylaspartate and N-acetylglutamate peak; NAA + NAAG peak) on odd-numbered acquisitions. Radiofrequency power and shim settings were optimized with the standard Phillips spectroscopy preparation routine.
The obtained spectra from 4.4 to 1.8 ppm were analysed using LCModel (Linear Combination of Model spectra; Provencher 1993), in which metabolite levels are estimated through automatic curve fitting (see Figure S1). Glutamate levels were deducted from spectra obtained from the difference between odd and even scans (diff spectrum). Creatine and
phosphocreatine (Cre + PCr) was used as an internal reference. Since Cre +PCr is only
represented in the even spectrum, both metabolite complexes were inversely referenced to NAA + NAAG (Wadell et al 2011):
Referenced Glu = Glu/(NAA + NAAG) / ((NAA + NAAG)/(Cre + PCr)) |---| |---| diff spectrum even spectrum
Spectra of which the peaks of interest had a Cramér-Rao Lower Bound (CRLB) >20% were considered unreliable and excluded from further analysis.
Cue Reactivity Task
During the cue reactivity task the participants watched a total of 75 full-color pictures, of which 30 were smoking-related (cues), 30 depicted random objects (noncues), and 15 depicted animals (controls). Half of the cues as well as non-cues were active (e.g. a person smoking; a person writing), half were passive (e.g. a cigarette package on a table; pen and paper on a table). To keep participants attentive, they were instructed to press a button if a picture with an animal appeared. Imaging data obtained during these animal pictures were not used for analysis. The task started with a fixation cross, of which presentation time varied around a mean of 4 seconds to allow jittering. Each picture was presented for 4 s.
fMRI Acquisition and Analyses
A gradient-echo echo-planar (EPI) sequence sensitive to blood oxygenation level-dependent (BOLD) contrast (TR = 2000 ms; TE=27.63 ms; matrix size 80 x 80; voxel size 3x3x3 mm; 37 slices; slice gap 0.3 mm) was used to acquire 302 images during the cue reactivity task (cr-fMRI) and 200 images during resting state (rs-fMRI).
fMRI data were preprocessed using SPM8 (Statistical Parametric Mapping; Wellcome Trust Centre for Neuroimaging, London, UK) and Matlab 2012a (The MathWorks, Inc., Natick, MA, USA). For each subject, the images were reoriented, slice time corrected, realigned, and
unwarped. The T1-weighted scans were coregistered to the functional scans and segmented. Last, the fMRI scans were normalized to MNI space and smoothed with an 8 mm full width at half maximum Gaussian kernel. Scans were excluded if the subject had moved > 3 mm in any direction. For the cr-fMRI first-level fixed effects analysis, each subjects’ event onsets for the cues and noncues were convolved with the canonical hemodynamic response function. Six
movement regressors were added to the General Linear Model (GLM). Finally, cue > noncue contrasts were computed.
Due to inconsistent placement of the acquisition box during scanning, various parts of the brain were not included in the field of view (FOV), amongst which a substantial part of the occipital cortex. For an exploratory analysis in the occipital cortex, an analysis with a subdivision of participants of which the FOV included this area, was performed. A mask was generated to preclude display of activation outside this area was generated with the Talairach Deamon Lobes atlas (Lancaster, et al 2000) , within the WFU Pickatlas toolbox (Maldjian, et al 2003).
In the of the NAC trial assessment, region of interest (ROI) analyses were performed on areas in which activation differed between smokers and nonsmokers. ROIs were generated around the coordinates peak voxels of clusters of activation that appeared on the smokers and nonsmokers contrast. The 3 mm spheres were drawn using the MarsBaR toolbox
(http://marsbar.-sourceforge.net/).
Functional Connectivity Analyses
Preprocessed rs-fMRI data were analysed using REST (Resting-State fMRI Data Analysis Toolkit; Xiao-Wei et al., http://resting-fmri.sourceforge.net). Linear detrending and temporal bandpass filtering for .01 – .08 Hz was applied. Physiological nuisance was removed by regressing out global white matter and CSF temporal signals. Likewise, motion parameters were treated as nuisance variables and removed via regression. The dACC region of interest (ROI) was produced to mirror the dACC VOI used in the MRS procedure. To this end, a mask was generated with MarsBaR, defined as a 20 x 35 x 15 mm box, which was drawn on the MNI brain (x = 0, y = 7, z = 36) and tilted (pitch = 0.6) to mirror the dACC VOI used in MRS. The nucleus accumbens ROI mask consisted of bilateral 3 mm spheres with centres at x = ± 9.1, y = 8.6, z = -7.8 (see Figure 2).The locations of these spheres are MNI converted duplicates from the ones employed
by Motzkin et al. (2014) who found a decreased functional connectivity between the bilateral Nacc and dACC in people with substance use disorders, using these voxels as seeds. Averaged time courses of these ROIs were correlated using Pearson’s correlation analysis (see Figure S2). Finally, z-scores were obtained with Fisher’s Z transformation.
Impulsivity Tasks
Behavioral impulsivity was measured with two computerized tasks, which are thought to represent the two main behavioural components of impulsivity (Reynolds, et al 2006): impulsive
disinhibition and impulsive decision making. The first was assessed with the Stop Signal Task
(SST; Logan, et al 1984) , the second with the Delay Discounting Task (DDT;Kirby, et al 1999).
Stop Signal Task. The SST consisted of 252 trials in which a picture of an airplane represented a
Go signal. Forty-eight of these Go signals was followed by a Stop signal: an airplane with a cross placed on top. Participants were instructed to press either the right or left arrow, depending on the direction of the airplane, as fast and accurate as possible. If a Stop signal occurred,
participants were to inhibit their response. The Go trials were displayed for 1000 ms, preceded by a fixation cross for 500 ms. The time between the Go and the Stop signal (stop-signal delay; SSD) was automatically adjusted during the task so that the percentage of correct inhibited responses approximated 50. Stop signals occurred at random time points.
Typically, the stop signal reaction time (SSRT) is computed by subtracting the mean SSD from the mean reaction time on Go trials (Go-RTs). This approach is based on the assumption that the percentage of correctly inhibited responses is 50. However, this is not always the case. Importantly, strategically slowing the Go response to increase the chance of correct inhibition can result in more than 50% correct inhibitions and an artificially shorter the mean SSRT. This would make it appear that a participant is better in inhibiting their responses than they actually are (Leotti and Wager 2010). Distinctly long mean Go-RTs, co-occurring with correct inhibition rates >50% and comparatively low mean SSRTs had us suspect that several participants had applied this particular strategy. Therefore, we analysed each participants’ integrated SSRT (SSRTi) instead of mean SSRT, for which the mean SSD is subtracted from the Go-RTs at the percentile that corresponds to the percentage of correct inhibitions. This approach has been shown to result in data less influenced by strategic slowing (Boehler, et al 2012) .
Delay Discounting Task. The DDT consisted of 138 dilemmas in which the participant had to
(usually higher) reward at a later time point (Later). The test comprised of 6 blocks of different time delays (0, 7, 30, 90, 180, 365 days). Each block consisted of 23 trials with different Now rewards between € 0.00 – 10.50 and Later rewards always being € 10.00. Blocks were scrambled and trials were ordered randomly.
The point at which the participant equally preferred the Now and Later reward (point of indifference) was obtained for each time delay and plotted. A hyperbolic function is often applied to obtain the discounting rate. However, this led to extreme non-normality of the data as well as inhomogeneity of variances. Therefore, we instead calculated and analysed the area under the curve (AUC; Myerson, et al 2001). This method has the additional advantage that it is not based
on any assumptions regarding the algebraic shape of the data points. A smaller AUC corresponds to steeper discounting, which is indicative of more impulsive decision making.
Statistical Analyses
SPSS 20.0 (SPSS, Chicago, IL) was used for analyses of referenced Glu levels, functional connectivity z-scores, average BOLD activity in ROIs, behavioural data, and questionnaires. First, scores exceeding ± 3 standard deviations of the group mean were excluded. Smokers and nonsmokers were compared with independent t-test, or Mann-Whitney U tests if data was
non-normally distributed and transformation did not result in improvement. The NAC and placebo group comparisons were made with repeated measures ANOVAs (RM ANOVAs), with group as between factor and session as within. SPSS was also used for the computation of Spearman’s rho correlations (the data distribution of Glu deviated from normal), survival analysis with Kaplan-Meier, and dose-response analyses with linear regression. As in NAC and addiction is still in its infancy and even small differences between groups could be clinically relevant, no adjustments for multiple comparisons were made; the α-level was set at 0.05 for each test
SPM8 was used for second level analyses of whole brain BOLD activity. Independent two-sample t-tests were used to compare the two groups on the cue > noncue contrast. The
difference between session 1 and 2 of the NAC and placebo group contrasts were compared using a full factorial model, with group (NAC vsplacebo) as a between factor and session (1 vs 2) as within. As the nature of these tests was exploratory, significance level was set at uncorrected
RESULTS
Differences between Smokers and Nonsmokers Sample
A nonsmoker with alcohol and cannabis dependence (as measured by the MINI), and a
nonsmoker that portrayed hazardous drinking (as measured by the AUDIT) were excluded from all analyses (for results with these participants included, see Supplemental Information on Exclusion Criteria). For the remaining participants, exclusion took place for each outcome measure separately, in order to maximize statistical power. Information on group sizes per outcome measure can be found in Table S1.
Demographics and clinical characteristics can be found in Table 1. BIS-11 score was higher for smokers (t(55) = 2.04, p = 0.046), while the higher BAS-drive score for smokers did not reach
significance (t(55) = 1.79, p = 0.079). Scores on clinical subscales can be found in Table S2.
Group means of each outcome measure can be found in Table S3.
Neuroimaging
Glutamate levels. Since seven out of nine spectra of the putamen voxel had a CRLB >20%, the BG
voxel was employed for the remaining participants. However, as a quarter of the remaining data still did not meet the reliability criterion, all BG data was disregarded. The MRS data obtained
Table 1 Demographics and Clinical Characteristics of Smokers and Nonsmokers
Smokers Nonsmokers T/U (df) p
M (SD) M (SD) Age 35.08 (9.83) 39.93 (13.11) 267.500 (61) 0.135 Education (SOI) 4.20 (0.76) 4.53 (0.99) 277.500 (58) 0.176 IQ (NLV) 105.29 (8.25) 108.93 (8.57) -1.48 (61) 0.144 BIS-11 66.84 (11.91) 59.93 (10.20) 2.04 (55) 0.046 BIS 19.26 (3.97) 17.43 (3.48) 1.58 (55) 0.119 BAS-drive 11.97 (2.50) 10.71 (2.64) 1.79 (55) 0.079 BAS-fun 11.87 (2.05) 11.36 (1.78) 230.500 (55) 0.183 BAS-reward 16.54 (2.39) 17.29 (1.90) 292.500 (55) 0.873
Abbreviations: SOI, standaard onderwijs indeling; NLV, Nederlandse Leestest voor Volwassenen; BIS-11, Baratt Impulsiveness Scale; BIS, Behavioral Inhibition System; BAS, Behavioral Activation System; BAS-drive, BAS Drive subscale; BAS-fun, BAS Fun Seeking subscale; BAS-reward, BAS Reward Responsiveness subscale.
Figure 3 Referenced glutamate (Glu) in the dorsal Figure 4 Functional connectivity (FC) during anterior cingulate of smokers and nonsmokers (mean resting state between the anterior cingulate cortex ± SE). Glu was significantly higher for smokers than and nucleus accumbens of smokers and non-smokers for nonsmokers. (mean ± SE). The lower FC for smokers approached significance.
from the dACC voxel was much more reliable, and referenced levels of Glu in the dACC were found to be higher for smokers than for nonsmokers (t(55) = 2.42, p = 0.019; see Figure 3).
Functional connectivity. Fisher’s z transformed correlation coefficients between dACC and Nacc
BOLD activity during resting state were non-significantly higher for nonsmokers than for nonsmokers (t(59) = -1.81, p = 0.075; see Figure 4).
Cue reactivity. A two-sample t-test of smokers > nonsmokers for the cue > noncue showed
bilateral clusters of activation in the anterior calcarine gyri (BA 30; Table 3; Figure 5). As portrayed in Figure 5, estimation of contrasts showed that the effect in the left calcarine gyrus was due to smokers exhibiting more BOLD activation during cue than during noncue events, while nonsmokers exhibited less BOLD activation during cue than during noncue events. The effect in the right calcarine gyrus was due to smokers exhibiting equal BOLD activation during cue and noncue events, while nonsmokers again exhibited less BOLD activation during cue than during noncue events.
Since the field of view of 17 scans did not comprise the entire occipital lobe, the above mentioned analyses lack information on a substantial part of the visual cortex. To explore the possibility that clusters of activation in this area were missed, analyses were reran on the subset of scans in which the full occipital lobe was included. This revealed an extra cluster of activation for smokers > nonsmokers than was shown by the initial analyses (BA 17; Table 3; Figure 6).
Table 3 BOLD Activity in Smokers > Nonsmokers for the Cue > Noncue Contrast
Group Region L/R BA MNI Cluster
size (voxels) T x y z Smokers > Nonsmokers, total sample Anterior Calcarine Gyrus L 30 -12 -58 10 67 3.55 Anterior Calcarine Gyrus R 30 15 -58 7 103 3.41 Smokers > Nonsmokers, occipital region Inferior Occipital Gyrus L 17 -12 -97 -8 14 4.01
Abbreviations: L/R, left/right; BA, Brodmann Area; MNI Montreal Neurological Institute;
p < 0.001 (uncorrected), k ≥ 10 voxels.
For the sample in which the occipital lobe was included, only activations within the occipital lobe were visualized.
Figure 5 BOLD activity for the cue > noncue contrast is higher for smokers than for nonsmokers in the left and right calcarine gyrus at uncorrected p = 0.001, k = 10 voxels. As is shown in the graphs, this is due to smokers
having either higher (left calcarine gyrus) or equal (right calcarine gyrus) BOLD activity (mean ± SE) for the cue > noncue contrast, while nonsmokers have lower BOLD activity for cues than noncues.
Figure 6 BOLD activity for the cue > noncue contrast is higher for smokers than for nonsmokers in a cluster of voxels in the occipital cortex at uncorrected p = 0.001, k = 10 voxels. Analysis performed in subset of participants in
which the occipital cortex was included in the field of view. Only activations within the occipital lobe were visualized.
Impulsivity Tasks
No difference between smokers and nonsmokers was found on SSRTi. Likewise, the groups did not differ in AUC as obtained by the DDT.
Explorative Correlations
Explorative correlations between referenced dACC Glu and the other outcome measures were calculated for the smokers group, to gauge in what way the hypothesized NAC induced Glu decrease would affect the other outcome measures. A significant, negative correlation between referenced Glu and activity in a ROI drawn around the peak voxel in the left calcarine gyrus (ρ =
-0.34, p = 0.032; see Figure 7), as well as around the peak in the right calcarine gyrus (ρ = -0.33, p
Figure 7 Correlation in smokers between referenced Figure 8 Correlation in smokers between referenced
glutamate (Glu) and average BOLD activation glutamate (Glu) and average BOLD activation
associated with the cue > noncue contrast in the left associated with the cue > noncue contrast in the right calcarine region of interest . calcarine region of interest.
= 0.041; see Figure 8). No correlations between Glu and functional connectivity, SSRTi, or delay discounting were found (See Table S4).
Effect of N-acetylcysteine on Smokers Sample
For 11 of the 19 NAC participants and 14 of the 20 placebo participants we were able to do a pill count. One NAC and 2 placebo participants did not meet the inclusion criterion of ≥ 2/3 of the total required pill intake for the total of two weeks and/or for the second week of the trial. Hence these participants were excluded from further analyses. The remaining participants in the
Table 4 Demographics and Clinical Characteristics of the NAC and Placebo Group
NAC Placebo T/U (df) p
M (SD) M (SD)
Age 36.89 (9.98) 36.89 (9.1) 0.548 (36) 0.587
Education (SOI) 4.18 (0.73) 4.27 (0.83) 164.500 (34) 0.827
IQ (NLV) 105.83 (7.45) 105.28 (9.04) 0.494 (36) 0.624
Cigarettes per day 23.59 (6.07) 21.78 (6.55) 124.500 (35) 0.167 Cigarettes day before 25.93 (8.90) 22.24 (8.42) 110.000 (28) 1.000 Years smoking 18.88 (9.44) 18.28 (9.52) 0.265 (35) 0.793 FTND 6.00 (1.94) 5.78 (1.73) 0.429 (36) 0.670 AUDIT 6.33 (2.85) 7.72 (3.64) -0.916 (36) 0.366 RCQ-D 44.33 (4.79) 41.55 (3.73) 2.046 (36) 0.048 OCSS 19.67 (3.50) 16.50 (5.48) 2.135 (36) 0.040 BIS-11 62.44 (9.56) 76.61 (11.05) -.1773 (36) 0.085 BIS 18.72 (3.92) 19.44 (3.99) -0.943 (36) 0.352 BAS-drive 11.94 (2.78) 11.89 (2.30) -0.248 (36) 0.805 BAS-fun 11.06 (2.24) 12.67 (1.71) -2.593 (36) 0.014 BAS-reward 17.11 (2.78) 17.05 (2.15) -0.234 (36) 0.816
Abbreviations: SOI, standaard onderwijs indeling; NLV, Nederlandse Leestest voor Volwassenen; FTND, Fragerström Test for Nictone Dependence; AUDIT, Alcohol Use Disorder Identification; RCQ-D, Dutch version of the Readiness to Change Questionnaire; OCSS, Obsessive Compulsive Smoking Scale; Test BIS-11, Baratt Impulsiveness Scale; BIS, Behavioral Inhibition System; BAS, Behavioral Activation System; BAS-drive, BAS Drive subscale; BAS-fun, BAS Fun Seeking subscale; BAS-reward, BAS Reward Responsiveness subscale.
NAC group ingested an average of 2228 mg/day (SD: 429 mg/day), and the placebo group 2106 mg/day (SD: 237 mg/day).
The total number of participants per outcome measure and reasons for exclusion can be found in Table S5. Demographics and clinical characteristics of the NAC and placebo group are depicted in Table 4. The NAC group scored higher than de placebo group on the RCQ-D (t(36)
= 2.05, p = 0.048) and on the OCSS (t(36) = 2.14, p = 0.004). The placebo group scored
non-significantly higher on the BIS-11 (t(36) = -0.18, p = 0.085) and significantly higher on the
BAS-fun (t(36) = -2.59, p = 0.014). Scores on clinical subscales can be found in Table S6. Group
means of the outcome measures discussed below can be found in Table S7.
Neuroimaging
Glutamate levels. A RM ANOVA did not show a main effect of session on referenced Glu in the
dACC, nor an interaction between group x session. The data was non-normally distributed. Normality could not be obtained via transformation due to opposing skews; the data of the NAC group during session 1 was negatively skewed, while the data of the placebo group during this session was positively skewed.
Functional Connectivity. There was no main effect of session on functional connectivity between the
Nacc and dACC. Also, no interaction effect of group x session was found.
Cue reactivity. There was no group x session interaction effect on BOLD activation for the
whole-brain analysis. For the left calcarine ROI, there was a main effect of session (F(1,30) = 5.29, p=
0.029), as BOLD activation to cue > noncue decreased from session 1 to 2. No group x session interaction effect was found (Figure 9). Also, no main effect of session or group x session interaction effect was found for the right calcarine ROI.
Impulsivity Tasks
SSRTi data were square root transformed to obtain normality. There was no main effect of session, nor an interaction effect of group x session on SSRTi. For AUC as obtained with the DDT, no main effect of session was found. There was a significant group x session interaction (F(1,32) = 5.08, p = 0.031), as AUC decreased somewhat from session 1 to 2 in the NAC group,
while it increased in the placebo group (Figure 10). The AUC data of the placebo group was leptokurtotic for both sessions, and the data of session 2 barely met the criterion of homogeneity of variance.
Figure 9 Average BOLD activity (mean ± SE) in the Figure 10 Area under the curve (AUC; mean ± left calcarine ROI for the cue > noncue contrast. SE) resulting from the delay discounting task. An A main effect of session was shown, as BOLD activity of interaction effect was found, as AUC decreased both groups decreased over time. from session 1 to 2 for the NAC group, while it increased for the placebo group.
Craving and Smoking
QSU. There was a significant main effect of session (F(1,34) = 5.53, p = 0.027), as QSU score
decreased from session 1 to 2. A similar effect was found on the QSU-1 subscale (F(1,34) =
13.95, p = 0.001; see Figure 11), but not on the QSU-2 subscale. No group x session interaction
effect on the QSU or its subscales was found.
Smoking. There was main effect on number of cigarettes smoked per day (F(1,30) = 276.93, p <
0.001), as both groups significantly reduced cigarette use from before to during the trial (see Figure 12). No group x session interaction was found. The data was non-normally distributed, as
Figure 11 Scores (mean ± SE) on the ‘desire Figure 12 Cigarettes smoked per day (mean ± and intention to smoke’ subscale of the SE). There was a main effect of time, as smoking Questionnaire of Smoking Urges (QSU-1). decreased for both groups.
There was a main effect of session, as scores of both groups decreased from session 1 to 2.
it was positively skewed for all four groups. Also, Kaplan-Meier analyses showed no difference in survival function between the NAC and placebo group (Table S8; Figure S3).
Explorative dose-response regressions
Linear regressions were performed to explore dose-response effects. Participants that were initially excluded from analyses due to low pill intake, were included in the regression analyses to obtain a broader dosage spectrum, as well as larger group sizes. Some chariness on the
interpretation of these data is implied, as group sizes still were especially small. That said, a significant effect of dosage on dACC Glu levels was found within the NAC group (N = 11, β =
-0.69, p = 0.018), as well as on BOLD activity in the left calcarine (N = 11, β = -0.70, p = 0.016),
and right calcarine ROI (N = 11, β = -0.66, p = 0.029). A trend was found for SSRTi (N = 10, β
= -0.61, p = 0.060). No effects of dosage were found on functional connectivity, DDT score,
QSU, or cigarettes per day. See Figures 12-15 for slopes of the significant regressions. All regression scores can be found in Table S9.
Secondly, to examine the possibility that the dose-response effect is caused by compliance to the trial rather than the dosage itself, a dose-response regression was performed for the placebo group. No effect was found (N = 12, β = -0.03, p = 0.930). Regression scores can be
found in Table S10.
Figure 12 Effect of dosage of N-acetylcysteine Figure 13 Effect of dosage of N-acetylcysteine (NAC) on referenced glutamate (Glu) in smokers. (NAC) on average BOLD activity in the left A significant effect of dosage was found, as Glu calcarine ROI smokers. A significant effect
increased with a low dosage of NAC, while it of dosage was found, as activity increased with a low decreased with a high dosage. dosage of NAC, while it decreased with a high dosage.
Figure 14 Effect of dosage of N-acetylcysteine Figure 15 Effect of dosage of N-acetylcysteine (NAC) on average BOLD activity in the right on integrated stop signal reaction time (SSRTi) in calcarine ROI in smokers. A significant smokers. A trend towards a dosage effect was found, effect of dosage was found, as activity increased as SSRTi increased with a low dosage of NAC, while with a low dosage of NAC, while it decreased with it decreased with a high dosage
a high dosage.
DISCUSSION
In this study, smokers and nonsmokers were compared on glutamate levels in the putamen/BG and dACC, FC between the dACC and NAcc, brain activity during CR, and impulsivity.
Subsequently, the effect of NAC on smokers, on abovementioned measures, as well as craving and smoking cessation, was assessed in a RCT.
Results indicate that smokers have higher levels of Glu in the dACC than nonsmokers and a trend towards lower functional connectivity between the dACC and the NAcc;
corroborating our expectations. We were unable to obtain reliable data from the putamen/BG regarding glutamate levels. Against our expectation, activity in the motor cortex while watching cues was not higher for smokers than for nonsmokers. The activity in the visual cortex, on the other hand, was. Last, the anticipated increased impulsivity in smokers was not found.
Whereas NAC was predicted to decrease the differences between smokers and nonsmokers, this cannot be concluded from the results. Furthermore, NAC did not aid in reducing craving or quitting smoking. Results of the DDT indicated that the NAC group became móre impulsive while the placebo group became less impulsive over time.
In order to optimally interpret the implications of the study, the methods and results of the smokers and nonsmokers comparisons will be discussed first, after which the results of the NAC trial will be discussed.
Differences between smokers and nonsmokers. Unfortunately, we were unable to obtain a reliable Glu
signal in the BG area. While Glu MRS studies more often employ frontal voxels, ventral striatal Glu levels and BG Glu + glutamine (Glx) have been reported (Ernst and Chang 2008; Bauer, et al
2013, respectively). In the study of Bauer et al., the same CRLB < 20% exclusion criterion was used as in the current study. They also excluded nearly a fourth of the participants for the data of the Glu signal. However, the variance of the remaining data was vastly smaller than that of the BG Glu signal obtained here. Likewise, the variance of the Glx signal in the BG as reported by Ernst and Chang was relatively narrow. Therefore, if one were to decide to examine Glu
metabolism in the BG area with MRS, it might be wise to place the voxel in the ventral direction rather than dorsal, and assess the Glx rather than the Glu signal. Still, the inhomogeneous nature of the structure might not accommodate the collection of a reliable signal.
The difference in dACC Glu levels between smokers and nonsmokers is congruent with the difference in Glu between cocaine dependents and healthy controls, as found by Schmaal et al
(2012b) and with the Glu difference between alcoholics and controls as found by Lee et al (2007).
On the other hand, dACC Glu levels have also found to be decreased in addicts (e.g. Yang, et al
2009; Yücel, et al 2007). These opposing findings may partly be due to varying influences of
glutamine in the reported Glu signal (Thoma, et al 2011), as well as to the fact that the placement
of the different VOIs within the dACC only partially overlap and dACC is not functionally uniform (Rushworth, et al 2004). Studies in which Glu disparities between addicts and healthy
controls were found have also shown associations between Glu and impulsivity (Schmaal, et al
2012a), as well as relapse (Mashhoon, et al 2011). This affirms the idea, stemming from rodent
research (Kalivas 2009), that Glu plays a meaningful role in addiction.
On the other hand, the theorized relationship between Glu and rs-FC between the dACC and Nacc was not found. This could have to do with relatively small FC difference between smokers and nonsmokers. An explanation for this might be that the Nacc is a small structure, increasing the chance that the location of the employed ROI is does not coincide with the Nacc of all participants.However, the voxels around which the Nacc ROIs were drawn in the current study, mirrored the seed voxels in an FC study on substance abuse of Motzkin et al. (2014). In that study, activity in these seeds not only correlated with dACC activity, but also with extensive clusters of voxels that encompassed the ventral striatum. A more a more pronounced and relevant difference between smokers and nonsmokers in dACC-nACC FC might be expected during a cue reactivity task, however.
The increased activity in the anterior calcarine gyri during cue reactivity for smokers than for nonsmokers meets our expectation. The activity seems to be located in the retrosplenial
cortex, which is involved in the integration of visual information (Park and Chun 2009). Moreover, the secondary analyses in the full occipital cortex indicated that smokers also had more activation in the most posterior part of the cortex, V1. For activity in such downstream brain areas to occur, an upstream brain area must have recognized the cue as being significant and allocated activity in the occipital cortex. It is conceivable that such a process would be the neural basis of attentional bias towards substance related cues. Direct evidence on the matter is scarce however; so far, only Janes et al (2010) has related occipital activity during substance
related cues to attentional bias. More research is required to substantiate the relation between activity in the visual cortex and attentional bias. Finally, with regard to the location of the BOLD activity during cues, it should be mentioned that the field of view of the fMRI data was rather limited; causing areas such as the orbitofrontal cortex, hippocampal area, and supplemental motor area not to be fully included in the analyses.
Unlike Engelmann (2012), the activity in the visual cortex for cues was only found between groups, not within smokers. However, both groups, and especially the smokers group, revealed a pattern of activity during cues that closely resemble the default mode network (Buckner, et al 2008). This network is involved in self-reflection, and anticorrelated with areas
that are more outward or action oriented, such as the visual and motor cortex (Uddin, et al 2009).
It could be that activity associated with fast, automatic biases is inhibited by activity related to more pensive inwards activity, causing a difference in attention to only appear in group comparisons. This could also explain why the expected activity in the motor cortex was not shown. It might also explain the negative correlation found between calcarine BOLD activity and dACC Glu in smokers.
Results on the impulsivity tasks did not meet our expectations, as no differences were found between smokers and nonsmokers. In contrast, smokers did score higher than
nonsmokers on questionnaires of trait impulsivity. For the SST, we calculated the SSRTi instead of the mean SSRT in order to obtain data less influenced by participants who presumably strategically slowed down their response. While this did have the intended effect, the question remains whether the data of these participants justly represents their SSRT, as the strategy that they used in itself is still different from that used by the other participants. A possible reason for the lack of difference between the groups in DDT scores might be that smokers had a higher income than nonsmokers, which could have obscured differences in impulsivity (Green, et al
1996). Unfortunately, direct data on the matter is lacking, but there are indications that smokers generally participated in the study in order to quit smoking, while nonsmokers more often had financial motives.
In all, while we were not able to show differences in state impulsivity between smokers and nonsmokers, several differences were revealed regarding brain function and constituents, indicating that there is latitude for a Glu normalising agent to provide aid.
The effect of N-acetylcysteine on smokers. While the difference found in dACC Glu levels between
smokers and nonsmokers was in accordance with dACC Glu difference found by Schmaal et al
(2012b), the lack of effect of NAC on dACC Glu was not. An important aspect concerning this matter is that Schmaal et al measured Glu levels an hour after NAC ingestion, when NAC should
approach peak plasma concentration (Holdiness 1991). Participants in the current study did not receive specific instructions regarding the timing of pill ingestion other than that they should take two pills twice daily. It could therefore be possible that NAC has a phasic effect on Glu levels in the brain, but not a tonic one. Subsequently, given that the theoretically most proximal effect of NAC does not appear to have taken place, it is not surprising that there were no effects found on other brain measures, SSRT, craving, or smoking. The difference in change in DDT score over time was unexpected, on the other hand. Scores on trait impulsivity were higher for the placebo than for the NAC group, perhaps indicating that there was more room for improvement within the placebo group. Moreover, taking into account the other results, the level of significance, and the non-normality of the data, it is conceivable that this is a spurious result.
The dose-response regressions portray a different story than the aforementioned results. From these slopes, it appears that high dosages do have the intended effects, but low dosages have the exact opposite effects. While such a biphasic response curve is quite uncommon in neurobiological literature, it does fit the description of a hormetic response (Calabrese 2008). However, in hormesis, the decrease in Glu at high dosages would be explained as a neurotoxic effect, instead of the supposedly healthier state. Evidently, a fair amount of scepticism is implied, as these results are based on small sample sizes and heavily influenced by the extreme values of a single participant. However, it does seem remarkable that the courses of the different regressions are consistent with each other. Given that the tipping point of the slope seems to be at 2000 – 2400 mg/day, a somewhat bold conclusion might be that a NAC dosage over 2400 mg/day could have the intended effects.
Apart from dosage, there are other factors that could have influenced the outcomes. Inherent to the set-up of the study, there may have been effects of withdrawal on Glu (Baker, et al 2003) , as well as on the other outcome measures (e.g. Carter and Tiffany 2001; Field, et al
had quit smoking before session 1 took place. Also, direct effects of nicotine on Glu (Lambe, et al
2003) were unaccounted for in the current study, as well as possible phasic effects of NAC. One might aspire to investigate whether NAC can be beneficial in overcoming smoking addiction at a higher dosage. Mardikian (2007) found that an intake of 3600 mg/day for the duration of a month was well tolerated. If such a study were to take place, it might be advisable to not ask participants to quit smoking, as to exclusively investigate the effects of NAC on brain measures and possibly impulsivity. By prioritizing these measures, results will be less influenced by intervening variables. Importantly, in such a study, NAC intake should be more closely monitored. Furthermore, if one does intent to examine the effect of NAC on quitting, a prolonged duration of the trial would be recommended. It seems that in the current study, the difference between session 1 and 2 were thus large for smoking and craving, that it could have masked any effect of NAC.
Concluding, while there are indications of glutamatergic dysregulation in the brains of smokers, as well as diminished frontal control over the NAcc and increased cue reactivity, a NAC dosage of 2400 mg/day does not seem to restore these aspects and consequently does not aid smoking cessation. More research is required to determine if pharmacogenic normalization of glutamatergic homeostasis is obtainable and if it can be used in overcoming addiction.
REFERENCES
Adhikari B, Kahende J, Malarcher A, Pechacek T, Tong V (2009). Smoking-Attributable Mortality, Years of Potential Life Lost, and Productivity Losses-United States, 2000-2004 (Reprinted from MMWR, vol 57, pg 1226-1228, 2008). Jama-Journal of the American Medical Association 301: 593-594.
Baker DA, McFarland K, Lake RW, Shen H, Tang X, Toda S, Kalivas PW (2003).
Neuroadaptations in cystine-glutamate exchange underlie cocaine relapse. Nat Neurosci 6:
743-749.
Bauer J, Pedersen A, Scherbaum N, Bening J, Patschke J, Kugel H, Heindel W, Arolt V, Ohrmann P (2013). Craving in alcohol-dependent patients after detoxification is related to glutamatergic dysfunction in the nucleus accumbens and the anterior cingulate cortex.
Neuropsychopharmacology 38: 1401-1408.
Bechara A, Noel X, Crone EA (2006). Loss of willpower: abnormal neural mechanisms of impulse control and decision making in addiction. Handbook of implicit cognition and addiction
215-232.
Boehler CN, Appelbaum LG, Krebs RM, Hopf J, Woldorff MG (2012). The influence of different Stop-signal response time estimation procedures on behavior–behavior and brain– behavior correlations. Behav Brain Res 229: 123-130.
Brody AL, Mandelkern MA, Olmstead RE, Jou J, Tiongson E, Allen V, Scheibal D, London ED, Monterosso JR, Tiffany ST (2007). Neural substrates of resisting craving during cigarette cue exposure. Biol Psychiatry 62: 642-651.
Buckner RL, Andrews‐Hanna JR, Schacter DL (2008). The brain's default network. Ann N Y
Acad Sci 1124: 1-38.
Calabrese EJ (2008). Neuroscience and hormesis: overview and general findings. CRC Crit Rev Toxicol 38: 249-252.
Carter BL and Tiffany ST (2001). The cue-availability paradigm: the effects of cigarette availability on cue reactivity in smokers. Exp Clin Psychopharmacol 9: 183.
Carver CS and White TL (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: the BIS/BAS scales. J Pers Soc Psychol 67: 319.
Cornish JL and Kalivas PW (2000). Glutamate transmission in the nucleus accumbens mediates relapse in cocaine addiction. The Journal of neuroscience: the official journal of the Society for Neuroscience 20: RC89-RC89.
Defuentes-Merillas L, Dejong CA, Schippers GM (2002). Reliability and validity of the Dutch version of the Readiness to Change Questionnaire. Alcohol Alcohol 37: 93-99.
Engelmann JM, Versace F, Robinson JD, Minnix JA, Lam CY, Cui Y, Brown VL, Cinciripini PM (2012). Neural substrates of smoking cue reactivity: a meta-analysis of fMRI studies. Neuroimage 60: 252-262.
Ernst T and Chang L (2008). Adaptation of brain glutamate plus glutamine during abstinence from chronic methamphetamine use. Journal of Neuroimmune Pharmacology 3: 165-172.
Field M and Cox WM (2008). Attentional bias in addictive behaviors: a review of its development, causes, and consequences. Drug Alcohol Depend 97: 1-20.
Field M, Kiernan A, Eastwood B, Child R (2008). Rapid approach responses to alcohol cues in heavy drinkers. J Behav Ther Exp Psychiatry 39: 209-218.
Field M, Santarcangelo M, Sumnall H, Goudie A, Cole J (2006). Delay discounting and the behavioural economics of cigarette purchases in smokers: the effects of nicotine deprivation.
Psychopharmacology (Berl ) 186: 255-263.
Gray KM, Watson NL, Carpenter MJ, LaRowe SD (2010). N-acetylcysteine (NAC) in young marijuana users: an open-label pilot study. The American journal on addictions/American Academy of Psychiatrists in Alcoholism and Addictions 19: 187.
Green L, Myerson J, Lichtman D, Rosen S, Fry A (1996). Temporal discounting in choice between delayed rewards: the role of age and income. Psychol Aging 11: 79.
Heatherton TF, Kozlowski LT, Frecker RC, Fagerström K (1991). The Fagerström test for nicotine dependence: a revision of the Fagerström Tolerance Questionnaire. Br J Addict 86:
Hester R and Garavan H (2004). Executive dysfunction in cocaine addiction: evidence for discordant frontal, cingulate, and cerebellar activity. The Journal of Neuroscience 24: 11017-11022.
Hitsman B, Shen B, Cohen RA, Morissette SB, Drobes DJ, Spring B, Schneider K, Evans DE, Gulliver SB, Kamholz BW (2010). Measuring smoking-related preoccupation and compulsive drive: evaluation of the obsessive compulsive smoking scale. Psychopharmacology (Berl ) 211:
377-387.
Holdiness MR (1991). Clinical pharmacokinetics of N-acetylcysteine. Clin Pharmacokinet 20:
123-134.
Hughes JR, Keely J, Naud S (2004). Shape of the relapse curve and long‐term abstinence among untreated smokers. Addiction 99: 29-38.
Janes AC, Pizzagalli DA, Richardt S, de B Frederick B, Holmes AJ, Sousa J, Fava M, Evins AE, Kaufman MJ (2010). Neural substrates of attentional bias for smoking-related cues: an fMRI study. Neuropsychopharmacology 35: 2339-2345.
Kalivas PW (2009). The glutamate homeostasis hypothesis of addiction. Nature Reviews Neuroscience 10: 561-572.
King B, Dube S, Kaufmann R, Shaw L, Pechacek T (2011). Vital Signs: Current Cigarette Smoking Among Adults Aged greater than or equal to 18 Years- United States, 2005-2010. Morb Mortal Weekly Rep 60: .
Kirby KN, Petry NM, Bickel WK (1999). Heroin addicts have higher discount rates for delayed rewards than non-drug-using controls. J Exp Psychol : Gen 128: 78.
Knackstedt LA, LaRowe S, Mardikian P, Malcolm R, Upadhyaya H, Hedden S, Markou A, Kalivas PW (2009). The role of cystine-glutamate exchange in nicotine dependence in rats and humans. Biol Psychiatry 65: 841-845.
Lambe EK, Picciotto MR, Aghajanian GK (2003). Nicotine induces glutamate release from thalamocortical terminals in prefrontal cortex. Neuropsychopharmacology .
Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey L, Kochunov PV, Nickerson D, Mikiten SA, Fox PT (2000). Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp 10: 120-131.
Lee E, Jang DP, Kim JJ, An SK, Park S, Kim IY, Kim SI, Yoon KJ, Namkoong K (2007). Alteration of brain metabolites in young alcoholics without structural changes. Neuroreport 18:
1511-1514.
Leotti LA and Wager TD (2010). Motivational influences on response inhibition measures.
Journal of Experimental Psychology: Human Perception and Performance 36: 430.
Littel M, Franken IH, Muris P (2011). Psychometric properties of the brief Questionnaire on Smoking Urges (QSU-Brief) in a Dutch smoker population. Netherlands Journal of Psychology 1-20.
Logan GD, Cowan WB, Davis KA (1984). On the ability to inhibit simple and choice reaction time responses: a model and a method. Journal of Experimental Psychology: Human Perception and Performance 10: 276.
Maldjian JA, Laurienti PJ, Kraft RA, Burdette JH (2003). An automated method for
neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage 19:
1233-1239.
Mardikian PN, LaRowe SD, Hedden S, Kalivas PW, Malcolm RJ (2007). An open-label trial of N-acetylcysteine for the treatment of cocaine dependence: A pilot study. Prog
Neuro-Psychopharmacol Biol Psychiatry 31: 389-394.
Mashhoon Y, Janes AC, Jensen JE, Prescot AP, Pachas G, Renshaw PF, Fava M, Evins AE, Kaufman MJ (2011). Anterior cingulate proton spectroscopy glutamate levels differ as a function of smoking cessation outcome. Prog Neuro-Psychopharmacol Biol Psychiatry 35: 1709-1713.
McFarland K, Lapish CC, Kalivas PW (2003). Prefrontal glutamate release into the core of the nucleus accumbens mediates cocaine-induced reinstatement of drug-seeking behavior. The Journal of neuroscience 23: 3531-3537.
Mescher M, Merkle H, Kirsch J, Garwood M, Gruetter R (1998). Simultaneous in vivo spectral editing and water suppression. NMR Biomed 11: 266-272.
Motzkin JC, Baskin‐Sommers A, Newman JP, Kiehl KA, Koenigs M (2014). Neural correlates of substance abuse: Reduced functional connectivity between areas underlying reward and cognitive control. Hum Brain Mapp .
Myerson J, Green L, Warusawitharana M (2001). Area under the curve as a measure of discounting. J Exp Anal Behav 76: 235-243.
Park S and Chun MM (2009). Different roles of the parahippocampal place area (PPA) and retrosplenial cortex (RSC) in panoramic scene perception. Neuroimage 47: 1747-1756.
Patton JH and Stanford MS (1995). Factor structure of the Barratt impulsiveness scale. J Clin Psychol 51: 768-774.
Petry NM (2001). Delay discounting of money and alcohol in actively using alcoholics, currently abstinent alcoholics, and controls. Psychopharmacology (Berl ) 154: 243-250.
Provencher SW (1993). Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magnetic Resonance in Medicine 30: 672-679.
Reynolds B, Ortengren A, Richards JB, de Wit H (2006). Dimensions of impulsive behavior: Personality and behavioral measures. Personality and individual differences 40: 305-315.
Rushworth M, Walton M, Kennerley S, Bannerman D (2004). Action sets and decisions in the medial frontal cortex. Trends Cogn Sci (Regul Ed ) 8: 410-417.
Saunders JB, Aasland OG, Babor TF, de la Fuente, Juan R, Grant M (1993). Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption‐II. Addiction 88: 791-804.
Schmaal S, Berk L, Hulstijn KP, Cousijn J, Wiers RW, van den Brink, W (2011). Efficacy of N-acetylcysteine in the treatment of nicotine dependence: A double-blind placebo-controlled pilot study. Eur Addict Res 17: 211-216.
Schmaal L, Goudriaan AE, Meer J, Brink W, Veltman DJ (2012a). The association between cingulate cortex glutamate concentration and delay discounting is mediated by resting state functional connectivity. Brain and behavior 2: 553-562.
Schmaal L, Veltman DJ, Nederveen A, van den Brink W, Goudriaan AE (2012b).
N-acetylcysteine normalizes glutamate levels in cocaine-dependent patients: a randomized crossover magnetic resonance spectroscopy study. Neuropsychopharmacology 37: 2143-2152.
Schmand B, Bakker D, Saan R, Louman J (1991). The Dutch Reading Test for Adults: a measure of premorbid intelligence level. Tijdschr Gerontol Geriatr 22: 15-19.
Schoenmakers T, Wiers RW, Jones BT, Bruce G, Jansen A (2007). Attentional re‐training decreases attentional bias in heavy drinkers without generalization. Addiction 102: 399-405.
Sobell LC and Sobell MB (1992): Timeline follow-back. In: Anonymous Measuring alcohol consumption. Springer. pp 41-72.
Strack F and Deutsch R (2004). Reflective and impulsive determinants of social behavior.
Personality and social psychology review 8: 220-247.
Thoma R, Mullins P, Ruhl D, Monnig M, Yeo RA, Caprihan A, Bogenschutz M, Lysne P, Tonigan S, Kalyanam R (2011). Perturbation of the glutamate–glutamine system in alcohol dependence and remission. Neuropsychopharmacology 36: 1359-1365.
Uddin LQ, Clare Kelly A, Biswal BB, Xavier Castellanos F, Milham MP (2009). Functional connectivity of default mode network components: correlation, anticorrelation, and causality.
Hum Brain Mapp 30: 625-637.
van Vliet I and de Beurs E (2007). Het Mini Internationaal Neuropsychiatrisch Interview (MINI). Een kort gestructureerd diagnostisch psychiatrisch interview voor DSM-IV en
ICD-10-stoornissen [The Mini International Neuropsychiatric Interview (MINI). A short structured diagnostic psychiatric interview for DSM-IV and ICD-10 disorders]. Tijdschrift voor Psychiatrie 49:
393-397.
Waddell KW, Zanjanipour P, Pradhan S, Xu L, Welch EB, Joers JM, Martin PR, Avison MJ, Gore JC (2011). Anterior cingulate and cerebellar GABA and Glu correlations measured by ¹H J-difference spectroscopy. Magn Reson Imaging 29: 19-24.
Wiers RW, Gladwin TE, Hofmann W, Salemink E, Ridderinkhof KR (2013). Cognitive Bias Modification and Cognitive Control Training in Addiction and Related Psychopathology Mechanisms, Clinical Perspectives, and Ways Forward. Clinical Psychological Science 1: 192-212.
Yang S, Salmeron BJ, Ross TJ, Xi Z, Stein EA, Yang Y (2009). Lower glutamate levels in rostral anterior cingulate of chronic cocaine users—A< sup> 1 H-MRS study using TE-averaged PRESS at 3 T with an optimized quantification strategy. Psychiatry Research: Neuroimaging 174:
171-176.
Yin HH and Knowlton BJ (2006). The role of the basal ganglia in habit formation. Nature Reviews Neuroscience 7: 464-476.
Yücel M, Lubman DI, Harrison BJ, Fornito A, Allen NB, Wellard RM, Roffel K, Clarke K, Wood SJ, Forman SD (2007). A combined spectroscopic and functional MRI investigation of the dorsal anterior cingulate region in opiate addiction. Mol Psychiatry 12: 691-702.
SUPPLEMENTARY INFORMATION
Figure S1 Representative diff (top) and even (bottom) spectrum of dorsal anterior cingulate cortex voxel of a nonsmoker
Figure S2 Representative functional connectivity plot between the dorsal anterior cingulate cortex (dACC) and nucleus accumbens (Nacc) of a nonsmoker
Table S1 Exclusion Smokers and Nonsmokers on Different Outcome Measures
Outcome measure Group Excluded Reason Included
Demographics smokers 0 48
(3) Missing data for education. (45)
nonsmokers 0 15
Clinical Characteristics smokers 5 Did not attend session 1.5. 43
nonsmokers 0 15
Glu dACC smokers 4 CRLB > 20%. 43
1 Outlier.
nonsmokers 1 CRLB > 20%. 14
Glu BG smokers 9 Putamen measured. -
14 CRLB > 20%. nonsmokers 4 CRLB > 20%. FC smokers 2 Movement > 3 mm. 46 nonsmokers 0 15 CR smokers 2 Movement > 3 mm. 45 1 Time constraints. nonsmokers 0 15
CR - full occipital smokers 13 Occipital lobe outside FOV. 32
2 Movement > 3 mm.
1 Time constraints.
nonsmokers 3 Occipital lobe outside FOV. 12
SST smokers 1 Outlier. 47
nonsmokers 1 Outlier. 13
1 Time constraints.
DDT smokers 2 Time constraints. 46
nonsmokers 1 Time constraints. 14
Abbreviations: Glu, referenced glutamate; dACC, dorsal anterior cingulate cortex; CRLB, Cramer-Rao Lower Bound; BG, basal ganglia; FC, functional connectivity; CR, cue reactivity; FOV, field of view; SST, Stop Signal Task; DDT, Delay Discounting Task.
Table S2 Demographics and Clinical Characteristics Smokers and Nonsmokers, Subscales Included
Smokers Nonsmokers T/U (df) p
M (SD) M (SD) Age 35.08 (9.83) 39.93 (13.11) 267.500 (61) 0.135 Education (SOI) 4.20 (0.76) 4.53 (0.99) 277.500 (58) 0.176 IQ (NLV) 105.29 (8.25) 108.93 (8.57) -1.48 (61) 0.144 BIS-11 66.84 (11.91) 59.93 (10.20) 2.04 (55) 0.046 BIS-att 17.87 (4.67) 15.50 (4.22) 1.44 (55) 0.155 BIS-mot 23.95 (4.83) 21.79 (4.17) 1.62 (55) 0.111 BIS-np 25.06 (4.04) 22.64 (4.81) 1.89 (55) 0.064 BIS 19.26 (3.97) 17.43 (3.48) 1.58 (55) 0.119 BAS-drive 11.97 (2.50) 10.71 (2.64) 1.79 (55) 0.079 BAS-fun 11.87 (2.05) 11.36 (1.78) 230.500 (55) 0.183 BAS-reward 16.54 (2.39) 17.29 (1.90) 292.500 (55) 0.873
Abbreviations: SOI, standaard onderwijs indeling; NLV, Nederlandse Leestest voor Volwassenen; BIS-11, Baratt Impulsiveness Scale; BIS, Behavioral Inhibition System; BIS-mot; BIS Motor subscale; BIS-att, BIS Attentional subscale; BIS-np, BIS non-planning subscale; BAS, Behavioral Activation System; BAS-drive, BAS Drive subscale; BAS-fun, BAS Fun Seeking subscale; BAS-reward, BAS Reward Responsiveness subscale.
Tabel S3 Results Smokers and Nonsmokers
Outcome measure Smokers Nonsmokers t/U (df) p
M (SD) M (SD)
Glu dACC (referenced) 0.66 (0.13) 0.56 (0.12) 2.42 (55) 0.019
FC (Z) 0.43 (0.19) 0.53 (0.17) -1.81 (59) 0.075
SST (ms) 251.56 (39.09) 256.04 (35.36) -0.40 (58) 0.692
DDT (AUC) 1958.68 (1156.42) 1863.41 (1050.66) 302.00 (58) 0.727
Abbreviations: Glu, glutamate; dACC, dorsal anterior cingulate cortex; FC, functional connectivity; SST, Stop Signal Task; DDT, Delay Discounting Task.
Table S4 Exploratory Correlations with dACC Glu in Smokers
N Spearman's rho p
FC (Z) 41 -0.06 0.711
CR - left calcarine ROI (beta) 39 -0.34 0.032
CR - left calcarine ROI (beta) 39 -0.33 0.041
SST (ms) 42 0.01 0.975
DDT (AUC) 42 -0.02 0.878
Abbreviations: Glu, glutamate; dACC, dorsal anterior cingulate cortex; FC, functional connectivity; CR, cue reactivity; SST, Stop Signal Task; DDT, Delay Discounting Task.
Supplemental Information on Exclusion Criteria
In order to maintain continuity in the research, as well as to optimize resemblance of the sample to the general population, participants who exhibited a dependence on a substance other than cigarettes were initially not excluded from the study. This was the case for both smokers and nonsmokers, so that smoking addiction would be the only differing variable. However, this resulted in the inclusion of a nonsmoker with a dependence on alcohol as well as cannabis, and one that portrayed hazardous drinking As it stems from our theory that most outcome measures are affected by addiction in general rather than by smoking addiction in particular, these two participants were excluded from further analysis.
Comparison of smokers and nonsmokers with the inclusion of these two particiants resulted in a somewhat less pronounced difference in referenced Glu in the dACC (t(57) = 2.23, p 0.029) and no significant difference in functional connectivity (t(61) = -1.11, p = 0.279).The cue
> noncue contrast within nonsmokers did not show any clusters of activation, and the bilateral calcarine activity for the smokers > nonsmokers contrast only occurred with a threshold of p <
0.005 uncorrected, k ≥ 10 voxels (left peak MNI: -12, -58, 7; k = 32; t = 3.23; right peak MNI:
15, -58, 7; k = 31; t = 2.97). Again, no difference was shown in SSRTi (t(59) = -0.36, p =0.724 )
