Introduction

Brain imaging approaches are powerful tools to identify differences in brain structure and function in individuals with drug addiction compared to healthy controls (Goldstein and Volkow, 2002; Volkow et al, 1997b, 2003). Some of these modifications persist long after discontinuation of drug use and may underlie the persistent risk of relapse over time whereas others appear recovered after withdrawal from the drug (Goldstein and Volkow, 2002; Parvaz et al, 2011). For example, using 2-deoxy-2-[18F]fluoro-d-glucose (18FDG) positron emission tomography (PET) imaging, it has been found that during early withdrawal, individuals with cocaine addiction show global increases in functional brain activity whereas after protracted periods of withdrawal, brain activity decreases compared to naive subjects (Goldstein and Volkow, 2002; Parvaz et al, 2011; Volkow et al, 1991, 1997b). Neuroimaging studies have been instrumental to the understanding of the brain changes associated with addiction (Goldstein and Volkow, 2002; Koob and Volkow, 2010).

Whereas brain imaging in humans have greatly contributed to our understanding of the neurobiological dysfunctions associated with drug addiction, rodent models could be invaluable tools to rapidly test novel hypotheses that could be then translated back to humans. In animals, several studies have used brain imaging to investigate the changes in brain activity induced by drugs (Caprioli et al, 2013; Gould et al, 2014; Hammer et al, 1993; Hanlon et al, 2013; Howell, 2008; Macey et al, 2004). Using functional magnetic resonance imaging (fMRI), Gozzi et al found that after 10 days of abstinence rats with a history of extended access to cocaine self-administration show decreases in basal cerebral blood flow in fronto-cortical regions such as the orbitofrontal (OFC), prefrontal and anterior cingulate (ACC) cortices, and the nucleus accumbens (NAc) (Gozzi et al, 2011). In another study, Calipari et al used ex vivo 2-[14C]deoxyglucose methods and found a decrease in the activity of several brain areas 48 h after 5 days of extended access to cocaine self-administration, including the NAc and the ACC (Calipari et al, 2013). However until now, no longitudinal study has been performed to assess changes in brain activity that occur after short and long periods of abstinence from cocaine. Importantly, investigating changes in brain activity after long periods of withdrawal could enable obtaining information about brain recovery processes and about mechanisms underlying incubation of drug craving (Grimm et al, 2001; Hanlon et al, 2013; Pickens et al, 2011).

In the present study, we used PET with 18FDG to evaluate metabolic activity changes after short (1 week) and long (4 weeks) periods of abstinence from cocaine in rats with a history of cocaine self-administration using the escalation model (Ahmed and Koob, 1998). In this model, rats are given either short-access (ShA) or long-access (LgA) to cocaine self-administration for several weeks, allowing comparing recreational use in ShA rats to extended addiction-like escalated intake in LgA rats. Thus, the comparison of the brain metabolic maps obtained for ShA, LgA and naive rats, may provide insights into the short- and long-term cerebral metabolic changes after withdrawal after exposure to different regimens of cocaine intake.

Materials and methods

For detailed material and methods see Supplementary Information.

Subjects and Housing Conditions

Adult male Sprague-Dawley rats (Janvier Labs, France), experimentally naive at the beginning of the study, were used in this study. All experiments were conducted in accordance with European Union directives (2010/63/EU) for the care of laboratory animals and approved by the local ethics committees (COMETHEA).

General Experimental Design

Our general experimental design is schematized in Figure 1. During the first 7 days, rats were allowed to self-administer cocaine in 2-h training sessions. On the 8th day, rats were divided into two groups: one group had access to cocaine for 1 h (ShA, n=8) and the other group had access to cocaine for 6 h (LgA, n=8) to 20 sessions. At the end of the last self-administration session (day 20), rats were transferred from the animal facility at the University of Poitiers to the animal facility at University of Tours by an authorized transporter and underwent abstinence for 4 weeks. Metabolic imaging using 18FDG was performed in the same rats after 6–8 days (1 week) and 27–29 days (4 weeks) of abstinence. Naive rats of the same age and with analogous housing conditions were used as controls (n=8).

Figure 1
figure 1

Experimental design. After 7 days of cocaine self-administration training (2-h/day), rats were assigned to two groups short access (ShA), which have access to cocaine for 1-h/day and long access (LgA), which have access to cocaine for 6-h/day. At the end of the 20 days of cocaine self-administration, rats underwent a 4-week abstinence during which brain metabolic imaging using 18FDG was performed at 1 week and 4 weeks of abstinence. Naive rats were used as controls.

PowerPoint slide

Full size image

Brain Imaging

Local uptake of 18FDG reflects cerebral metabolic rates of glucose utilization and allows the investigation of regional brain metabolic status (Phelps et al, 1979; Sokoloff, 1977). Metabolic imaging using 18FDG was performed under basal conditions. Rats were habituated to the PET experimental procedures for 4 days before each scan, and fasted overnight before each scan. The day of brain-imaging acquisition, awake rats were injected with 18FDG (18.5 MBq/100 g i.p.; Cyclopharma, Tours), and placed in the habituation cage for 45 min. Then, they were anesthetized using isoflurane 4% (Baxter, Maurepas, France), placed on a heating pad (Minerve, Esternay, France) and centered in the field of view of the Explore VISTA-CT microPET camera (GE Healthcare, Velizy, France). A CT-scan was performed for attenuation correction of PET images and a list-mode PET acquisition of 30 min started 60 min after 18FDG injection. After data reconstruction using a 2-D OSEM algorithm, all images were co-registered and normalized for tissue activity in the whole brain. Quantitative results were expressed as mean±SD and were presented on Z-score maps. Analyses focused on brain areas known to be key nodes in addiction: the ACC, OFC, prelimbic (PrL), infralimbic (IL), insular and motor cortices, in addition to the dorsal striatum (DStr), NAc, ventral pallidum (VP), substantia nigra/ventral tegmental area (SN/VTA), amygdala (Amyg), and hippocampus (Hipp) (Volkow and Baler, 2014). It should be noted that areas that were too small to be identified using microPET 18FDG imaging were not included in our analysis.

Statistical Analyses

For self-administration, two-way repeated measures ANOVA with time as a within-subject factor and cocaine exposure (LgA or ShA) as a between-subject factor was used. Results showing significant overall changes were subjected to a Student-Newman-Keuls post hoc test.

For micro-PET data, a voxel-based analysis was used to assess the differences in cerebral 18FDG uptake between the averaged brains of LgA/ShA vs control rats at each stage of withdrawal. The regions of interest were derived from Schiffer’s templates (Schiffer et al, 2006) using PMOD v3.2 software (PMOD Technologies Ltd, Switzerland) and applied to Z-score maps to obtain the Z-score values in these areas. Inter-group comparison was performed using a two-tail unpaired Student t-test. Intra-group comparisons were performed using a two-tail paired t-test to reveal within-rat differences over the course of cocaine withdrawal. Differences were considered significant when p<0.05.

Results

Self-Administration

Cocaine self-administration started with seven 2-h sessions (Figure 2, left in all graphs a–d). After the initial training, rats were divided into two groups and were allowed to self-administer cocaine for 1-h (ShA) and 6-h (LgA) for 20 additional sessions (Figure 2, right in all graphs a–d). In the ShA group, active nose-pokes (Figure 2a) and cocaine intake (Figure 2c) during the 20 sessions were stable whereas in the LgA group, the number of active nose-pokes (Figure 2a) and cocaine intake (Figure 2c) significantly increased over time. Figure 2d shows the cocaine intake during the first hour of access to the drug in ShA and LgA. During the first 7 days of self-administration ShA and LgA rats had similar levels of cocaine intake. In the ShA group the cocaine intake was stable over time, from 8.0±1.6 at day 8 to 8.8±1.3 mg/kg of cocaine at day 27. In contrast, in the LgA group, the cocaine intake significantly escalated over time, from 10.4±1.7 at day 8 to 14.8±0.9 mg/kg of cocaine at day 27. Statistical analysis revealed a significant effect of time [F(20 280)=2.83, p<0.0001] and a significant GroupXTime interaction [F(20 280)=2.14, p<0.01].

Figure 2
figure 2

Cocaine self-administration training. Number of active (a) and inactive (b) nose-pokes, total cocaine intake (c) and cocaine intake during the first hour of sessions (d) during the training phase in which all animals had access to cocaine for 2 h (left part of the graphs) and during the escalation phase in which rats were divided into short access (ShA, n=8, 1-h sessions) and long access (LgA, n=8, 6-h sessions) groups (right part of the graphs). *p<0.05 and **p<0.01 compared to session 1 of LgA self-administration, ##p<0.0001 LgA compared to ShA group.

PowerPoint slide

Full size image

Brain Metabolic Activity in Long-Access and Short-Access Rats During Abstinence in Areas Involved in Reward, Salience, Motivation and Drive

Nucleus accumbens

In the NAc, 18FDG uptake was not altered in the ShA group (Figure 3) but was reduced in LgA rats vs controls after short periods of abstinence (Table 1) and recovered after 4 weeks of abstinence (Figure 4, Tables 1 and 2). The direct comparison between LgA and ShA rats confirmed that the metabolic activity of the NAc was reduced after 1 but not after 4 weeks of abstinence in LgA rats compared to ShA rats (Figure 5, Table 1).

Figure 3
figure 3

Changes in metabolic activity in ShA rats compared to naive controls. (a) Summary of the significant increases (red) and decreases (blue) in 18FDG uptake observed in ShA rats vs controls (n=8/group) after 1 week (upper part) and 4 weeks (lower part) of cocaine withdrawal presented on representative coronal plates of the Paxinos and Watson atlas (Student’s two-tailed t-test; p<0.01). (b) Examples of the significant differences in 18FDG uptake observed between ShA and control rats after 1 week (upper part) and 4 weeks (lower part) of cocaine abstinence presented on coronal images of Z-score maps fused with an MRI template (increases in 18FDG uptake from dark red to yellow, decreases in 18FDG uptake from black to light blue; Student’s two-tailed t-test; p<0.01). ACC, anterior cingulate cortex; Amyg, amygdala; DHipp, dorsal hippocampus; DMStr, dorsomedial striatum; DLStr, dorsolateral striatum; IL, infralimbic cortex; Ins, insula; Mot, motor cortex; NAc, nucleus accumbens; OFC, orbitofrontal cortex; PrL, prelimbic cortex; SN/VTA, substantia nigra/ventral tegmental area; VHipp, ventral hippocampus; VP, ventral pallidum.

PowerPoint slide

Full size image
Table 1 Statistical Significances for Inter-Group Comparisons
Full size table
Figure 4
figure 4

Changes in metabolic activity in LgA rats compared to naive controls. (a) Summary of the significant increases (red) and decreases (blue) in 18FDG uptake observed in LgA rats vs controls (n=8/group) after 1 week (upper part) and 4 weeks (lower part) of cocaine abstinence presented on representative coronal plates of the Paxinos and Watson atlas (Student’s two-tailed t-test; p<0.01). (b) Examples of the significant differences in 18FDG uptake observed between LgA and control rats after 1 week (upper part) and 4 weeks (lower part) of cocaine abstinence presented on coronal images of Z-score maps fused with an MRI template (increases in 18FDG uptake from dark red to yellow, decreases in 18FDG uptake from black to light blue; Student’s two-tailed t-test; p<0.01). ACC, anterior cingulate cortex; Amyg, amygdala; DHipp, dorsal hippocampus; DMStr, dorsomedial striatum; DLStr, dorsolateral striatum; IL, infralimbic cortex; Ins, insula; Mot, motor cortex; NAc, nucleus accumbens; OFC, orbitofrontal cortex; PrL, prelimbic cortex; SN/VTA, substantia nigra/ventral tegmental area; VHipp, ventral hippocampus; VP, ventral pallidum.

PowerPoint slide

Full size image
Table 2 Statistical Significances for Intra-group Comparisons
Full size table
Figure 5
figure 5

Comparisons of the changes in metabolic activity in LgA vs ShA rats. Summary of the significant increases (red) and decreases (blue) in 18FDG uptake observed in LgA vs ShA rats (n=8/group) after 1 week (upper part) and 4 weeks (lower part) of cocaine abstinence presented on representative coronal plates of the Paxinos and Watson atlas (Student’s two-tailed t-test; p<0.01).

PowerPoint slide

Full size image

Ventral pallidum

No significant differences between groups were found in the VP at any time point (Figures 3 and 4).

Dorsal striatum

In the DStr, 18FDG uptake was increased in both ShA and LgA rats in the medial part (DMStr) after short periods of abstinence (Table 1) and decreased in the lateral part (DLStr) after long periods of abstinence compared to controls (Figures 3 and 4, Table 1). Thus, the metabolic activity was restored in the DMStr and decreased in the DLStr during cocaine abstinence in both groups (Table 2). Importantly, direct comparison of LgA to ShA rats showed that the metabolic activity of the DLStr was similar after 1 week of abstinence but it was lower in LgA compared to ShA rats after 4 weeks (Figure 5, Table 1), suggesting that extended access to cocaine produces quantitatively bigger disruptions in the activity of this brain area.

Substantia nigra/ventral tegmental area

In the SN/VTA, increased 18FDG uptake was observed after 1 week of abstinence in both ShA and LgA rats compared to controls (Figures 3, and 4, Table 1), but this effect was stronger in LgA than in ShA rats (Figure 5, Table 1). After 4 weeks of abstinence, increased 18FDG uptake was still observed and it was significantly stronger in the SN/VTA of LgA rats (Figure 3, Tables 1 and 2), whereas a decreased uptake was observed in ShA rats (Figure 4, Tables 1 and 2). Direct comparison of LgA and ShA rats confirmed that the metabolic activity of the SN/VTA was higher in LgA compared to ShA rats (Figure 5, Table 1).

Orbitofrontal cortex

In the OFC, a decreased metabolic activity compared to controls was found in ShA rats after 1 week of abstinence but this returned to control levels after 4 weeks of abstinence (Figure 3, Tables 1 and 2). No change in the metabolic activity of this region was detected in LgA rats (Figure 4). Direct LgA/ShA comparison showed that the metabolic activity in the OFC of LgA rats was increased compared to ShA rats after 1 week of abstinence (Figure 5, Table 1) but this effect was due to metabolic hypoactivity in ShA rats rather than a change compared to control.

Motor cortex

In the Motor cortex, increased 18FDG uptake was found in ShA (Figure 3, Table 1) but not LgA rats compared to controls after 1 week of abstinence (Figure 4). However, the metabolic activity of this region returned to control levels after 4 weeks of abstinence (Figures 3 and 4, Table 2). Direct LgA/ShA comparison highlighted a lower metabolic activity in the Motor cortex of LgA compared to ShA rats after 1 week of abstinence (Figure 5, Table 1) but this effect was due to metabolic hyperactivity in ShA rats rather than a change compared to controls.

Brain Metabolic Activity in Long-Access and Short-Access Rats after Abstinence in Areas Involved in Inhibitory Control, Executive Functions, and Interoception

Prelimbic and infralimbic cortices

No change in the metabolic activity of the PrL was observed in any group at any time point (Figures 3 and 4). On the other hand, the IL shows increases in 18FDG uptakes in ShA (Figure 3, Table 1) and decreases in LgA rats (Figure 3, Table 1) after 1 week of abstinence but both modifications disappeared after 4 weeks of abstinence (Figures 3 and 4, Table 2). Direct LgA/ShA comparison confirmed that the IL of LgA rats show a hypoactive metabolism compared to ShA rats specifically after a short period of abstinence (Figure 5, Table 1).

Anterior cingulate cortex

The 18FDG uptake was reduced in the ACC in both ShA and LgA rats compared to controls after both short and long periods of abstinence (Figures 3 and 4, Table 1). Whereas the decreases observed in LgA rats persisted overtime, in ShA rats the effect was significantly weaker after 28 days of cocaine abstinence (Table 2) indicating some level of recovery in this group. When we compared directly LgA to ShA rats, we found that the hypoactive metabolism was more pronounced in ShA rats after 1 week of abstinence (Table 1), while the contrary was observed after 4 weeks of abstinence with LgA rats showing less activity of the ACC compared to ShA rats (Figure 5, Table 1). This suggests that extended self-administration of cocaine produced more intense and more persistent disruptions in the activity of this brain area.

Insular cortex

In the insula, the 18FDG uptake was reduced in LgA, but not in ShA rats after short and long periods of abstinence (Figures 3 and 4, Table 1). Direct LgA/ShA comparison confirmed that the metabolic activity in this area was consistently lower in LgA compared to ShA rats (Figure 5, Table 1). The effect observed in LgA rats persisted over the course of cocaine withdrawal (Table 2).

Brain Metabolic Activity in Long-Access and Short-Access Rats after Abstinence in Areas Involved in Contextual and Emotional Memories

Hippocampus

After short periods of abstinence, metabolic activity was reduced in both ShA and LgA rats in the dorsal part of the Hipp (DHipp) and in the ventral part (VHipp; Figures 3 and 4, Table 1).

After 4 weeks of abstinence, in ShA rats the metabolic activity of the DHipp was still reduced (Table 1) and not significantly different from was observed after 1 week (Table 2), whereas the VHipp became hyperactive (Tables 1 and 2). In contrast, both the DHipp and the VHipp became hyperactive in LgA rats compared to controls after 4 weeks of abstinence (Figure 4, Tables 1 and 2). When we compared directly LgA to ShA rats, the metabolic activity of the VHipp was lower in LgA rats compared to ShA rats after 1 week of abstinence suggesting that hypoactivity of this region was more pronounced in LgA than in ShA rats (Figure 5, Table 1). In addition, after 4 weeks of abstinence, the metabolic activity of the DHipp was higher in LgA compared to ShA rats (Figure 5, Table 1).

Amygdala

In ShA rats, the metabolic activity of the Amyg, was increased compared to controls after short periods of abstinence (Table 1) but recovered after 4 weeks of abstinence (Figure 3, Table 2). Conversely, in LgA rats the metabolic activity of the Amyg was normal after short periods of abstinence but increased after 4 weeks of abstinence (Figure 4, Tables 1 and 2). Direct comparison of LgA and ShA rats, confirmed that the metabolic activity of the Amyg in LgA rats was lower compared to ShA rats after 1 week of abstinence but became higher after 4 weeks of abstinence (Figure 5, Table 1).

Discussion

We show that voluntary intake of cocaine produces changes in basal brain metabolic activity that depends on the intensity of cocaine self-administration and on the duration of abstinence. Indeed, escalation of cocaine self-administration produces cerebral changes that are quantitatively and qualitatively different from those found after ShA cocaine self-administration. Importantly, the neuroadaptations found in this study are consistent with data described in humans and support the idea that cocaine exposure is associated with specific disruptions in interconnected meso-striato-cortical, limbic and fronto-cortical circuits that are involved in reward, motivation, salience attribution, executive control, stress reactivity and negative affect (Goldstein and Volkow, 2002; Koob and Volkow, 2010; Parvaz et al, 2011; Volkow and Baler, 2014; Volkow et al, 2003).

Longitudinal Brain Dysregulations During Abstinence from Cocaine Self-Administration

Several animal studies have investigated the changes in brain functioning associated with drug exposure (Caprioli et al, 2013; Gould et al, 2014; Hammer et al, 1993; Hanlon et al, 2013; Howell, 2008; Macey et al, 2004) but only two studies have focused on withdrawal from extended access to cocaine using brain imaging in rats (Calipari et al, 2013; Gozzi et al, 2011). Whereas those studies were limited to the early consequences of cocaine withdrawal on brain activity, here we investigated longitudinally both early and long-term changes in metabolic activity related to cocaine abstinence in rats exhibiting two different patterns of cocaine voluntary intake (stable low intake and escalated high intake). Although the direct comparison of the results is complicated because of the different experimental protocols and imaging approaches used (fMRI vs PET), a common finding in all studies is that cocaine-exposed rats displayed decreased metabolic activities in the ACC, medial prefrontal cortex and the NAc after short periods of withdrawal (Calipari et al, 2013; Gozzi et al, 2011). It is important to note that some of these adaptations were found both after limited and extended access to cocaine. However, the intensity of metabolic modification was significantly higher in LgA than in ShA. In addition, whereas some brain regions show spontaneous recovery of normal metabolic activity after protracted abstinence, in LgA rats, a few brain structures (the DStr, the ACC, the SN/VTA, the Amyg, the Hipp and the insula) show persistent changes in metabolic activity that may contribute to long-lasting risks of relapse. These results are consistent with separate previous studies that have shown that withdrawal from cocaine self-administration is associated with neuroadaptations in the DStr (Hearing et al, 2008; Pomierny-Chamiolo et al, 2015), ACC (Pomierny-Chamiolo et al, 2015; Zavala et al, 2007; Zorrilla et al, 2001), VTA (mostly with no change in the SN (Arroyo et al, 2000; Chen et al, 2008; Grimm et al, 2003; Lu et al, 2003), Amyg (Lee et al, 2013; Lu et al, 2005a; Zorrilla et al, 2001) and Hipp (Garcia-Fuster et al, 2012; Noonan et al, 2008; Pomierny-Chamiolo et al, 2015; Thompson et al, 2004). Importantly, each of these regions has been shown to be involved in cocaine craving and relapse in animal models (Cosme et al, 2015; Fuchs et al, 2005, 2006; Grimm and See, 2000; Lu et al, 2005b; McFarland and Kalivas, 2001; Torregrossa et al, 2013). Thus, these results suggest that extended access cocaine self-administration is associated with simultaneous dysregulations in interconnected networks that can lead to increased risks of relapse.

Similarities between Rat and Human Studies

Over a decade ago, on the basis of clinical and brain-imaging data in humans, Goldstein and Volkow proposed the Impaired Response Inhibition and Salience Attribution (I-RISA) model of addiction that postulates that addiction is the result of a change in the balance of different functions and in the activity of corresponding brain areas (Goldstein and Volkow, 2002). In this model, the six main functions involved in drug addiction are: (1) reward/saliency involving the NAc, the VP and the dopaminergic system; (2) memory/learning/habits involving the DStr, the Amyg and the Hipp; (3) inhibitory control/executive functions involving the prefrontal cortex and the ACC; (4) motivation/drive involving the OFC, the motor cortex and the dopaminergic system (SN/VTA); (5) interoception involving the insula and ACC and (6) aversion involving the Amyg (Goldstein and Volkow, 2002; Koob and Volkow, 2010; Parvaz et al, 2011; Volkow and Baler, 2014; Volkow et al, 2003). In particular, in this model, reductions in the activity of prefrontal and frontal cortical areas induce dysregulated top–down processes ultimately leading to difficulties in controlling and inhibiting drug-seeking behavior and increased risks of relapse (for review see Goldstein and Volkow (2002)). For example, long-term abstinent cocaine abusers (more than 6 weeks) show frontal hypo-metabolism (Volkow et al, 1992) and this disruption is supposed to favor compulsive drug taking (Volkow et al, 1993). This hypo-metabolism is associated with decreased dopamine D2 receptor availability (Volkow et al, 1990, 1993), reflecting a dysregulation of dopamine receptors following alterations in the activity of the dopaminergic system, described in detoxified cocaine abusers (Volkow et al, 1997a, 1997c). Our data demonstrate that rats that have extended access to cocaine show a profile of brain metabolic activity that is consistent with this model. In fact, we found a persistent reduction in the metabolic activity of ACC associated with an increase in the activity of the SN/VTA and a decrease in the activity of the DStr, which could be secondary to hyperactivity of the dopaminergic system and activation of dopamine receptors. In addition, after long periods of abstinence, we found that the Amyg and Hipp are hyperactive which is consistent with the critical role of memory, negative affect and stress-related process in relapse (Koob and Volkow, 2010, 2016). Finally, we found a decrease in the activity of the insula that reveals a dysregulation in the functioning of this region that has been shown to be critical for interoceptive processes and relapse (Goldstein et al, 2009).

Methodological Considerations

When interpreting the results of this study, several aspects should be considered. First, our measures reflect metabolic activity under resting states. Thus, it is possible that dysregulation in the ability of certain brain regions to respond to external or internal stimuli may have passed unnoticed in the present study. This is particularly important because most of the recent studies in humans have used fMRI approaches to highlight the changes in reactivity of the brain of individuals with addiction to several manipulations such as exposure to drug cues (Goldstein and Volkow, 2002; Volkow and Baler, 2014; Wilson et al, 2004). In addition, PET imaging allows a resolution of one millimeter (Caprioli et al, 2013). For this reason, some regions that were too small to be identified individually, such as the SN/VTA or different nuclei of the Amyg, had to be pooled. Thus, it is possible that certain changes attributed to one region may be actually due to changes in one of their sub-regions. Conversely, it is possible that lack of effects in some regions may be in part due to differences in the activity of the sub-regions. Whereas in this study we focused our attention to the abstinence phase of addiction, brain-imaging studies could be potentially used longitudinally to compare brain functioning in the same individual before, during, and after exposure to cocaine. Unfortunately, because the behavioral equipment and the PET scan were not in the same location, we were not able to perform such studies. Future studies will be needed to allow this additional important comparison. Another important consideration is that given that imaging procedure was conducted twice in each rat, the effects of repeated manipulations such as fasting, habituation, and general anesthesia may have at least in part contributed to the effects measured in the second scan. However, analysis of brain activity in control rats at different time points did not show significant effect of retesting (data not shown). A final aspect to consider is that in our study, control naive animals did not undergo surgery and were not exposed to operant cages. While we cannot rule out the possibility that these manipulations may have contributed to differences in metabolic activity measured in ShA and LgA rats vs controls, the contribution of these experimental differences is likely limited compared to exposure to cocaine. Furthermore, it should be considered that from a translational point of view, cocaine naive individuals do not share with individuals with cocaine addiction a wide variety of drug-related experiences and may indeed resemble in many ways to our naive control groups. Finally, whereas the comparison to naive control is an important aspect of this work, our main focus concerns the differences between ShA and LgA rats that had very similar experimental histories.

Conclusions

We show that extended access intake of cocaine in the escalation model (Ahmed and Koob, 1998) produce a complex regional and temporal pattern of changes in basal brain metabolic activity that are strikingly similar to those reported in humans and framed into the I-RISA model of addiction (Goldstein and Volkow, 2002). These functional dysregulations would lead to increased sensitivity to drugs and drug-related cues, difficulties in regulating emotional responses and deficits in cognitive functions that render individuals with addiction persistently vulnerable to relapse. The combination of animal models of addiction and brain imaging approaches represents a unique tool to investigate the brain mechanisms underlying the effects of novel therapeutic interventions.

Funding and disclosure

This study was funded by the INSERM, the University of Poitiers, the University of Tours and the Fondation pour la Recherche Medicale (FRM, DPA20140629806 grant to MS). C. Nicolas was a recipient of a PhD fellowship by the Poitou-Charentes region. The authors declare no conflict of interest.