Functional connectivity analysis of the neural circuits of opiate craving: “more” rather than “different”?

Functional connectivity analysis of the neural circuits of opiate craving: “more” rather than “different”?

NeuroImage 20 (2003) 1964 –1970 www.elsevier.com/locate/ynimg Functional connectivity analysis of the neural circuits of opiate craving: “more” rath...

375KB Sizes 1 Downloads 50 Views

NeuroImage 20 (2003) 1964 –1970

www.elsevier.com/locate/ynimg

Functional connectivity analysis of the neural circuits of opiate craving: “more” rather than “different”? Mark R.C. Daglish,a,b,c Aviv Weinstein,a Andrea L. Malizia,a,b Susan Wilson,a Jan K. Melichar,a,b Anne Lingford-Hughes,a,b Judith S. Myles,c,d Paul Grasby,b and David J. Nutta,e,* a

Psychopharmacology Unit, University of Bristol, UK MRC Clinical Sciences Centre, Hammersmith Hospital, London, UK c Bristol Specialist Drug Service, Blackberry Hill Hospital, Bristol, UK d St. George’s Hospital Medical School, London, UK e Avon & Wiltshire Mental Health Partnership NHS Trust, UK b

Received 26 February 2003; revised 18 July 2003; accepted 18 July 2003

Abstract We investigated the functional connectivity of brain regions activated during opiate craving. Previously we used recorded autobiographical scripts to induce opiate craving in 12 abstinent opiate-dependent subjects while they were undergoing positron emission tomography (PET) scanning using the regional cerebral blood flow (rCBF) tracer H215O. SPM99 was used to examine the connectivity patterns associated with the primary brain regions activated in response to drug-craving memories (anterior cingulate, AC) and correlated with opiate craving (orbitofrontal cortex, OFC). Two separate connectivity patterns were identified associated with the OFC and AC regions. The AC region was associated with activity in the left temporal region. The left OFC region activity correlated with activity in the right OFC, and left parietal and posterior insular regions. There was also a positive association with the hippocampus and brainstem. Both the AC and OFC regions showed a negative association with posterior visual areas. We suggest that the patterns of functional connectivity reflect the ability of drug-related stimuli to activate attentional and memory circuits to a greater degree than non-drug-related stimuli. This argues that neural circuits of dependence and craving are not specific “craving” or “addiction” brain regions but are “normal” circuits activated to a greater degree. © 2003 Elsevier Inc. All rights reserved.

Introduction Drug abuse is a chronic condition characterized by remissions and relapses. It remains unclear why abstinence can be so difficult to maintain for individuals who were previously dependent. Craving is a term often used by such individuals to explain their relapses to heroin use, but it has proved difficult to define this term rigorously in a scientific context (Robinson and Berridge, 1993; Pickens and Johanson, 1992). Cue responsivity, or cue exposure, has been

* Corresponding author. Psychopharmacology Unit, School of Medical Sciences, University Walk, Bristol, BS8 1TD, UK. Fax: ⫹44-0-117-9277057. E-mail address: [email protected] (D.J. Nutt). 1053-8119/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2003.07.025

repeatedly used to study craving. The term refers to the ability of situations, drug paraphernalia, and mood states or memories associated with previous drug use to effect the desire to use the drug (Childress et al., 1988). Neuroimaging techniques have been used to study these processes in cocaine, nicotine, and opiate dependence. Several brain areas are consistently reported to show activations across paradigms, particularly the dorso-lateral prefrontal cortex (Grant et al., 1996; Maas et al., 1998; Brody et al., 2002; Bonson et al., 2002), anterior cingulate (AC) cortex (Maas et al., 1998; Daglish et al., 2001; Childress et al., 1999; Brody et al., 2002), amygdala (Grant et al., 1996; Childress et al., 1999; Bonson et al., 2002), and orbitofrontal cortex (OFC) (Volkow et al., 1999; Daglish et al., 2001; Wang et al., 1999; Sell et al., 2000; Childress et al.,

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

1965

Fig. 1. Brain regions (Brodmann areas) associated with the left AC region. Solid lines with labels indicate a direct relationship between brain regions with the label specifying the analyses when this relationship was demonstrated. Dotted lines denote a modulation of the relationship by scan condition or subject group

1999; Brody et al., 2002; Grant et al., 1996; Bonson et al., 2002). Previously we used H215O positron emission tomography (PET) to examine changes in regional cerebral blood flow (rCBF) in response to opiate-related stimuli and associated with craving for opiates (Daglish et al., 2001). We demonstrated that individuals who were previously dependent on opiates (even those who had remained abstinent from opiates for at least one year) exhibited craving for opiate drugs after exposure to cues that consisted of autobiographical scripts (Daglish et al., 2001). Furthermore, these stimuli were associated with increased rCBF in the left anterior cingulate cortex and left orbito-frontal cortex activity was correlated with subjective opiate craving (Daglish et al., 2001). Cognitive subtraction studies of the above type identify brain regions involved in select psychological processes, but are not designed to demonstrate functional connectivity within the brain. Applying connectivity methods originally used to image the neural circuitry of visual attention (Friston et al., 1997) it is possible to examine the functional circuitry involved in opiate craving. We report here a connectivity analysis of the data from our previous study (Da-

glish et al., 2001). In particular, we examine the patterns of rCBF associated with responses in the AC and OFC, the two brain areas identified in our previous study.

Method Subjects The subject group is as previously described (Daglish et al., 2001) with 12 opiate-free previously heroin-dependent subjects. This study was approved by the local ethics committees and the Administration of Radioactive Substances Advisory Committee (ARSAC). After complete description of the study, written informed consent was obtained. Protocol Subjects recorded two audio taped scripts prior to the scanning session. Each script was of 2 min duration and recounted a single specific episode from their past. In one case the script concerned an episode when they experienced a strong craving for opiates; the other was of a neutral

1966

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

Fig. 2. Brain regions (Brodmann areas) associated with the left OFC region Solid lines with labels indicate a direct relationship between brain regions with the label specifying the analyses when this relationship was demonstrated. Dotted lines denote a modulation of the relationship by scan condition or subject group.

episode. Subjects were specifically told to focus on describing the feelings, both emotional and visceral, of the opiate craving they experienced during the episode they were recounting. Subjects then underwent a 12-run PET scan of rCBF using the tracer H215O, with a slow bolus of 333 MBq infused over 20 s. There was a 10-min gap between runs, with the total scan session lasting 2 h. A brain-dedicated Siemens/CTI ECAT 953b PET camera was used, operating in 3D mode. During each of the 12 runs the subject listened to one of the two previously recorded audio-taped scripts. Each drugrelated and neutral script was presented six times, in random order. Image acquisition began 90 s after the beginning of the script to allow craving to be induced in response to the stimulus. Images were acquired in a single 90-s frame. After each run the subjects completed six visual analogue scales (VAS) of “craving” for heroin, “urge to use” heroin, “happy,” “sad,” “anxiety,” and “vividness” of the script. Image analysis The PET images of rCBF were analyzed using Statistical Parametric Mapping (SPM99, Wellcome Department of

Cognitive Neurology, London). The 12 images for each subject were realigned to a mean image for that subject and normalized to a standardized template H215O PET image in MNI space, using windowed sinc interpolation. The resultant realigned normalized images were then smoothed with a Gaussian function at 12 mm full width half-maximum. Our previously reported categorical analysis of these images showed activation in the left anterior cingulate gyrus in response to the opiate-related stimulus compared to the neutral stimulus in all subjects (Daglish et al., 2001). Activation in the left orbito-frontal region was also shown to be correlated with subjective craving reported by the eight (66%) subjects who craved. For each of these two regions, SPM was used to extract the first eigenvariate from all voxels within a radius of 10 mm of the peak of activation for each scan. These two variables were then entered into further SPM analyses. In each case the covariate was centered around the subject mean and modeled with a covariate by subject interaction. The time in minutes from the start of the scanning session and the global mean were entered into the model allowing a subject-specific fit. For each eigenvariate, three separate analyses were per-

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

1967

Fig. 3. Scatter plots showing the association between left anterior cingulate gyrus (X axis) and Brodmann Area 21 (Y axis) activity for all 12 subjects individually, with linear trendline fitted.

formed: a simple association, an interaction with scan condition, and an interaction with subject group. Subject group was determined post hoc by whether the subject demonstrated a craving response to the stimuli. In the first analysis the eigenvariate alone was entered into SPM99 as described above. This tests for brain regions where activity is functionally connected with the AC or OFC region. In the next analyses two separate methods were used to test for brain regions where the scan condition or subject group modulated the correlation of activity with the covariate. First an interaction variable was calculated by multiplying the eigenvariate by 1 or ⫺1 depending on the condition or group. This was entered into an SPM99 analysis along with the main variable of condition or group. In the second interactions method, the conditions or groups and the eigenvariate were entered into a combined analysis where the variable was centered around the condition or

group mean and interactions tested for by contrasting the difference between the covariance of the brain activity between the two conditions. The threshold for statistical significance for all analyses was set at P ⬍ 0.05 after correction for multiple comparisons for either peak change in rCBF or the size of the activated cluster. In all cases the analysis uses 12 scans per subject with no differentiation made between intra- and intersubject variance measures. As such, this is a fixedeffects analysis and consequently any statistical inference extends only as far as the subjects in this study.

Results Two networks of neural activation related to the primary areas found from the previous analysis are shown in Figs. 1

1968

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

Fig. 4. Scatter plots with linear trendlines showing the change in association between left OFC (X axis) and Brainstem (Y axis) activity between the craving (■) and neutral (⫹) conditions for all 12 subjects.

and 2. All P values reported are corrected for multiple comparisons by SPM99. The first network (Fig. 1) is of regions where rCBF is associated with activity in the left AC region (Brodmann Area, BA 32). A positive association with middle temporal gyrus was found on the ipsilateral (left) side (⫺62, ⫺8, ⫺18mm, BA21, cluster P ⫽ 0.003, voxel P ⫽ 0.003). Fig. 3 shows the consistency of this association across all 12 subjects. There was a consistent negative relationship between the AC and posterior visual areas (BA 18 right (28, ⫺92, 8, cluster P ⬍ 0.001, voxel P ⫽ 0.026) and BA 19 bilateral (left ⫺16, ⫺86, 18, cluster P ⬍ 0.001, voxel P ⫽ 0.014) (right 28, ⫺72, 34, cluster P ⫽ 0.032, voxel P ⫽ 0.257)). In BA 19 on both sides there was a significant modulation of the negative association with AC activity by the scan condition and BA 18 on the right side. Finally for this circuit there was a significant modulation of the association between the AC region and left BA 3 (⫺34, ⫺24,

36, cluster P ⫽ 0.713, voxel P ⫽ 0.043) depending on the presence or absence of a subjective craving response, i.e., an interaction with subject group. The circuit associated with rCBF in the left orbito-frontal region (BA 11) is depicted in Fig. 2. As with activity in the AC region there was a negative association between the OFC and posterior visual cortex bilaterally (BA 17/18 left (⫺14, ⫺86, 2 cluster P ⬍ 0.001, voxel P ⫽ 0.081) and BA 17 right (8, ⫺80, 14, cluster P ⬍ 0.001, voxel P ⫽ 0.041)). A similar negative association was seen in right BA 37 (52, ⫺60, 0, cluster P ⫽ 0.02, voxel P ⬍ 0.001). There was a significant association between the left and the right OFC (BA 11) only for those who craved in response to the drug-related stimulus (26, 48, ⫺18, cluster P ⫽ 0.230, voxel P ⫽ 0.038). The same pattern was seen in the left parietal region BA 48 (⫺36, 20, ⫺2, cluster P ⫽ 0.002, voxel P ⫽ 0.194). This region also showed a significant modulation of the association dependent on the subject

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

group. For all subjects there was a positive association between activity in the left OFC (BA 11) and left posterior insular cortex (BA 41, ⫺26, ⫺38, 10, cluster P ⫽ 0.006, voxel P ⬍ 0.001). The scan condition was shown to have a significant modulatory effect on the association between activity in the left OFC (BA 11) and two subcortical regions. Specifically, this result was seen in the brainstem— centered in the area of the red nucleus (0, ⫺16, ⫺8, cluster P ⫽ 0.170, voxel P ⫽ 0.025) and the left hippocampus (⫺20, ⫺36, 4, cluster P ⫽ 0.092, voxel P ⫽ 0.034). In these two areas no main association with either the OFC activity or the scan condition was observed. Fig. 4 shows the association between the left OFC and brainstem regions for each subject, in particular the change in slope between the craving and neutral conditions in the 12 subjects.

Discussion We previously reported increased activity in the left AC (BA 32) in response to drug-related stimuli in drug-free former heroin users that was independent of their subjective response to the stimulus (Daglish et al., 2001). In addition, left OFC (BA 11) activity was dependent on the level of subjective craving response. The analysis of rCBF reported in this article extends the above findings by defining brain regions functionally connected to these areas. Areas found in this analysis that positively correlate with activity in the AC were deep in the posterior central gyrus (BA 3) and middle temporal gyrus (BA 21). Neither of these areas is classically thought to be connected to the AC region. However, the functions of these areas do have a possible link with the task condition, which may explain this apparent functional connectivity. Brodmann area 21 is involved in auditory sensory input and this experiment used auditory stimuli. The activation in the posterior central gyrus is in the area representing intraabdominal sensation (Penfield and Rasmussen, 1950). It is therefore plausible that both of these areas are activated by opiate-related auditory stimuli that may evoke a visceral response in addition to any cognitive response. The opiate-related stimuli are likely to engage a greater auditory attention to the stimulus, which would explain the increase in BA 21 and the negative correlation between the AC region and visual areas via a cross-modality suppression of visual areas (Fig. 1). The nodes in the OFC connection network that activated proportionally with craving also have face validity when considering the nature of the experimental situation. Hippocampus and temporal regions are implicated in memory or episodic memory. The stimuli used here were autobiographical episodes of craving. The more the script evoked activation in the memory areas the more subjective craving was described. The brainstem (red nucleus) region is usually considered to be involved in the integration of motor activ-

1969

ity. This region also has projections from the anterior cingulate cortex and cerebellar dentate gyrus. The spatial resolution of PET is not sufficient to differentiate the red nucleus from other nearby brainstem nuclei more often implicated in the opiate system, e.g., periaqueductal gray matter. However, a recent fMRI study of video-induced heroin craving, with the greater spatial resolution of fMRI, has also shown red nucleus to be activated in response to drug-related stimuli (Langleben et al., 2002). The negative association between activity in the OFC and posterior visual areas could again be related to cross-modal suppression. These results suggest that opiate craving and drug-related stimuli do not activate “special” brain circuits specific for drug dependence or craving. Instead, the connected networks reported in this study identify circuits related to attention, sensory processing, and memory. In other words, drug-dependence circuitry is perhaps associated with a greater degree of activation of regions associated with processing the autobiographical stimuli rather than activating “special” “addiction” regions of the brain. Our results add further support to the theories of addiction that suggest that what makes dependence is the ability of the drug and its related stimuli to “hi-jack” the neural circuits usually activated by attention and motivation and drive them stronger than “normal” rewards and stimuli (e.g., the incentive-sensitization model of Robinson and Berridge (1993). A number of caveats apply to these data. This is a study on a limited number of subjects using a fixed-effects model for analysis and it is thus unclear how far these results can be extrapolated beyond this group. The example plots shown in Figs. 3 and 4 provide an estimate of the consistency of the findings across all 12 subjects. It is also necessary to bear in mind the limited spatial and temporal resolution afforded by PET scanning. This technique of examining functional connectivity has allowed us to highlight a network of brain regions implicated in opiate dependence, but it does not allow us to study the nature of the interactions between them. At present this is a map of nodes whose level of activation shows an association. It is not possible to state a causal relationship between activity in connected brain region, or the strength of one, should it exist. This will require further study with techniques better suited to the study of effective connectivity, but where the experimental environment is even less conducive to the state of mind we are trying to study, e.g., fMRI.

Acknowledgments This work was, in part, funded by an MRC Programme Grant. The authors thank the staff of the drug treatment agencies for help in recruiting the subjects for this study, in particular Steve Barnes, Dr. Colin Brewer, and Dr. Simon Britten. A.L.M. was a Wellcome Training Fellow.

1970

M.R.C. Daglish et al. / NeuroImage 20 (2003) 1964 –1970

References Bonson, K.R., Grant, S.J., Contoreggi, C.S., Links, J.M., Metcalfe, J., Weyl, L., Kurian, V., Ernst, M., London, E.D., 2002. Neural systems and cue-induced cocaine craving. Neuropsychopharmacology 26 (3), 376 –386. Brody, A.L., Mandelkern, M.A., London, E.D., Childress, A.R., Lee, G.S., Bota, R.G., Ho, M.L., Saxena, S., Baxter Jr., L.R., Madsen, D., Jarvik, M.E., 2002. Brain metabolic changes during cigarette craving. Arch. Gen. Psychiatry 59 (12), 1162–1172. Childress, A.R., McLellan, A.T., Ehrman, R., O’Brien, C.P., 1988. Classically conditioned responses in opioid and cocaine dependence: a role in relapse? NIDA Res. Monogr. 84, 25– 43. Childress, A.R., Mozley, P.D., McElgin, W., Fitzgerald, J., Reivich, M., O’Brien, C.P., 1999. Limbic activation during cue-induced cocaine craving. Am. J. Psychiatry 156 (1), 11–18. Daglish, M.R.C., Weinstein, A., Malizia, A.L., Wilson, S., Melichar, J.K., Britten, S., Brewer, C., Lingford-Hughes, A., Myles, J.S., Grasby, P., Nutt, D.J., 2001. Changes in regional cerebral blood flow elicited by craving memories in abstinent opiate-dependent subjects. Am. J. Psychiatry 158 (10), 1680 –1686. Friston, K.J., Buechel, C., Fink, G.R., Morris, J., Rolls, E.T., Dolan, R.J., 1997. Psychophysiological and modulatory interactions in neuroimaging. NeuroImage 6, 218 –229. Grant, S., London, E.D., Newlin, D.B., Villemagne, V.L., Liu, X., Contoreggi, C., Phillips, R.L., Kimes, A.S., Margolin, A., 1996. Activation of memory circuits during cue-elicited cocaine craving. Proc. Natl. Acad. Sci. USA 93 (21), 12040 –12045.

Langleben, D.D., Wang, J., Gray, J., Fornash, A., O’Brien, C.P., Childress, A.R., 2002. Functional magnetic resonance imaging (fMRI) of regional cerebral blood flow during heroin-related cues in opiate-dependent subjects. Drug Alcohol Dep. 66 (Suppl. 1), S99. Maas, L.C., Lukas, S.E., Kaufman, M.J., Weiss, R.D., Daniels, S.L., Rogers, V.W., Kukes, T.J., Renshaw, P.F., 1998. Functional magnetic resonance imaging of human brain activation during cue-induced cocaine craving. Am. J. Psychiatry 155 (1), 124 –126. Penfield, W., Rasmussen, T., 1950. The Cerebral Cortex of Man. Macmillan, New York. Pickens, R.W., Johanson, C.E., 1992. Craving: consensus of status and agenda for future research. Drug Alcohol Dep. 30 (2), 127–131. Robinson, T.E., Berridge, K.C., 1993. The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Res. Brain Res. Rev. 18 (3), 247–291. Sell, L.A., Morris, J.S., Bearn, J., Frackowiak, R.S., Friston, K.J., Dolan, R.J., 2000. Neural responses associated with cue evoked emotional states and heroin in opiate addicts. Drug Alcohol Dep. 60 (2), 207–216. Volkow, N.D., Wang, G.-J., Fowler, J.S., Hitzemann, R., Angrist, B., Gatley, S.J., Logan, J., Ding, Y.-S., Pappas, N., 1999. Association of methylphenidate-induced craving with changes in right striato-orbitofrontal metabolism in cocaine abusers: implications in addiction. Am. J. Psychiatry 156 (1), 19 –26. Wang, G.J., Volkow, N.D., Fowler, J.S., Cervany, P., Hitzemann, R.J., Pappas, N.R., Wong, C.T., Felder, C., 1999. Regional brain metabolic activation during craving elicited by recall of previous drug experiences. Life Sci. 64 (9), 775–784.