The conditioning of environmental stimuli or contexts with rewarding or aversive events represents a fundamental learning process that becomes encoded in neurocircuits and subsequently drives appropriate behavior. Animals and humans perceive stimuli as rewarding or aversive; they can also learn through associative processes that avoiding or removing negative stimuli ameliorates unpleasant experiences. In subjects with drug dependence, for example, learned associations between contextual stimuli and a drug include associations linked to the reversal of the adverse withdrawal state by drug use. The ability to process meaningful stimuli related to this negative reinforcement learning is a vital neural function that is essential for maintaining stability, well-being, and survival. Therefore, how stimuli that drive behavior are represented in neurocircuits is a fundamental question.

Exposure to motivationally relevant stimuli elicits sparse patterns of neuronal activation known as neuronal ensembles or cellular assemblies, which can occur without coactivation of synaptically connected partners. Because learned associations between stimuli and subjectively rewarding or aversive experiences are a major factor in the chronically relapsing nature of compulsive alcohol seeking and use, a withdrawal-related learning (WDL) procedure was used to identify stimulus-reactive neurons that mediate the motivating effects of negative reinforcement learning. In the WDL procedure, environmental stimuli conditioned to reversal of aversive alcohol withdrawal states by alcohol self-administration (i.e., negative reinforcement) acquire conditioned incentive value and may represent a significant major factor in substance craving and relapse, leading to reward dysregulation and pathological hedonic allostasis.

The formation of sparse and distributed neuronal assemblies that encode learned associations that then mediate appetitively motivated behavior including drug-seeking responses is thought to be the basic unit of acquired learning. However, the neuronal substrates that specifically mediate the motivating effects of stimuli associated with the reversal of negative hedonic states such as dysphoria, anxiety, and sensitivity to stress after excessive or long-term substance use remain to be understood. The relevance of this understanding is illustrated by findings of a parallel behavioral study in which alcohol seeking induced by contextual stimuli associated with the reversal of adverse withdrawal states by alcohol (WDL) was qualitatively and quantitatively different from that in rats without a dependence and WDL history. Specifically, alcohol seeking in rats with WDL experience was stronger overall, and more importantly, it was compulsive in nature (i.e., resistant to punishment and increased effort requirements), whereas alcohol seeking in nondependent rats with a social drinking history was not. Therefore, using a rat reinstatement model, we sought to 1) identify stimulus-reactive neurons recruited by environmental stimuli linked to alcohol availability and consumption during withdrawal episodes following the development of alcohol dependence (negative reinforcement) versus alcohol availability in the nondependent state (positive reinforcement) and 2) establish whether alcohol seeking in rats with WDL experience recruits different neurons than in rats without this experience as well as the nature of these differences.

Methods and Materials

Animal Use and Care

All procedures were conducted in adherence to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Scripps Research. All animals were adult male Wistar rats (Charles River Laboratories) weighing approximately 450 grams.

Brains were obtained from rats trained and tested under the conditions described below in Behavioral Training.

Behavioral Training

The WDL and all other behavioral procedures were conducted as previously described. These procedures established compound contextual stimuli (i.e., both contextual stimuli and response-contingent discrete stimuli) to response-contingent alcohol availability during withdrawal in dependent rats versus alcohol availability in both nondependent and postdependent rats (Figure 1). The purpose of the WDL experimental condition (DEP-WDL, N = 13) (Figure 1, top row) was to establish the effects of these stimuli on reinstatement in the presence of motivational and environmental challenges (i.e., punishment by footshock and effort manipulations). Following ethanol (EtOH) self-administration training, rats were subjected to alcohol vapor inhalation or remained nondependent. After 3 weeks, rats were transiently removed from the vapor (or control) chambers, and after 8 hours of withdrawal, they were given the opportunity to operantly self-administer EtOH in the presence of the compound contextual stimuli and conditioned stimuli in 30-minute sessions. EtOH self-administration conditioning continued for 9 sessions, separated by 1 to 2 days during which rats remained undisturbed in the vapor chambers (or room air). Three control/comparison groups were included. First, a nondependent group (NDEP-WDL, N = 12) provided a comparison between the effects of a history of dependence and withdrawal (DEP-WDL) versus a history of only nondependence (NDEP-WDL) on stimulus-induced alcohol seeking, with all other conditions remaining equal. A no WDL (NWDL) group served the purpose of providing a control for the effects of a dependence history alone without alcohol reinforcement during withdrawal. Second, a dependent (DEP-NWDL, N = 9) control group served as a control for the effects of a history of dependence and withdrawal alone (without WDL) to establish whether such a history alone rather than WDL experience (DEP-WDL) accounts for the observed effects on stimulus-induced alcohol seeking. Third, a nondependent and no WDL (NDEP-NWDL, N = 9) group was tested in parallel with the DEP-NWDL group to provide a final comparison between the effects on alcohol seeking with a history of dependence and withdrawal alone (DEP-NWDL) versus a history of nondependence, with all other conditions remaining equal. Following dependence induction, DEP-NWDL rats were withdrawn from alcohol for 1 week and then given the opportunity to operantly self-administer EtOH in the presence of compound contextual stimuli. Paralleling the procedures for the WDL group, conditioning was conducted in 9 sessions, separated by 1 to 2 days during which rats remained undisturbed in the home cages. After completion of procedures in the dependence induction phase, rats remained in their home cages for 1 week, followed by daily reexposure to the self-administration operant chamber under extinction conditions. Subsequent tolerance of increased effort, resistance to punishment, and simple reinstatement tests were conducted across 14 days.

Figure 1. Illustration of the design including the WDL procedure. Rats were divided into 4 groups (DEP-WDL, NDEP-WDL, DEP-NWDL, NDEP-NWDL), 2 of which were subjected to alcohol dependence (DEP-WDL and DEP-NWDL) induction while the other 2 remained nondependent (NDEP-WDL and NDEP-NWDL). DEP-WDL (experimental group): history of alcohol SA, dependence, and withdrawal—WDL via repeated exposure to stimuli associated with alcohol availability during withdrawal states in the training phase. NDEP-WDL (control group 1): alcohol nondependent rats, tested in parallel with DEP-WDL rats; history of alcohol SA but no history of alcohol dependence/withdrawal; exposed to the same alcohol-associated stimuli as the DEP-WDL group but during alcohol availability in the nondependent state (i.e., no WDL experience). DEP-NWDL (control group 2): history of alcohol SA, dependence, and withdrawal—but no WDL experience (i.e., no history of stimulus presentation during alcohol availability in the training phase). NDEP-NWDL (control group 3): alcohol nondependent rats, tested in parallel with DEP-NWDL rats; history of alcohol SA; no history of alcohol dependence/withdrawal; no WDL but exposed to the same alcohol-associated stimuli as the DEP-NWDL group but during alcohol availability in the nondependent state. Control group 1 provided a comparison between the effects of a history of dependence and WDL (DEP-WDL) vs. a history of only nondependence (NDEP-WDL) on stimulus-induced alcohol seeking, with all other conditions remaining equal. Control group 2 served as a control for the effects of a history of dependence and withdrawal alone (without WDL) to establish whether such a history alone rather than WDL experience (DEP-WDL) accounts for the observed effects on stimulus-induced alcohol seeking. Control group 3 provided a final comparison between the effects on alcohol seeking of a history of dependence and withdrawal alone (DEP-NWDL) vs. a history of nondependence, with all other conditions remaining equal. DEP, dependent; ∗E, euthanasia; EtOH, ethanol; NDEP, nondependent; NWDL, no WDL; SA, self-administration; WDL, withdrawal-related learning.

Tissue Preparation

Brains from our previous study were used to analyze neuronal activation in the current study. All rats were sacrificed with CO₂ and transcardially perfused (4% paraformaldehyde in 0.1-mM sodium tetraborate) 90 minutes after reexposure to contextual stimuli in the operant chamber (without alcohol availability). This time frame allows for the visualization of c-Fos protein expression in activated neurons. Then, brains were harvested and placed in 30% sucrose before cutting 40 to 50 μm thick tissue sections using a Leica microtome. In this study, whole brains were sectioned at the same thickness, and all tissue sections for this study were collected together into appropriate wells in a 24-well plate for immunostaining. Immunolabeling was achieved with an anti-c-Fos antibody (#2250 c-Fos [9F6] rabbit monoclonal antibody; 1:5000; Cell Signaling). The c-Fos family of nuclear oncogenes includes c-Fos, c-FosB, FRA1, and FRA2. Here, c-Fos antibody was used that detects endogenous levels of total c-Fos protein. The antibody does not cross-react with other c-Fos proteins, and it was characterized by Western blot analysis. A donkey anti-rabbit Alexa Fluor 488 was used to visualize c-Fos-positive neurons (#A21206; Life Technologies). First, tissue sections were incubated in blocking solution (5% donkey serum, 0.25% triton, and 0.05% sodium azide in 0.01M phosphate-buffered saline [PBS]) for 1 hour before being left to incubate in primary antibody for 72 hours. Next, sections were rinsed 6 times in 1× PBS and then allowed to incubate in the secondary antibody (1:800) for 4 hours. Final rinses and DAPI (1:1000; Thermo Fisher Scientific) staining were achieved before the tissue was mounted on glass slides and cover slipped.

Quantitative Analysis

Imaging details are provided in the Supplement. Representative sampled neuroanatomical regions are delineated in Figure 2A–F, with Figure 2D, F showing c-Fos-expressing cortical-, amygdalar-, and paraventricular thalamic-activated neurons. c-Fos-expressing neurons within contoured regions were detected automatically with the cell detection function in NeuroInfo-rat (MBF Bioscience) (Figure 3 and Supplement). Using the c-Fos dataset from the current study, NeuroInfo-rat was built specifically for c-Fos-expressing cells in rat tissue. All automated cell detections were validated and edited by an experimenter before finalization. Three independent experimenters edited automated cell detections, and all 3 finalized cell count values were averaged. Experimenters were blinded to the treatment conditions. c-Fos-positive cell counts were normalized over the total surface area of each region of interest to obtain c-Fos density (c-Fos counts/mm²) as a proxy for neuronal activation.

Figure 2. Representative confocal images showing brain regions analyzed for c-Fos-positive stimulus-reactive neurons. (A) Low-magnification coronal section depicting the PL and IL, with outlined areas used for quantification of c-Fos-positive nuclei. (B) Low-magnification image indicating regions sampled in the dSTRI and NAc. (C) Low-magnification image showing regions sampled in the PVT, CeA, and BLA. (D) Higher magnification view of the PL and IL showing c-Fos-positive nuclei. Representative c-Fos-positive nuclei (red arrows) with approximate diameters of 10 μm in the PL/IL (E) and PVT (F). Nuclei smaller than 7 μm were excluded from quantification via automated thresholding in NeuroInfo-rat (MBF Bioscience), and thus, any small speckle was excluded from the counts. All c-Fos-positive neurons were visualized with an Alexa Fluor 488 antibody and pseudocolored for improved contrast and visibility. BLA, basolateral amygdala; CeA, central amygdala; dSTRI, dorsal striatum; IL, infralimbic cortex; NAc, nucleus accumbens; PL, prelimbic cortex; PVT, paraventricular nucleus of the thalamus.

Figure 3. Automated detection of c-Fos-positive cells using conventional and AI-enhanced methods. (A) c-Fos-expressing neurons identified using a standard cell detection algorithm available in NeuroInfo-rat (MBF Bioscience). Details are provided in the Supplement. (B) Improved detection using an AI-based method incorporating machine learning, which reduces false positives from edge artifacts (white arrows) and enhances detection of low-intensity (dimmer) c-Fos-positive nuclei. Note the 2 green arrows in (A) pointing to c-Fos-positive cells (>7 μm) that were not detected by the standard cell detection algorithm but were captured using AI-assisted image analysis algorithm integrated within the NeuroInfo-rat software platform, which incorporates machine learning principles for cell recognition based on pixel intensity, morphology, and pattern recognition in (B). Red arrows represent c-Fos-positive neurons, which were visualized by an Alexa Fluor 488 antibody and pseudocolored yellow for improved contrast and visibility. AI, artificial intelligence.

Statistical Analysis

The design consisted of 4 experimental groups: 1) DEP-WDL (N = 13), 2) NDEP-WDL (N = 12), 3) DEP-NWDL (N = 9), and 4) NDEP-NWDL (N = 9) as described above. To study the effects of the respective contextual stimulus exposure on neuronal activity, c-Fos density was used as the dependent variable. While the assumptions that underlie parametric tests were met for independent random sampling, normal distributions and homogeneity of variance were not observed in these data. The Shapiro-Wilk test indicated that the distribution of c-Fos-positive values in the WDL group deviated significantly from normality (W = 0.90, p ≤ .001) (see histograms in the Supplement). Moreover, Levene’s test for homogeneity of variance revealed a significant difference in variances across groups (F_3,39 = 3.32, p ≤ .029). For these reasons, the treatment effects on c-Fos density were analyzed using nonparametric tests, and, where appropriate, these were followed by post hoc comparisons with Bonferroni corrections.

Results

Overall Neuronal Activation Associated With Alcohol Seeking Is Increased in the DEP-WDL Group Relative to Nondependent Groups

To examine whether WDL stimulus exposure had an effect on neuronal activity, c-Fos density was quantified in key brain regions with an established role in alcohol seeking and craving. A Kruskal-Wallis test was conducted to examine the differences in c-Fos density among the following 4 groups: DEP-WDL, NDEP-WDL, DEP-NWDL, and NDEP-NWDL. The results of this test indicated a significant difference in c-Fos density among the treatment groups (H₃ = 7.5, p ≤ .040). Post hoc analysis using pairwise comparisons revealed an increased overall density of activated neurons in the DEP-WDL group compared with both nondependent groups (Dunn’s test with Bonferroni correction, DEP-WDL vs. NDEP-WDL p ≤ .014, DEP-WDL vs. NDEP-NWDL p ≤ .024) (Figure 4A). No significant difference was found between the DEP-WDL and DEP-NWDL groups (p > .05, nonsignificant [n.s.]) or between the DEP-NWDL and NDEP-NWDL groups (p > .05, n.s.).

Figure 4. (A) Plot summarizing the overall c-Fos-activated neurons for all 4 experimental groups: DEP-WDL (N = 13), NDEP-WDL (N = 12), DEP-NWDL (N = 9), and NDEP-NWDL (N = 9). Inset shows group conditions. (B) Measurement of c-Fos activity in the DS (N = 15), PL (N = 14), and PVT (N = 15) revealed an increased number of activated neurons in the PVT of DEP-WDL rats (Kruskal-Wallis, Dunn’s test with Bonferroni correction p ≤ .019). Increased neuronal activity in the DS was also present in both dependent groups (DEP-WDL, DEP-NWDL) relative to the nondependent (NDEP-WDL, NDEP-NWDL) groups (Kruskal-Wallis, Dunn’s test with Bonferroni correction p ≤ .046). Bars represent mean and SEM. Asterisks denote statistical significance. DEP, dependent; DS, dorsal striatum; EtOH, ethanol; NDEP, nondependent; NWDL, no WDL; PL, prelimbic cortex; PVT, paraventricular nucleus of the thalamus; SA, self-administration; WDL, withdrawal-related learning.

Withdrawal Learning Experience–Dependent Neuronal Activation Is Increased in the Paraventricular Nucleus of the Thalamus and Dorsal Striatum

To better understand the neuroanatomical localization of these stimulus-reactive neurons, c-Fos density was initially analyzed in 3 brain regions with an established role in alcohol/drug addiction: the dorsal striatum (DS), prelimbic cortex (PL), and paraventricular nucleus of the thalamus (PVT). A Kruskal-Wallis test was conducted to examine the differences in neuronal activation among the 4 treatment groups and across the 3 brain regions. The results revealed a significant treatment group difference in c-Fos density in the DS and PVT (DS: H₃ = 8.5, p ≤ .010; PVT: H₃ = 10.5, p ≤ .012) but not in the PL (H₃ = 4.4, p > .05, n.s.). Multiple pairwise comparisons using Dunn’s test with Bonferroni correction revealed that the DEP-WDL group showed an increased density of activated neurons specifically in the PVT compared with all other treatment groups (DEP-WDL vs. NDEP-WDL p ≤ .003; DEP-WDL vs. DEP-NWDL p ≤ .045; DEP-WDL vs. NDEP-NWDL p ≤ .011) (Figure 4B). In addition, increased neuronal activation was observed in the DS of DEP-WDL rats relative to rats in the nondependent groups, NDEP-WDL and NDEP-NWDL (p ≤ .008 and p ≤ .003, respectively); however, no difference was found between the DEP-WDL and DEP-NWDL groups for this striatal region (Figure 4B). Because these data revealed that WDL experience was a critical factor in the observed neuronal activity increases, especially in the PVT, we focused the remaining analyses on activity-dependent neuronal changes between the DEP-WDL and NDEP-WDL groups.

For a more comprehensive analysis across different neuroanatomical brain regions, c-Fos counts were extended to 4 additional brain regions with an established role in drug addiction including the infralimbic cortex (IL), nucleus accumbens (NAc), central nucleus of the amygdala (CeA), and basolateral amygdala (BLA) (Figure 5). To gain further insights into the functional changes across neuroanatomical regions and the recruitment of c-Fos-positive neurons as a result of the respective learning experience during the withdrawal state in dependent animals versus nondependent controls, a Mann-Whitney U test was used to compare neuronal activity in the DEP-WDL and NDEP-WDL groups across all 8 brain regions: the DS, PL, PVT, IL, NAc, CeA, BLA, and lateral hypothalamus (LH). The results revealed a significant increase in the density of activated neurons between the DEP-WDL and NDEP-WDL groups following WDL experience in the PVT, DS, and CeA (PVT: U = 1, p ≤ .021; DS: U = 0, p ≤ .014; CeA: U = 0, p ≤ .021) but no difference in c-Fos density for the PL, IL, NAc, BLA, or LH between the DEP-WDL and NDEP-WDL groups (Figure 5).

Figure 5. Recruitment of c-Fos-positive neurons in the DS (N = 9), PVT (N = 8), and CeA (N = 8), but not in the IL (N = 9), PL (N = 8), NAc (N = 11), BLA (N = 11), or LH, following WDL (Mann-Whitney U; DS p ≤ .016, PVT p ≤ .029, CeA p ≤ .29). Asterisks denote statistical significance. BLA, basolateral amygdala; CeA, central amygdala; DEP, dependent; DS, dorsal striatum; IL, infralimbic cortex; LH, lateral hypothalamus; NAc, nucleus accumbens; NDEP, nondependent; NWDL, no WDL; PL, prelimbic cortex; PVT, paraventricular nucleus of the thalamus; WDL, withdrawal-related learning.

Discussion

Differential Neuronal Recruitment in Alcohol-Seeking Rats With Versus Without a History of WDL

Our previous behavioral findings established that environmental stimuli conditioned specifically to amelioration of withdrawal (i.e., the negative reinforcing effects of alcohol) exert more powerful control over alcohol seeking than stimuli conditioned only to the positive reinforcing effects of alcohol. More specifically, reinstatement of alcohol seeking in rats with a history of WDL was not only greater but also impervious to motivational challenges including punishment of responding by electrical footshock and increased effort requirements than behavior induced by stimuli conditioned to alcohol in the nondependent state. The current findings extend these behavioral observations to the neuroanatomical level and implicate a role for clusters of neurons in the PVT, CeA, and DS in the potent motivating effects of WDL.

The findings confirm significant differences in c-Fos density across brain regions activated during context-induced compulsive alcohol seeking in rats with a WDL history versus rats without such a history that do not show compulsive behavior. More specifically, c-Fos-positive neurons were recruited in the PVT, DS, and CeA of WDL animals in which contextual stimuli were associated with the reversal of adverse withdrawal effects (i.e., negative hedonic states) (Figure 4, Figure 5, Figure 6). In contrast, exposure to stimuli associated with the hedonically positive aspects of alcohol consumption (i.e., in nondependent or postdependent rats without WDL experience) did not recruit the same number of c-Fos-positive neurons in the PVT and produced only mild, although significant, activation in the DS of dependent (DEP-NWDL) compared with nondependent (NDEP-NWDL) animals. These observations support the hypothesis that the conditioned effects of contextual stimuli associated with the reversal of withdrawal distress are differentially represented in the brain compared with the effects of stimuli associated with all other learning conditions in both alcohol-dependent and nondependent rats.

Figure 6. Neuroanatomical summary illustration of the recruitment of active neurons associated with context-induced alcohol seeking in rats with WDL experience and dependent drinking (B) vs. alcohol seeking induced by the same context in nondependent rats with a social drinking history (A). BLA, basolateral amygdala; CeA, central amygdala; DEP, dependent; DS, dorsal striatum; IL, infralimbic cortex; LH, lateral hypothalamus; NAc, nucleus accumbens; NDEP, nondependent; PL, prelimbic cortex; PVT, paraventricular nucleus of the thalamus; WDL, withdrawal-related learning.

In the PVT, increased c-Fos density was observed only in the WDL group but not in 3 relevant control groups (NDEP-WDL, DEP-NWDL, NDEP-NWDL). This prominent neuronal activation pattern found exclusively in the PVT of the WDL group suggests that this nucleus plays an important role in 1) the learning or acquisition of the negative contingency between alcohol consumption and the dysphoric effects of withdrawal and 2) the resulting development of compulsive drug seeking (Figures 4B, 5, and 6). The PVT is a key hub for neural circuits implicated in drug addiction. Moreover, this nucleus has an established role in emotional responses to anxiety and stress. Stress increases neuronal activity in the PVT, and the PVT was shown to be critical for stress-induced reinstatement of oxycodone seeking. The dysphoric effects of alcohol withdrawal are strongly associated with stress. Therefore, this finding is consistent with a major role of withdrawal stress in WDL learning and compulsive drug seeking following WDL acquisition. The stress-sensitive PVT area projects axonal fibers to the CeA, a region associated with negative emotion, stress, alcohol dependence, and particularly the aversive effects of alcohol withdrawal. Not unexpectedly, therefore, exposure to the WDL-associated stimulus context resulted in enhanced recruitment of c-Fos-positive neurons in the CeA, suggestive of increased CeA neuronal activity compared with NDEP-WDL rats, in which the stimulus context was associated only with the hedonically positive experience of alcohol consumption (Figure 5). Given that chemogenetic inhibition of the PVT to CeA projection alleviates stress responses, the current findings confirm a possibly major role of the CeA-projecting PVT system in behavioral responses to stress.

c-Fos-Positive Neurons Are Recruited During Alcohol Seeking in the DS of Dependent Animals

Independent of the WDL experience, exposure to stimuli conditioned to the hedonically negative aspects of alcohol withdrawal in dependent rats resulted in increased striatal c-Fos density relative to nondependent rats. The development of dependence is associated with the emergence of habitual behavior that is thought to be mediated by striatal brain regions, in particular the DS . This enhanced activation of neurons in the DS of the DEP-WDL group may be explained by a progressive engagement of dorsal striatal regions in habit formation that emerges during the development of substance dependence as proposed by the spiraling hypothesis [i.e., neural processes through which the ventral striatum comes to exert control over dorsal striatal processes mediated by so-called spiraling striato-nigro-striatal circuitry]. A pattern of activated neurons in the DS was also noted in the brains of postdependent (DEP-NWDL) animals (Figure 4B). These rats had a history of dependence but without WDL experience and were trained to associate the stimulus context with alcohol availability following completion of alcohol withdrawal. The neuronal activation in the DS of these DEP-NWDL animals was smaller than that in the WDL history group but significantly different from nondependent controls (NDEP-NWDL). Thus, DS neurons were recruited and became activated to some degree not only as a consequence of the WDL experience or negative reinforcement learning but also other dependence associated or experiential factors independent of WDL.

Addiction as a State of Reward Dysregulation and Hedonic Allostasis

It can be assumed that rats in all nondependent (NDEP-WDL) groups experienced a positive hedonic state during alcohol self-administration. In contrast, rats in the WDL group (DEP-WDL) experienced a profound negative hedonic state during withdrawal. Upon stimulus reexposure, increased neuronal activation occurred exclusively in the PVT of DEP-WDL animals (Figure 4B). Stimuli-reactive neurons in the PVT were associated specifically with WDL experience and likely directly linked to the reversal of the stressful aspect of alcohol withdrawal in rats with WDL experience. By contrast, the DS ensemble of active neurons, although differential, was observed in both dependent animals with WDL experience (DEP-WDL) and animals with a history of dependence but not WDL (DEP-NWDL). Considering our previous observations, these new findings suggest that the PVT may have a broader role in behavior motivated by hedonically negative states beyond alcohol-seeking behavior by providing a neuroanatomical hub for the development of hedonic allostasis, a chronic deviation from normal hedonic homeostasis and a state associated with reward dysregulation, stress responses, abnormal motivation, and addiction.

The current conditioning regimen that elicited behavior motivated by both positive and negative hedonic states may provide a tool for investigating aspects of the opponent process hypothesis of motivation and the development of hedonic allostasis. According to these hypotheses, the experience of a positive hedonic state results in a transient opposing negative hedonic state until equilibrium or hedonic homeostasis is restored. Over time, repeated experience of hedonically positive stimuli or substances (e.g., alcohol, opioids, sugar, skydiving, sexual activity, nicotine, gambling) leads to an allostatic state in which both the initial positive and resulting opponent negative processes still occur but eventually drop below the baseline level of normal hedonic homeostasis into negative hedonic territory, i.e., hedonic allostasis. Although these hypotheses of motivation are widely accepted, the corresponding neuronal mechanisms that mediate these processes remain to be established. Based on the current findings, one may speculate that the PVT plays a significant role in opponent processes and the development of hedonic allostasis. The prominent activation of the PVT following WDL stimulus exposure implicates this nucleus in mediating the powerful motivating effects of stimuli linked to alcohol (or other reinforcers) that reverse negative hedonic states—negative reinforcement—and thereby drive further alcohol consumption, exacerbating the negative hedonic process that ultimately conveys incremental hedonic valence to alcohol, with eventual progression toward an allostatic state. Given that marked neuronal activation in the PVT was associated exclusively with alcohol seeking elicited by the WDL-paired stimuli, this neuroanatomical region may provide a prime target to study the neuronal mechanisms that mediate opponent processes and the development of hedonic allostasis.

Limitations and Their Implications

The results are limited to information from male rats and require extension to female rats. The study was also somewhat constrained by the amount of tissue available from the original behavioral study. Finally, neuronal activation to identify stimulus-reactive neurons was measured by c-Fos protein expression, a marker for recent neuronal activity. However, populations of c-Fos-positive neurons can become activated by contextual stimuli through a mechanism that involves the expansion of multisynaptic boutons independently of the coactivation state of postsynaptic partners. That is, stimulus-reactive presynaptic boutons of projection neurons may recruit other neurons that were not engaged during the learning task to drive subsequent behavior. Such a mechanism could contribute to enhanced network-level responsiveness in the PVT of DEP-WDL rats by recruiting additional neurons, thus extending activation beyond traditional coactive engram ensembles. This activation might also have been influenced by the animal’s previous behavioral experiences, resulting in spontaneous c-Fos expression around the time of compound contextual stimulus presentation. Therefore, the stimulus-evoked signal is likely superimposed on a background level of c-Fos activity shaped by the animal’s unique behavioral trajectory throughout the experiment.

Conclusions

The findings suggest that the activated neurons identified here serve as a neural substrate for compulsive drug seeking resulting from 1) negative reinforcement learning associated with hedonic allostasis (PVT), 2) the development of habitual behavior over the progression of dependence (DS), and 3) stress memories associated with the stimulus context in which withdrawal and reversal of withdrawal were experienced (CeA). To confirm a role for these stimulus-activated neurons specifically in mediating WDL experience, confirmatory experiments will be required such as the silencing of each reactive group of neurons with cell type–specific methods. Finally, these findings undoubtedly have implications for maladaptive behavior linked to reward dysregulation and the processing of emotionally salient stimuli that drives behavior beyond substance use disorders. Research to establish the neurobiological alterations that link the conditioned incentive value of substance-associated contexts and their role in exacerbating drug seeking will also require extension to other classes of maladaptive behavior. These include, but are not limited to, systems that regulate fear conditioning, anxiety disorders, traumatic avoidance learning, and possibly predatory behavior.

Understanding Learned Responses to Stimuli

The way the brain learns to connect environmental cues or situations with rewards or unpleasant events is a basic process. This learning is stored in brain circuits and then guides behavior. Both animals and humans recognize stimuli as either rewarding or negative. They also learn that avoiding or removing negative stimuli can reduce uncomfortable experiences. For instance, individuals with drug dependence learn to associate certain places or situations with using a drug to stop the unpleasant feelings of withdrawal. The brain's ability to process these important cues, which relate to negative reinforcement learning, is crucial for stability, well-being, and survival. Therefore, understanding how the brain represents stimuli that drive behavior is a fundamental area of study.

When exposed to important stimuli that drive motivation, specific groups of neurons, known as neuronal ensembles or cellular assemblies, become active. This activation can happen even without direct connections between these neurons. Previous research has shown that learned connections between stimuli and rewarding or negative experiences are a major reason for the ongoing cycle of alcohol seeking and use. To investigate this, a procedure called withdrawal-related learning (WDL) was used. This procedure helps identify neurons that react to stimuli and are involved in the motivating effects of negative reinforcement learning. In the WDL procedure, environmental cues that have been linked to reducing severe alcohol withdrawal symptoms through alcohol self-administration (a process of negative reinforcement) gain a strong incentive value. These cues can be a significant factor in substance craving and relapse, which can lead to problems with reward processing and a chronic state of imbalance in pleasure.

The formation of these distinct, spread-out groups of neurons that store learned connections and then drive motivated behaviors, including drug seeking, is thought to be the basic unit of learned actions. However, the specific brain structures that cause the motivating effects of stimuli linked to reducing negative emotional states—such as sadness, anxiety, and stress sensitivity after heavy or long-term substance use—are not fully understood. The importance of this knowledge is highlighted by studies showing that alcohol seeking caused by environmental cues linked to withdrawal relief was different in rats with a history of dependence and WDL compared to those without. Specifically, rats with WDL experience showed stronger alcohol seeking, and more importantly, this behavior was compulsive; it resisted punishment and increased effort. In contrast, alcohol seeking in rats without dependence but with a history of social drinking was not compulsive. This study aimed to identify the specific neurons activated by environmental cues linked to alcohol during withdrawal (negative reinforcement) versus alcohol availability in a non-dependent state (positive reinforcement). It also sought to determine if alcohol seeking in rats with WDL experience activates different neurons compared to rats without this experience, and to understand the nature of these differences.

Methods and Materials

Animal Use and Care

All animal procedures followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and received approval from the Institutional Animal Care and Use Committee of Scripps Research. All animals were adult male Wistar rats, each weighing about 450 grams. Brains for analysis were obtained from rats previously trained and tested under specific behavioral conditions.

Behavioral Training

The WDL and other behavioral procedures were conducted as described in earlier research. These methods established complex environmental cues (both general context and specific cues) linked to alcohol availability during withdrawal in dependent rats, compared to alcohol availability in both non-dependent and post-dependent rats. The goal of the WDL experimental condition (DEP-WDL) was to observe how these cues affected alcohol seeking, even when faced with motivational and environmental challenges like foot shock punishment and increased effort.

After initial alcohol self-administration training, rats either underwent alcohol vapor inhalation to induce dependence or remained non-dependent. Three weeks later, rats were temporarily removed from the vapor chambers. After 8 hours of withdrawal, they had opportunities to self-administer alcohol in the presence of the established cues during 30-minute sessions. This conditioning continued for nine sessions, with 1 to 2 days of rest in between, during which rats stayed in their vapor chambers or room air.

Three control groups were included for comparison. The first was a non-dependent group (NDEP-WDL), which allowed for comparison between rats with a history of dependence and withdrawal (DEP-WDL) versus those with only a non-dependent history (NDEP-WDL) regarding stimulus-induced alcohol seeking, with all other factors being the same. A "no WDL" (NWDL) group served as a control for the effects of dependence alone, without alcohol reinforcement during withdrawal. The second was a dependent control group (DEP-NWDL), which helped determine if a history of dependence and withdrawal alone, rather than the WDL experience, accounted for the observed effects on alcohol seeking. The third was a non-dependent and no WDL group (NDEP-NWDL), tested alongside the DEP-NWDL group, providing a final comparison between alcohol seeking in rats with a history of dependence and withdrawal alone (DEP-NWDL) versus those with a history of non-dependence.

Following dependence induction, DEP-NWDL rats underwent 1 week of alcohol withdrawal before having opportunities to self-administer alcohol in the presence of compound contextual stimuli. Similar to the WDL group, conditioning involved nine sessions, separated by 1 to 2 days of rest in home cages. After the dependence induction phase, rats remained in their home cages for 1 week. This was followed by daily re-exposure to the self-administration chamber under extinction conditions (without alcohol). Subsequent tests for tolerance to increased effort, resistance to punishment, and simple reinstatement were conducted over 14 days.

Tissue Preparation

Brains from a previous study were used to analyze neuronal activation. All rats were humanely euthanized and perfused 90 minutes after re-exposure to contextual stimuli in the operant chamber (without alcohol). This timing allows for the visualization of c-Fos protein expression, a marker for recently activated neurons. Brains were then removed and stored in sucrose before being cut into 40 to 50 μm thick sections using a microtome. All sections were collected for immunostaining. Immunolabeling involved an anti-c-Fos antibody, which detects natural levels of total c-Fos protein and does not react with other c-Fos family members. A donkey anti-rabbit Alexa Fluor 488 antibody was used to visualize c-Fos-positive neurons. Tissue sections were first incubated in a blocking solution for 1 hour, then in the primary antibody for 72 hours. After rinsing, sections were incubated in the secondary antibody for 4 hours. Final rinses and DAPI staining were performed before mounting tissues onto glass slides.

Quantitative Analysis

For imaging details, other research should be consulted. Representative sampled brain regions were analyzed, showing c-Fos-expressing activated neurons in cortical, amygdalar, and paraventricular thalamic areas. Neurons expressing c-Fos within defined regions were automatically detected using NeuroInfo-rat software. This software was specifically designed for c-Fos-expressing cells in rat tissue using data from the current study. All automated cell detections were verified and adjusted by an experimenter. Three independent experimenters edited the automated cell detections, and their final cell count values were averaged. The experimenters were unaware of the treatment conditions. The counts of c-Fos-positive cells were normalized by the total surface area of each region of interest to calculate c-Fos density (c-Fos counts per mm²), which serves as a measure of neuronal activation.

Statistical Analysis

The study included four experimental groups: DEP-WDL, NDEP-WDL, DEP-NWDL, and NDEP-NWDL. To investigate how exposure to specific contextual stimuli affected neuronal activity, c-Fos density was used as the primary outcome measure. While the study met assumptions for independent random sampling, the data did not show normal distributions or consistent variance. A Shapiro-Wilk test indicated that the distribution of c-Fos-positive values in the WDL group was not normal. Additionally, Levene’s test for homogeneity of variance showed a significant difference in variances across groups. Because of these findings, the effects of treatment on c-Fos density were analyzed using non-parametric statistical tests. Where appropriate, these analyses were followed by post hoc comparisons with Bonferroni corrections.

Results

Overall Neuronal Activation Associated With Alcohol Seeking Is Increased in the DEP-WDL Group Relative to Nondependent Groups

To determine if WDL stimulus exposure affected neuronal activity, c-Fos density was measured in key brain regions known to be involved in alcohol seeking and craving. A Kruskal-Wallis test showed a significant difference in c-Fos density among the four treatment groups. Further analysis using pairwise comparisons indicated that the DEP-WDL group had a higher overall density of activated neurons compared to both non-dependent groups (NDEP-WDL and NDEP-NWDL). No significant differences were found between the DEP-WDL and DEP-NWDL groups, or between the DEP-NWDL and NDEP-NWDL groups.

Withdrawal Learning Experience–Dependent Neuronal Activation Is Increased in the Paraventricular Nucleus of the Thalamus and Dorsal Striatum

To better pinpoint the brain areas where these stimulus-reactive neurons are located, c-Fos density was first analyzed in three brain regions important in alcohol/drug addiction: the dorsal striatum (DS), prelimbic cortex (PL), and paraventricular nucleus of the thalamus (PVT). A Kruskal-Wallis test revealed significant differences in c-Fos density in the DS and PVT among the four treatment groups, but not in the PL. Multiple pairwise comparisons showed that the DEP-WDL group had a higher density of activated neurons specifically in the PVT compared to all other treatment groups. Additionally, increased neuronal activation was observed in the DS of DEP-WDL rats compared to rats in the non-dependent groups (NDEP-WDL and NDEP-NWDL). However, there was no difference in the DS between the DEP-WDL and DEP-NWDL groups. Given that these data indicated WDL experience was a crucial factor in the observed increases in neuronal activity, especially in the PVT, the remaining analyses focused on activity-dependent neuronal changes between the DEP-WDL and NDEP-WDL groups.

For a more complete analysis across different brain regions, c-Fos counts were extended to four additional areas known to be involved in drug addiction: the infralimbic cortex (IL), nucleus accumbens (NAc), central nucleus of the amygdala (CeA), and basolateral amygdala (BLA). To further understand the functional changes and the recruitment of c-Fos-positive neurons resulting from the specific learning experience during withdrawal in dependent animals versus non-dependent controls, a Mann-Whitney U test was used. This test compared neuronal activity in the DEP-WDL and NDEP-WDL groups across all eight brain regions: the DS, PL, PVT, IL, NAc, CeA, BLA, and lateral hypothalamus (LH). The results showed a significant increase in the density of activated neurons in the PVT, DS, and CeA for the DEP-WDL group compared to the NDEP-WDL group following WDL experience. No differences in c-Fos density were found for the PL, IL, NAc, BLA, or LH between these two groups.

Discussion

Differential Neuronal Recruitment in Alcohol-Seeking Rats With Versus Without a History of WDL

Previous behavioral findings demonstrated that environmental cues specifically linked to easing withdrawal symptoms (the negative reinforcing effects of alcohol) have a more powerful influence on alcohol seeking than cues only linked to the positive reinforcing effects of alcohol. Specifically, when alcohol seeking returned in rats with a history of WDL, it was not only stronger but also resistant to challenges like foot shock punishment and increased effort. This study extends these behavioral observations to the brain level, suggesting that groups of neurons in the PVT, CeA, and DS play a role in the strong motivating effects of WDL.

The findings confirm notable differences in the density of c-Fos-positive neurons across brain regions activated during context-induced compulsive alcohol seeking in rats with a WDL history compared to rats without such a history, who do not exhibit compulsive behavior. Specifically, c-Fos-positive neurons were activated in the PVT, DS, and CeA of WDL animals where environmental cues were linked to reversing negative withdrawal effects. In contrast, exposure to cues associated with the pleasurable aspects of alcohol consumption (in non-dependent or post-dependent rats without WDL experience) did not activate the same number of c-Fos-positive neurons in the PVT. It only produced mild, though significant, activation in the DS of dependent (DEP-NWDL) rats compared to non-dependent (NDEP-NWDL) animals. These observations support the idea that the learned effects of environmental cues associated with relieving withdrawal distress are represented differently in the brain compared to the effects of cues associated with other learning conditions in both alcohol-dependent and non-dependent rats.

c-Fos-Positive Neurons Are Recruited During Alcohol Seeking in the DS of Dependent Animals

Regardless of WDL experience, exposure to stimuli linked to the negative aspects of alcohol withdrawal in dependent rats led to increased c-Fos density in the striatum compared to non-dependent rats. The development of dependence is associated with the emergence of habitual behavior, believed to involve striatal brain regions, particularly the DS. This increased activation of neurons in the DS of the DEP-WDL group might be explained by a gradual involvement of dorsal striatal areas in habit formation as substance dependence progresses. This concept is supported by the "spiraling hypothesis," which suggests that brain processes in the ventral striatum eventually influence dorsal striatal processes. A pattern of activated neurons in the DS was also observed in the brains of post-dependent (DEP-NWDL) animals. These rats had a history of dependence but no WDL experience and were trained to associate the stimulus context with alcohol availability after completing alcohol withdrawal. The neuronal activation in the DS of these DEP-NWDL animals was less pronounced than in the WDL history group but significantly different from non-dependent controls (NDEP-NWDL). This suggests that DS neurons were activated to some extent not only due to WDL experience or negative reinforcement learning but also other dependence-related or experiential factors independent of WDL.

Addiction as a State of Reward Dysregulation and Hedonic Allostasis

Rats in all non-dependent (NDEP-WDL) groups likely experienced pleasure during alcohol self-administration. In contrast, rats in the WDL group (DEP-WDL) experienced a significant negative emotional state during withdrawal. When re-exposed to the stimuli, increased neuronal activation occurred only in the PVT of DEP-WDL animals. Stimuli-reactive neurons in the PVT were specifically linked to WDL experience and likely directly related to relieving the stressful aspects of alcohol withdrawal in rats with WDL experience. In contrast, while the active neuron group in the DS showed differences, it was observed in both dependent animals with WDL experience (DEP-WDL) and animals with a history of dependence but not WDL (DEP-NWDL).

Considering previous observations, these new findings suggest that the PVT may have a broader role in behaviors motivated by negative emotional states beyond alcohol seeking. It may serve as a brain center for the development of "hedonic allostasis," a chronic shift from normal pleasure balance, leading to problems with reward, stress responses, abnormal motivation, and addiction.

The current conditioning method, which triggered behavior motivated by both positive and negative emotional states, could be a tool for studying aspects of the "opponent process hypothesis" of motivation and the development of hedonic allostasis. According to these ideas, experiencing a positive emotional state leads to a temporary opposing negative emotional state until balance is restored. Over time, repeated experiences with pleasurable stimuli or substances cause a state where both the initial positive and the opposing negative processes still occur but eventually drop below normal levels of pleasure, leading to hedonic allostasis. Although these theories of motivation are widely accepted, the brain mechanisms mediating these processes are still unknown. Based on current findings, the PVT may play a significant role in these opposing processes and the development of hedonic allostasis. The strong activation of the PVT following WDL stimulus exposure suggests this brain region is involved in mediating the powerful motivating effects of stimuli linked to alcohol (or other rewards) that reverse negative emotional states—a process known as negative reinforcement. This drives further alcohol consumption, worsening the negative emotional process and eventually contributing to an allostatic state. Since significant neuronal activation in the PVT was exclusively associated with alcohol seeking triggered by WDL-paired stimuli, this brain region could be a prime target for studying the neuronal mechanisms underlying opponent processes and the development of hedonic allostasis.

Limitations and Their Implications

The findings are limited to information from male rats and need to be extended to female rats. The study was also somewhat constrained by the amount of brain tissue available from the original behavioral study. Finally, neuronal activation was measured by c-Fos protein expression, which indicates recent neuronal activity. However, groups of c-Fos-positive neurons can be activated by environmental cues through a mechanism that involves the expansion of multisynaptic connections, even without simultaneous activation of connected partners. This means that stimulus-reactive connections from projection neurons might recruit other neurons not engaged during the learning task to drive subsequent behavior. Such a mechanism could contribute to increased network-level responsiveness in the PVT of DEP-WDL rats by recruiting additional neurons, extending activation beyond typical active groups of neurons. This activation might also have been influenced by the animal's prior behavioral experiences, resulting in spontaneous c-Fos expression around the time of compound contextual stimulus presentation. Therefore, the signal triggered by the stimulus is likely combined with a baseline level of c-Fos activity shaped by each animal's unique behavioral history throughout the experiment.

Conclusions

The findings suggest that the activated neurons identified in this study serve as the brain basis for compulsive drug seeking. This behavior results from several factors: negative reinforcement learning linked to hedonic allostasis (PVT), the development of habitual behavior during the progression of dependence (DS), and stress memories associated with the environmental context where withdrawal and its relief were experienced (CeA). To confirm the specific role of these stimulus-activated neurons in mediating WDL experience, further experiments would be required, such as silencing each group of reactive neurons using cell type-specific methods. Finally, these findings have implications for problematic behaviors related to reward dysregulation and the processing of emotionally significant stimuli that drive behavior beyond substance use disorders. Research to identify the brain changes that connect the learned value of substance-related contexts and their role in worsening drug seeking will also need to be expanded to other types of problematic behaviors. These include, but are not limited to, systems that control fear conditioning, anxiety disorders, traumatic avoidance learning, and possibly predatory behavior.

Summary

This study investigated how the brain processes learning related to avoiding unpleasant experiences, particularly in the context of alcohol dependence. Researchers explored how specific environmental cues, when associated with relief from alcohol withdrawal (a process called negative reinforcement), trigger distinct patterns of brain activity that drive compulsive alcohol seeking. The primary goal was to identify which groups of neurons become active in response to these cues in rats with and without a history of alcohol dependence and withdrawal-related learning (WDL). The findings suggest that certain brain regions, like the paraventricular nucleus of the thalamus (PVT), dorsal striatum (DS), and central amygdala (CeA), are specifically involved in the powerful drive for alcohol seeking linked to avoiding withdrawal.

Methods and Materials

Animal Use and Care

All experimental procedures followed the National Institutes of Health guidelines for animal care and were approved by the Institutional Animal Care and Use Committee. Adult male Wistar rats, weighing about 450 grams, were used in the study. Brains were collected from rats that had completed behavioral training and testing.

Behavioral Training

The study utilized a withdrawal-related learning (WDL) procedure to examine how environmental cues associated with alcohol availability influence alcohol seeking in dependent rats versus non-dependent rats. Four groups of rats were established:

1. Dependent-WDL (DEP-WDL): Rats became alcohol dependent and learned to associate specific environmental cues with receiving alcohol, which helped relieve withdrawal symptoms (negative reinforcement).

2. Non-Dependent-WDL (NDEP-WDL): These non-dependent rats experienced the same cues associated with alcohol availability, but without withdrawal symptoms. This group served to compare effects when there was no history of dependence or withdrawal.

3. Dependent-No WDL (DEP-NWDL): These rats were alcohol dependent but did not experience the learning phase where cues were linked to alcohol availability during withdrawal. This group helped determine if dependence alone, without WDL, influenced alcohol seeking.

4. Non-Dependent-No WDL (NDEP-NWDL): These were non-dependent rats with no WDL experience, serving as a general control.

After dependence induction (or control conditions), rats underwent training sessions where they could self-administer alcohol in specific environmental contexts. Following these phases, rats were tested for their alcohol-seeking behavior under various motivational challenges, including increased effort and punishment.

Tissue Preparation

Brains from previous studies were used for this analysis. Rats were humanely euthanized and their brains prepared 90 minutes after re-exposure to the environmental cues. This timeframe allows for the detection of c-Fos protein, a marker of recent neuronal activity. Brains were then preserved, sliced, and stained using an antibody that specifically identifies c-Fos protein. This staining process made active neurons visible for analysis.

Quantitative Analysis

Specific brain regions known to be involved in addiction, such as parts of the cortex, amygdala, and thalamus, were selected for analysis. The number of c-Fos-expressing neurons within these regions was automatically detected using specialized software, with human review and editing to ensure accuracy. The density of c-Fos-positive cells (number of cells per square millimeter) was calculated to represent the level of neuronal activation in each region. To avoid bias, the experimenters who analyzed the brain tissue did not know which treatment group each sample belonged to.

Statistical Analysis

The study involved four experimental groups. The researchers examined c-Fos density as the main measure of neuronal activity. Since the data did not meet the assumptions required for standard statistical tests (e.g., normal distribution), nonparametric statistical methods were used. These methods, along with adjustments for multiple comparisons, helped identify significant differences in neuronal activity among the groups.

Results

Overall Neuronal Activation Associated With Alcohol Seeking Is Increased in the DEP-WDL Group Relative to Nondependent Groups

When examining overall neuronal activity across various brain regions, a significant difference was found among the four treatment groups. Specifically, the DEP-WDL group, which had experienced withdrawal-related learning, showed a higher overall density of active neurons compared to both non-dependent groups (NDEP-WDL and NDEP-NWDL). No significant differences in overall neuronal activity were observed between the two dependent groups (DEP-WDL and DEP-NWDL) or between the two non-dependent groups (DEP-NWDL and NDEP-NWDL).

Withdrawal Learning Experience–Dependent Neuronal Activation Is Increased in the Paraventricular Nucleus of the Thalamus and Dorsal Striatum

To pinpoint the location of these active neurons, c-Fos density was analyzed in several brain regions linked to alcohol addiction, including the dorsal striatum (DS), prelimbic cortex (PL), and paraventricular nucleus of the thalamus (PVT). Significant differences in neuronal activation were found in the DS and PVT, but not the PL.

Specifically, the DEP-WDL group showed significantly increased activity in the PVT compared to all other groups. Additionally, increased neuronal activity was observed in the DS of DEP-WDL rats compared to non-dependent rats (NDEP-WDL and NDEP-NWDL). However, there was no significant difference in DS activity between the DEP-WDL and DEP-NWDL groups. Because WDL experience was a key factor in the observed neuronal activity, particularly in the PVT, further analyses focused on comparing the DEP-WDL and NDEP-WDL groups across a wider range of brain regions.

This extended analysis included the infralimbic cortex (IL), nucleus accumbens (NAc), central nucleus of the amygdala (CeA), and basolateral amygdala (BLA), along with the previously examined regions. A significant increase in active neuron density was found in the PVT, DS, and CeA in the DEP-WDL group compared to the NDEP-WDL group. No significant differences were found in the PL, IL, NAc, BLA, or lateral hypothalamus (LH) between these two groups.

Discussion

Differential Neuronal Recruitment in Alcohol-Seeking Rats With Versus Without a History of WDL

Previous behavioral studies have shown that environmental cues associated with reversing alcohol withdrawal have a stronger impact on alcohol seeking than cues linked to the positive effects of alcohol alone. This study extends those findings to the brain level, suggesting that specific clusters of neurons in the PVT, CeA, and DS play a role in the powerful motivating effects of withdrawal-related learning (WDL). The findings confirm that rats with a WDL history, which exhibit compulsive alcohol seeking, show distinct patterns of neuronal activity in these regions compared to rats without such a history. This supports the idea that the brain processes stimuli linked to withdrawal relief differently from stimuli associated with other types of alcohol learning.

c-Fos-Positive Neurons Are Recruited During Alcohol Seeking in the DS of Dependent Animals

Regardless of whether rats had experienced WDL, dependent animals showed increased c-Fos density in the dorsal striatum (DS) compared to non-dependent rats when exposed to alcohol-related cues. This increased activation in the DS of dependent rats, including those without WDL experience, suggests that this brain region becomes progressively involved in the formation of habitual behavior as alcohol dependence develops. This supports the idea that dependence itself can lead to changes in DS activity, contributing to learned alcohol-seeking behaviors.

Addiction as a State of Reward Dysregulation and Hedonic Allostasis

The study observed that increased neuronal activity in the PVT occurred exclusively in rats that experienced a negative emotional state during alcohol withdrawal, which was then relieved by alcohol. This suggests that the PVT is specifically linked to learning how alcohol reverses the stressful aspects of withdrawal, a process known as negative reinforcement. This role for the PVT aligns with the concept of hedonic allostasis, a chronic shift from normal emotional balance that contributes to reward dysregulation and addiction. The prominent activation of the PVT after WDL implies a significant role for this region in the brain's response to negative emotional states and the progression towards a state where seeking relief from distress drives further substance use.

Limitations and Their Implications

These findings are based on studies in male rats, so further research is needed to determine if the same patterns occur in female rats. The study was also limited by the amount of brain tissue available from the original behavioral experiments. While c-Fos protein expression is a useful marker for recent neuronal activity, it indicates that neurons have been active, but not necessarily that they were directly involved in the initial learning task. Other mechanisms, such as the recruitment of additional neurons by existing connections, could also contribute to the observed activity. Furthermore, background neuronal activity influenced by a rat's unique experiences could also affect c-Fos expression, meaning the observed signal is superimposed on this existing activity.

Conclusions

The identified activated neurons appear to form a neural foundation for compulsive drug seeking. This behavior is likely driven by: 1) negative reinforcement learning related to hedonic allostasis (involving the PVT), 2) the development of habitual behaviors during dependence progression (involving the DS), and 3) stress memories linked to the context of withdrawal and its relief (involving the CeA). To confirm these roles, future research could involve selectively silencing these active neuron groups. These findings have broader implications for understanding various maladaptive behaviors associated with reward dysregulation and the processing of emotionally significant stimuli, extending beyond substance use disorders to areas like fear conditioning, anxiety, and traumatic avoidance learning.

Summary

The brain learns by linking situations or places with experiences that are either pleasant or unpleasant. This basic learning process is stored in brain circuits and guides behavior. Both animals and people understand stimuli as rewarding or aversive. They can also learn that avoiding or getting rid of negative things helps reduce unpleasant feelings. For example, in individuals with drug dependence, certain places or situations become linked to the drug because using it relieves the bad feelings of withdrawal. The brain's ability to process these meaningful cues related to negative reinforcement learning is crucial for maintaining stability, well-being, and survival. Therefore, understanding how the brain represents these cues that influence behavior is a key area of study.

When exposed to important cues, small groups of brain cells, called neuronal ensembles, become active. These groups can activate even if the cells are not directly connected. The learned connections between cues and rewarding or unpleasant experiences are a main reason why addictive alcohol use often leads to relapse. To study this, a specific method called withdrawal-related learning (WDL) was used. This procedure helps identify brain cells that react to cues and drive behavior through negative reinforcement. In the WDL process, environmental cues become linked to getting rid of unpleasant alcohol withdrawal symptoms by drinking alcohol. These cues gain a learned value, which can greatly contribute to craving and relapse, causing problems with how pleasure is regulated and leading to an imbalance in well-being.

Small, spread-out groups of brain cells form to store learned connections that drive desired behaviors, including seeking drugs. This is considered the fundamental way learning happens. However, the specific brain pathways that cause the motivating effects of cues linked to reversing negative emotional states, such as sadness, anxiety, and stress sensitivity after heavy or long-term substance use, are not yet fully understood. Previous research showed that rats with a history of alcohol dependence and withdrawal-related learning (WDL) sought alcohol differently when exposed to cues that relieved withdrawal. Their alcohol seeking was stronger and more compulsive, meaning it continued despite punishment or requiring more effort, unlike rats without dependence. This study, using a rat model of relapse, aimed to identify brain cells that become active due to environmental cues linked to alcohol. The study compared cues tied to alcohol availability during withdrawal (negative reinforcement) with cues tied to alcohol availability when not dependent (positive reinforcement). Researchers also wanted to see if alcohol seeking in rats with WDL experience activated different brain cells than in rats without this experience, and to understand these differences.

Methods and Materials

Animal Use and Care

All procedures followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and received approval from the Institutional Animal Care and Use Committee of Scripps Research. Adult male Wistar rats, weighing approximately 450 grams, were used in the study. Brains were collected from rats that underwent training and testing as described in the behavioral training section.

Behavioral Training

The withdrawal-related learning (WDL) and all other behavioral procedures were carried out as described in earlier studies. These methods created complex cues (including both environmental signals and immediate responses) linked to alcohol availability. This was done either during withdrawal in dependent rats or when rats were not dependent or had recovered from dependence. The main WDL experimental group (DEP-WDL, 13 rats) aimed to see how these cues affected alcohol seeking even when rats faced challenges like foot shocks or increased effort.

After learning to self-administer ethanol (alcohol), some rats breathed alcohol vapor to become dependent, while others remained nondependent. After three weeks, rats were temporarily taken out of the vapor chambers. Eight hours into withdrawal, they could self-administer alcohol for 30-minute sessions in the presence of specific cues. This training continued for nine sessions, with 1 to 2 days between sessions during which rats remained undisturbed.

Three control groups were also included. The first was a nondependent group (NDEP-WDL, 12 rats). This group helped compare the effects of having a history of dependence and withdrawal (DEP-WDL) versus only being nondependent (NDEP-WDL) on alcohol seeking caused by cues, with all other conditions being the same. A "no WDL" (NWDL) group provided a control for dependence history alone, without alcohol reinforcement during withdrawal.

The second control group was dependent but without WDL experience (DEP-NWDL, 9 rats). This group checked if dependence history alone, rather than the WDL experience, explained the observed effects on alcohol seeking. The third group was nondependent and had no WDL (NDEP-NWDL, 9 rats). This group was tested alongside the DEP-NWDL group to compare alcohol seeking between a history of dependence without WDL and a history of nondependence.

After dependence was induced, DEP-NWDL rats were withdrawn from alcohol for one week. They then had the chance to self-administer alcohol in the presence of complex cues. Like the WDL group, this training involved nine sessions. After this, rats stayed in their home cages for one week, followed by daily re-exposure to the alcohol self-administration chamber without alcohol (extinction). Tests for tolerance to increased effort, resistance to punishment, and simple relapse were conducted over 14 days. These procedures helped illustrate the study design.

Tissue Preparation

Brains from a previous study were used to examine brain cell activity in this research. All rats were humanely euthanized and then had their brains flushed with a special preserving solution 90 minutes after being re-exposed to the alcohol-related cues. This specific timing allowed researchers to see the expression of c-Fos protein, a marker for recently active brain cells. Brains were then removed and prepared for slicing into thin sections, 40 to 50 micrometers thick, using a specialized cutting device.

All brain sections were collected for a process called immunostaining. This involved labeling the tissue with an anti-c-Fos antibody, which specifically detects total c-Fos protein in the cells without reacting with other similar proteins. A second antibody, Alexa Fluor 488, was then used to make the c-Fos-positive (active) neurons visible. The tissue sections went through a series of steps: incubation in a blocking solution, then in the primary c-Fos antibody for 72 hours, followed by rinsing, and then incubation in the secondary antibody for 4 hours. Finally, after more rinses, DAPI staining was applied, and the tissue was mounted on slides.

Quantitative Analysis

Details about the imaging can be found in the Supplement. Brain regions chosen for analysis were carefully outlined. Images showed activated neurons that expressed c-Fos protein in areas like the cortex, amygdala, and paraventricular thalamus. These c-Fos-expressing neurons within the specified brain regions were automatically detected using a cell detection function in NeuroInfo-rat software.

The NeuroInfo-rat software was specifically designed for detecting c-Fos-expressing cells in rat brain tissue, based on the data from this study. After automated detection, all cell counts were reviewed and corrected by a researcher. Three different researchers edited the automated detections, and their final cell counts were averaged. The researchers did not know which treatment group each sample belonged to. To represent brain cell activity, the number of c-Fos-positive cells was divided by the total surface area of each region of interest, providing a c-Fos density measurement.

Statistical Analysis

The study included four experimental groups: DEP-WDL (13 rats), NDEP-WDL (12 rats), DEP-NWDL (9 rats), and NDEP-NWDL (9 rats), as previously described. To understand how exposure to cues affected brain cell activity, c-Fos density was the main measurement used. Although the data met some requirements for standard statistical tests, they did not show normal distribution or equal variances across groups.

Specific statistical tests confirmed these issues. The Shapiro-Wilk test showed that the c-Fos values in the WDL group were not normally distributed. Levene’s test also found significant differences in variances among the groups. Because of these findings, statistical analyses for the effects on c-Fos density used nonparametric tests. When necessary, these were followed by specific comparisons to pinpoint differences, adjusted using Bonferroni corrections.

Results

Overall Neuronal Activation

To see if exposure to withdrawal-related learning (WDL) cues affected brain cell activity, the density of c-Fos was measured in important brain areas known to be involved in alcohol seeking and craving. A statistical test, the Kruskal-Wallis test, showed a notable difference in c-Fos density among the four treatment groups (DEP-WDL, NDEP-WDL, DEP-NWDL, and NDEP-NWDL).

Further analysis revealed that the DEP-WDL group had a higher overall density of activated neurons compared to both nondependent groups (NDEP-WDL and NDEP-NWDL). This indicates increased brain activity in rats that experienced both dependence and withdrawal-related learning when exposed to alcohol cues. However, no significant differences were found between the DEP-WDL and DEP-NWDL groups, or between the DEP-NWDL and NDEP-NWDL groups.

Brain Activity Linked to Withdrawal Learning

To pinpoint the locations of these cue-reactive brain cells, c-Fos density was first analyzed in three brain regions known to be important in alcohol addiction: the dorsal striatum (DS), prelimbic cortex (PL), and paraventricular nucleus of the thalamus (PVT). A Kruskal-Wallis test showed significant differences in brain cell activity (c-Fos density) in the DS and PVT across the groups, but not in the PL.

Further comparisons indicated that the DEP-WDL group had a much higher density of activated neurons specifically in the PVT compared to all other study groups. This suggests a strong role for the PVT in rats experiencing both dependence and withdrawal-related learning. Additionally, more brain cell activation was seen in the DS of DEP-WDL rats compared to nondependent groups (NDEP-WDL and NDEP-NWDL). However, there was no difference in DS activity between the DEP-WDL and DEP-NWDL groups. Since the WDL experience was key to these increases in brain activity, especially in the PVT, later analyses focused on comparing neuronal changes between the DEP-WDL and NDEP-WDL groups.

For a broader analysis across various brain regions, c-Fos counts were also taken in four other areas known for their role in drug addiction: the infralimbic cortex (IL), nucleus accumbens (NAc), central nucleus of the amygdala (CeA), and basolateral amygdala (BLA). To better understand the brain changes and activation of c-Fos-positive neurons resulting from learning experiences during withdrawal in dependent animals versus nondependent controls, a Mann-Whitney U test compared brain cell activity in the DEP-WDL and NDEP-WDL groups across eight brain regions.

The findings showed a notable increase in the density of activated neurons in the PVT, DS, and CeA for the DEP-WDL group after their withdrawal-related learning experience. However, there was no difference in c-Fos density between the DEP-WDL and NDEP-WDL groups in the PL, IL, NAc, BLA, or lateral hypothalamus (LH).

Discussion

Different Brain Activity in Alcohol Seeking

Previous behavioral research showed that environmental cues specifically linked to relieving withdrawal symptoms (the negative reinforcing effects of alcohol) have a stronger influence on alcohol seeking than cues linked only to the pleasant effects of alcohol. Specifically, alcohol seeking in rats with a history of withdrawal-related learning (WDL) was not only more intense but also resistant to challenges, such as punishment by footshock and increased effort, compared to seeking behavior caused by cues in a nondependent state. These current findings expand on those observations by showing specific brain areas, namely clusters of neurons in the paraventricular nucleus of the thalamus (PVT), central amygdala (CeA), and dorsal striatum (DS), are involved in the strong motivating effects of WDL.

The results confirm notable differences in brain activity across regions activated during compulsive alcohol seeking triggered by cues in rats with a WDL history, compared to rats without that history who did not show compulsive behavior. Specifically, active c-Fos-positive neurons were found in the PVT, DS, and CeA of WDL animals. In these animals, environmental cues were linked to reversing unpleasant withdrawal effects (negative emotional states).

In contrast, exposing rats to cues associated with the pleasant aspects of alcohol consumption (in nondependent or postdependent rats without WDL experience) did not activate the same number of c-Fos-positive neurons in the PVT. This exposure resulted in only mild, though still significant, activation in the DS of dependent rats (DEP-NWDL) compared to nondependent animals (NDEP-NWDL). These findings support the idea that the learned effects of cues linked to relieving withdrawal distress are processed differently in the brain than the effects of cues linked to other types of learning, in both alcohol-dependent and nondependent rats.

In the PVT, increased c-Fos density was observed only in the WDL group, not in the three control groups. This distinct pattern of brain cell activation solely in the PVT of the WDL group suggests this area plays a key role in two ways: first, in learning the negative link between alcohol use and the unpleasant effects of withdrawal, and second, in the development of compulsive drug seeking that follows.

The PVT is a critical center for brain circuits involved in drug addiction and has a known role in emotional responses to anxiety and stress. Stress increases PVT activity, and this area has been shown to be essential for stress-induced relapse to oxycodone seeking. The unpleasant feelings of alcohol withdrawal are strongly linked to stress. Thus, this finding supports the idea that withdrawal stress plays a significant part in WDL learning and compulsive drug seeking. The PVT, which is sensitive to stress, sends signals to the CeA, a brain region connected with negative emotions, stress, alcohol dependence, and particularly the unpleasant effects of alcohol withdrawal. As expected, exposure to WDL-related cues led to more active c-Fos-positive neurons in the CeA. This suggests higher CeA activity compared to NDEP-WDL rats, where the cues were only linked to the pleasant experience of drinking alcohol. Given that blocking the connection from the PVT to the CeA reduces stress responses, these findings support a significant role for the PVT-CeA system in how the body responds to stress.

Brain Activity in Dependent Animals

Regardless of the WDL experience, exposing dependent rats to cues linked to the unpleasant aspects of alcohol withdrawal led to increased c-Fos density in the striatum compared to nondependent rats. The development of dependence is linked to the formation of habitual behaviors, which are believed to be controlled by striatal brain regions, especially the dorsal striatum (DS). This increased activity of neurons in the DS of the DEP-WDL group might be due to the gradual involvement of dorsal striatal areas in creating habits as substance dependence develops. This aligns with the "spiraling hypothesis," which suggests how one part of the striatum begins to control other parts during dependence.

A pattern of activated neurons in the DS was also observed in postdependent animals (DEP-NWDL). These rats had a history of dependence but no WDL experience; they were trained to link cues with alcohol after withdrawal. The brain cell activation in the DS of these DEP-NWDL animals was less than in the WDL history group but still significantly different from nondependent controls (NDEP-NWDL). This suggests that DS neurons become active not only because of the WDL experience or negative reinforcement learning but also due to other factors related to dependence or past experiences, independent of WDL.

Addiction, Reward, and Well-being

It is believed that rats in all nondependent groups (NDEP-WDL) experienced a pleasant state while self-administering alcohol. In contrast, rats in the WDL group (DEP-WDL) experienced a deeply unpleasant state during withdrawal. When re-exposed to cues, increased brain cell activity occurred only in the PVT of DEP-WDL animals. These cue-reactive neurons in the PVT were specifically linked to the WDL experience and likely connected to reversing the stressful feelings of alcohol withdrawal in rats with that experience.

In contrast, active brain cell groups in the DS, while showing differences, were observed in both dependent animals with WDL experience (DEP-WDL) and those with a history of dependence but no WDL (DEP-NWDL). Considering previous observations, these new findings suggest that the PVT might play a broader role in behaviors driven by unpleasant states beyond just alcohol seeking. It could be a key area for the development of "hedonic allostasis," which is a long-term shift away from normal feelings of pleasure and well-being, linked to issues with reward regulation, stress, unusual motivation, and addiction.

The training methods used in this study, which caused behaviors driven by both pleasant and unpleasant states, could be useful for exploring the "opponent process hypothesis" of motivation and the development of hedonic allostasis. This hypothesis suggests that experiencing a pleasant state temporarily leads to an opposing unpleasant state until balance is restored. With repeated exposure to pleasant stimuli or substances like alcohol or opioids, this balance shifts. Both the initial pleasant feeling and the subsequent unpleasant opposing process still happen, but eventually, the overall state falls below normal well-being, creating a chronic state of imbalance called hedonic allostasis.

While these ideas about motivation are widely accepted, the specific brain mechanisms that control them are still unclear. Based on these findings, it is suggested that the PVT plays a significant role in these opposing processes and the development of hedonic allostasis. The strong activation of the PVT after exposure to WDL cues points to this brain area as crucial for the powerful motivating effects of cues. These cues are linked to alcohol (or other rewards) that reverse negative emotional states—a process known as negative reinforcement. This, in turn, drives more alcohol consumption, worsening the negative emotional process and eventually leading to a chronic state of imbalance.

Since significant brain cell activation in the PVT was linked solely to alcohol seeking caused by WDL-paired cues, this brain region could be an important target for studying the neural mechanisms behind opponent processes and the development of hedonic allostasis.

Limitations of the Study

The findings of this study are based only on male rats, and further research is needed to see if the same results apply to female rats. The study also had some limitations due to the amount of brain tissue available from the initial behavioral experiments.

Additionally, brain cell activation, used to identify cue-reactive neurons, was measured by c-Fos protein expression, a sign of recent activity. However, groups of c-Fos-positive neurons can become active in response to cues through a process where nerve endings connect with many other neurons, even those not directly involved in the initial learning. This means that nerve cells responding to cues might activate other neurons that were not active during the learning process to drive behavior. Such a process could increase overall brain activity in the PVT of DEP-WDL rats by bringing in more neurons, going beyond the typical groups of active cells. This activation could also have been affected by the animal’s past experiences, leading to some c-Fos expression simply from general activity. Therefore, the activity seen from the cues likely occurs alongside a baseline level of c-Fos activity influenced by each animal's unique experiences throughout the study.

Conclusions

The findings indicate that the active neurons identified in this study form a brain basis for compulsive drug seeking. This behavior results from three main factors: first, negative reinforcement learning linked to a chronic imbalance in well-being (in the PVT); second, the development of habitual behavior as dependence progresses (in the DS); and third, stress memories tied to the situations where withdrawal and its relief were experienced (in the CeA).

To fully confirm the role of these cue-activated neurons in mediating the WDL experience, further experiments are needed. These might involve selectively turning off specific groups of these reactive neurons. Ultimately, these findings have clear implications for understanding problematic behaviors related to issues with reward regulation and the processing of emotionally significant cues that drive actions beyond just substance use disorders. Future research should also explore the brain changes that connect the learned value of substance-related environments to increased drug seeking, and extend this to other types of problematic behaviors, such as fear conditioning, anxiety disorders, traumatic avoidance learning, and even predatory behavior.

Summary

The brain learns to connect places or things with good or bad feelings. This learning helps animals and people know what to do. For example, people with drug addiction learn that using a drug can stop the bad feelings of withdrawal. This ability to learn from bad feelings and change behavior is very important for staying healthy and safe. Scientists want to understand how the brain stores these learned connections.

When something important happens, certain groups of brain cells become active. This study looked at how different brain cells turn on when rats learn about alcohol. The researchers used a method called Withdrawal-Related Learning (WDL) to see which brain cells were active when rats learned that alcohol could stop their bad withdrawal feelings. This kind of learning is a big reason why people keep craving and using alcohol. When a person is in a state of addiction, the brain's reward system can become unbalanced, leading to a constant state of feeling bad.

Understanding how active brain cells store these learned connections is key. Especially, scientists want to know which brain parts help with learning to use a substance to stop bad feelings like sadness, worry, and stress after heavy use. Earlier studies showed that rats with WDL experience looked for alcohol much more strongly and compulsively than rats without this experience. Their urge to drink was so strong it was hard to stop, even with punishment. So, this study used rats to find out: 1) which brain cells become active when rats seek alcohol during withdrawal (when alcohol helps stop bad feelings) versus when they seek it just for a good feeling (without withdrawal), and 2) if different brain cells are involved in alcohol seeking for rats with WDL experience compared to those without it.

Methods and Materials

All study steps followed rules for animal care and were approved by a special committee. Adult male rats were used.

The rats went through different training steps. A key part was the WDL method, which taught rats that certain sights and sounds meant alcohol was available, and drinking it would make their withdrawal symptoms better. Some rats were made dependent on alcohol, meaning they would experience withdrawal. Other rats were not dependent. The study compared four groups of rats: 1) dependent rats with WDL experience (DEP-WDL), 2) non-dependent rats with WDL experience (NDEP-WDL), 3) dependent rats without WDL experience (DEP-NWDL), and 4) non-dependent rats without WDL experience (NDEP-NWDL). After drinking alcohol on their own, some rats breathed alcohol vapor for weeks to become dependent. Then, during short breaks from the vapor, they could drink alcohol when specific signals were present. This taught the dependent rats that alcohol could ease their withdrawal. Other groups had different training to compare the effects of dependence and WDL.

After the training, the rats were briefly put back into the places where they had learned about alcohol, but without alcohol. About 90 minutes later, their brains were collected. This timing allowed researchers to see a special protein called c-Fos, which shows which brain cells were active. Brains were cut into thin slices and treated with special colors to highlight the c-Fos protein.

Scientists then looked at specific parts of the brain under a microscope. They used special computer software to count the number of active cells (cells with c-Fos). This count showed how much brain activity there was in different areas. To make sure the counts were right, human experts checked and edited the computer's work. The experts did not know which treatment group each brain belonged to. The number of active cells in each area was then used to compare the groups. Special math tests were used to compare the groups because the data did not fit common statistical patterns. These tests helped to find important differences between the groups.

Results

Researchers looked at brain activity in key areas known to be involved in alcohol seeking. Special math tests showed a clear difference in overall brain activity among the four groups of rats. Specifically, dependent rats that learned to use alcohol to relieve withdrawal (DEP-WDL group) had more active brain cells than both groups of non-dependent rats. This means that learning to use alcohol to stop bad feelings led to more brain activity compared to just seeking alcohol when not dependent. There was no big difference in overall brain activity between the DEP-WDL group and the dependent rats without WDL, or between the two non-dependent groups.

To pinpoint where in the brain these changes happened, researchers looked at active cells in three specific brain areas: the dorsal striatum (DS), prelimbic cortex (PL), and paraventricular nucleus of the thalamus (PVT). The results showed that brain activity in the DS and PVT was different among the groups, but not in the PL. The DEP-WDL group had many more active cells in the PVT compared to all other groups. This suggests the PVT is very important for learning during withdrawal. In the DS, active cells were also higher in dependent rats with WDL compared to non-dependent rats. However, dependent rats without WDL also showed some increased activity in the DS.

Because learning during withdrawal (WDL) seemed to be a key factor, especially for the PVT, the researchers focused more closely on comparing the DEP-WDL group with the NDEP-WDL group across eight brain regions. These comparisons showed significantly more active cells in the PVT, DS, and central amygdala (CeA) for the DEP-WDL group. No significant differences were found in the other brain areas (IL, PL, NAc, BLA, or LH) between these two groups.

Discussion

Earlier studies showed that rats learning to drink alcohol to stop withdrawal feelings (WDL) sought alcohol more strongly and compulsively. These rats were harder to stop, even with punishment or extra effort. This new study shows that these differences are linked to specific brain areas. It points to groups of brain cells in the PVT, CeA, and DS as being important for this powerful, withdrawal-driven alcohol seeking.

The findings confirm that different brain areas become active when rats compulsively seek alcohol due to WDL, compared to rats that do not show this strong, uncontrollable behavior. Specifically, active brain cells appeared in the PVT, DS, and CeA of WDL animals. These were the animals where certain places and things were linked to making bad withdrawal feelings better. In contrast, when stimuli were linked to only the good feelings of drinking alcohol (in non-dependent rats or dependent rats without WDL experience), fewer active cells were seen in the PVT, and only a small increase was noted in the DS of dependent rats without WDL. This suggests that the brain reacts differently to cues that promise to stop bad feelings (withdrawal) versus cues that promise good feelings (pleasure from alcohol).

In the PVT, more active cells were found only in the WDL group, not in any of the other control groups. This pattern suggests the PVT is crucial for 1) learning that alcohol can reverse bad withdrawal feelings and 2) the strong, uncontrollable urge to seek drugs that follows. The PVT is a key brain area involved in drug addiction, stress, and anxiety. Alcohol withdrawal causes a lot of stress. This finding matches the idea that stress from withdrawal plays a big role in WDL and strong drug seeking. The PVT is connected to the CeA, another brain area linked to bad feelings, stress, and alcohol withdrawal. As expected, rats exposed to WDL cues also had more active cells in the CeA. This suggests that the PVT and CeA work together in the brain's response to stress and strong alcohol seeking.

Even without the WDL experience, dependent rats that were exposed to cues linked to the bad feelings of alcohol withdrawal showed more active cells in the DS compared to non-dependent rats. When a person becomes dependent on a substance, their behavior can become a habit, which is thought to involve brain areas like the DS. The increased activity in the DS of the DEP-WDL group might be because this brain area gets more involved in forming habits as addiction develops. Active cells were also seen in the DS of rats that had been dependent but had no WDL experience. These rats had learned to link cues with alcohol after withdrawal was over. The activity in the DS of these rats was less than in the WDL group, but still higher than in non-dependent rats. This means that DS cells become active not just because of WDL, but also due to other factors linked to being dependent or having certain experiences, even without WDL.

It is believed that non-dependent rats felt good while drinking alcohol. However, rats in the WDL group felt very bad during withdrawal. When these rats saw the cues again, only the PVT in the WDL group showed more active cells. The active cells in the PVT were specifically linked to WDL and likely to how alcohol stopped the stress of withdrawal in these rats. In contrast, the active brain cells in the DS were seen in both dependent rats with WDL and dependent rats without WDL. These new findings suggest that the PVT might play a larger role in behaviors driven by bad feelings beyond just alcohol seeking. It could be a key area for how the brain gets stuck in a state of feeling bad, where the reward system is out of balance, and there's constant stress and abnormal motivation, which are all part of addiction.

The way this study linked cues to both good and bad feelings could help researchers understand how feelings balance out and how long-term bad feelings can develop. This idea suggests that after a good feeling, a short bad feeling often follows, and the body tries to return to normal. But over time, if someone keeps having good feelings from things like alcohol or other rewarding activities, both the good and the following bad feelings can get worse, leading to a long-term state of feeling bad. While these ideas are widely accepted, how they work in the brain is not fully understood. Based on this study, the PVT might be very important in these processes. The strong activity in the PVT after WDL suggests this brain area helps connect cues to alcohol's ability to stop bad feelings. This drives more alcohol use, which makes the bad feelings worse over time, leading to that long-term state of feeling bad. Since strong PVT activity was only linked to alcohol seeking caused by WDL cues, this brain area could be a main focus for studying how the brain handles these opposing feelings and develops long-term negative states.

These results only come from male rats, so more studies are needed to see if the same is true for female rats. The study also had a limited amount of brain tissue. Also, active brain cells were identified by looking at c-Fos protein, which shows recent activity. However, some brain cells can become active in ways that are not always tied to how they learned something. This means that other cells not directly involved in the learning task might also get activated and influence behavior. This activity could also be affected by a rat's past experiences, leading to some background c-Fos activity. So, the observed brain activity is a mix of the specific cues and the rat's overall experience.

These findings suggest that the active brain cells found in this study are key to the strong, uncontrollable urge for drugs. This urge comes from three main sources: 1) learning to stop bad feelings, which can lead to a long-term state of feeling bad (in the PVT), 2) the development of habitual behavior as addiction progresses (in the DS), and 3) stress memories linked to the places where withdrawal was experienced and eased (in the CeA). To prove these active cells are specifically involved in WDL, more studies are needed to turn off these cells and see what happens. Lastly, these findings can help us understand other unhealthy behaviors linked to unbalanced reward systems and how strong emotional cues drive behavior, not just in substance use. Future research should also look at how learned cues related to substances make drug seeking worse, and extend this to other behaviors like fear, anxiety, avoiding trauma, and possibly even predatory behavior.