Alcohol effects on associative and sensorimotor cortico-thalamo-basal ganglia circuits alter decision making and alcohol intake

Abstract

Much of the behavioral repertoire of humans and other vertebrates is learned and controlled through the function of brain circuits involving the cortex, thalamus and Basal Ganglia (for simplicity we will refer to this as the Cortico-Thalamo-Basal Ganglia, or CTBG, circuitry). As the name implies, these circuits include the different regions of cortex and thalamus, as well as BG subregions including the striatum, globus pallidus (GP), substantia nigra (SN)/ventral tegmental area (VTA), and the subthalamic nucleus (STN). This circuitry has developed evolutionarily to provide overarching control of actions following discrete environmental events as well as self-initiated actions. Several parallel CTBG circuits have been identified and linked to different aspects of action control under different circumstances. Research in experimental psychology and Neuroscience has established how different CTBG circuits contribute to control of actions based on environmental circumstances and past learning history. There is also a large and growing body of evidence that misused substances, including alcohol, act on cells within these circuits. These actions promote acute intoxication and drug seeking and contribute to changes in behavior induced by chronic alcohol exposure, withdrawal and relapse. Alcohol exposure also influences which of the different CTBG circuits has the strongest influence on behavior. This review will cover the relevant circuitry and describe the current state of knowledge as to how alcohol alters CTBG circuit function and control of behavior. Studies in rodents, non-human primates and humans will be discussed. Finally, ideas for future research directions in this area will be considered.

Introduction

Alcohol is one of the most widely consumed psychoactive substances in the world. Unfortunately alcohol misuse and alcohol use disorder (AUD) are widespread with devastating consequences for individuals, families, and society at large (https://www.niaaa.nih.gov/alcohols-effects-health/alcohol-topics/alcohol-facts-and-statistics). Alcohol impairs cognitive functions including memory and decision making, which likely contributes to decisions regarding alcohol intake and excessive drinking. Thus, it is important to understand the neural mechanisms that contribute to alcohol-induced cognitive function, behavioral change and ultimately alcohol intake. Work in both human volunteers and laboratory animals has revealed important roles for cortico-thalamo-basal ganglia (CTBG) circuits in these alcohol effects and control of drinking. The CTBG system contains several subcircuits that act in parallel but inter-communicate to combine sensory information and intrinsic neural activity to ultimately generate and pattern actions. As the next section of this review will describe, the different CTBG subcircuits control actions in response to different information resulting in differential relationships between environmental events and behavior. Research over the last few decades has provided increasingly detailed information about the role of the different CTBG circuits in the neurobehavioral actions of alcohol as well as control of alcohol drinking. The role of the “limbic” CTBG circuit in rewarding effects of alcohol and affective influences on alcohol-related behaviors is well known and has been the subject of many reviews (e.g. Koob, 2014). Thus, the present review will focus mainly on alcohol and the “associative” and “sensorimotor” CTBG circuits.

The review begins with a detailed description of the different brain regions as well as cells in the CTBG subcircuits and their functions in behavioral control. A dominant theme is the important role of the associative circuit in relatively rapid learning of new actions and the control of these actions based on associations between actions and outcomes (especially reward). Likewise, the role of the sensorimotor circuit in slower action learning based on associations between environmental stimuli (or context) and action production is also a key concept in this discussion. The effects on neuronal, synaptic and CTBG circuit function of acute and chronic alcohol exposure (including both forced administration and voluntary alcohol consumption) will then be reviewed with a focus on distinguishing effects in the different subcircuits. Studies from both laboratory animals and human experiments are discussed in this section. The main emphasis in the discussion of acute and chronic alcohol effects will be on actions in the cortex and striatum as the literature on these topics is relatively large. Less is known about alcohol effects in other CTBG areas, but the growing literature on effects in other BG areas will be reviewed. Unfortunately, little is known about alcohol effects in thalamic subregions of the CTBG circuitry, and thus this topic will not be discussed in much detail. This discussion will also focus mainly on neurons, as there is to date little information about alcohol effects on glia in the CTBG circuitry. The current state of knowledge about the roles of the different associative and sensorimotor subcircuits in alcohol effects on decision making and action control will then be reviewed, followed by discussion of how these two subcircuits contribute to alcohol intake. This discussion will include information from both laboratory animal and human studies. Finally, ideas for future directions in this research area will be discussed. In this context, caveats about previous findings generated using earlier generations of research tools will be noted in the review. This discussion is meant to highlight questions that can now be explored with newly-developed techniques.

Parallel cortico-thalamo-basal ganglia circuits

Anatomical division of the different CTBG subcircuits

The basal ganglia (BG) are recognizable in all vertebrate species, and some invertebrates, and demonstrate increasing coevolution with cortex in mammals (Reiner, 2017; Strausfeld & Hirth, 2013). In mammals this system is highly integrated with the many cortical regions, including not only the layered cerebral cortical regions, but also more diffusely organized pallial regions like the amygdala. In the mammalian CTBG circuits, output from the SN reticulata (SNr) goes not only to motor control regions but also provides feedback to thalamocortical projection neurons to ultimately influence cortical output, a major influence on cortical activity that ultimately controls action production (McElvain et al., 2021). The mammalian CTBG circuit also shows increased subregional diversity with specificity of afferent-efferent connections between discrete cortical, thalamic, and BG regions. In rodents, non-human primates and humans several distinct CTBG subcircuits appear to act in parallel while some points of crosstalk can be identified based on this connectivity.

The three largest CTBG subcircuits are the associative, limbic and sensorimotor circuits (Fig. 1). The associative circuit (sometimes call the prefrontal circuit) includes diverse cortical regions including, but not limited to, orbitofrontal (OFC), prelimbic (PlC), and posterior parietal cortex (PPC) which project to the dorsomedial striatum (DMS) in rodents (roughly equivalent to the Caudate nucleus in primates) (Alexander et al., 1991). Glutamatergic projections from midline thalamic nuclei such as the intralaminar nucleus also innervate this striatal subregion along with dopaminergic afferents from the medial portions of the SN pars compacta (SNc) (Haber et al., 2000; Hunnicutt et al., 2016). This circuit also includes the GP external segment (GPe) and the major output from the circuit is from the medial regions of the SN reticulata (SNr). The limbic circuit includes not only glutamatergic projections to the ventral striatum (also known as Nucleus accumbens, NAc) from infralimbic and other prefrontal areas, but also from the hippocampus and basolateral amygdala pallial regions (Zahm, 2006). Glutamatergic thalamic inputs to NAc come predominantly from the central medial, paratenial and paraventricular nuclei (Su & Bentivoglio, 1990; Vertes et al., 2015), while the dopaminergic input comes from the VTA. The ventral pallidum and VTA constitute the core and output regions of this circuit, but the NAc also projects to hypothalamic subregions, SNc and SNr (Humphries & Prescott, 2010). Glutamatergic projections from sensory and motor cortices to dorsolateral striatum (DLS, the Putamen in primates), are characteristic of the sensorimotor (sometimes called the motor) circuit (Alexander et al., 1991; Haber, 2017). Thalamic glutamatergic innervation of DLS comes from an array of midline nuclei (Hunnicutt et al., 2016), while dopaminergic innervation is provided by the lateral portion of the SNc (Haber et al., 2000). The STN, lateral GP external segment (GPe) and lateral regions of SNr are also key parts of the sensorimotor circuitry (Alexander et al., 1991).

Fig. 1.

Connectivity in the three major CTBG Circuits. A) In the associative circuit, cortical regions such as lOFC, mPFC and posterior parietal cortex send glutamatergic projections to DMS. The intralaminar and midline thalamus also send glutamatergic projections to DMS. The DMS sends GABAergic projections to medial regions of GPe and SNr. The medial STN sends glutamatergic projections to medial SNr, and has reciprocal connections with GPe, sending glutamatergic afferents to GPe and receiving GABAergic connections in return. The medial SNc sends dopaminergic projections to DMS. B) The sensorimotor circuit has a similar organization. The glutamatergic cortical afferents to DLS arise from sensory and motor cortices (including M1 and SSC). The thalamic glutamatergic afferents to DLS come from IL and midline thalamus. The projections to and from GPe, SNr and STN are connected as in the associative circuit, with the exception that lateral aspects of each subregion are connected. The lateral SNc sends dopaminergic projections to DLS. C) The limbic circuit can also be thought of as a “ventral circuit”. Glutamatergic afferents from cortical areas such as Insula, ILC and mOFC as well as BLA and ventral hippocampus project to the VS/NAc. The limbic glutamatergic thalamostriatal projections come from midline thalamus and target VS/NAc. The VS/NAc sends GABAergic projections to VP, VTA and LHP. Dopaminergic innervation of VS/NAc comes from VTA. In each diagram glutamatergic projections are shown in yellow, GABAergic projections in white, and dopaminergic projections in green. It is important to note that for purposes of clarity not all afferent connections could be shown in these diagrams (e.g. some corticostriatal connections are not shown). Abbreviations: BLA = basolateral amygdala, DS = dorsal striatum, GPe = globus pallidus external segment, IL = intralaminar thalamus, ILC = infralimbic cortex, LHP = lateral hypothalamus, lOFC = lateral orbitofrontal cortex, mOFC = medial orbitofrontal cortex, M1 = primary motor cortex, PLC = prelimbic cortex, PPC = posterior parietal cortex, SNr = substantia nigra pars compacta/reticulata, SSC = somatosensory cortex, STN = subthalamic nucleus, VTA = ventral tegmental area. This figure uses images generated with the Allen Brain Explorer beta version.

Additional CTBG circuits, including the oculomotor circuit (Alexander et al., 1991), have been defined, and the three major circuits can also be subdivided based on factors such as the different NAc functional domains (Floresco, 2015; Pennartz et al., 1994; Zahm, 1999) or differential corticostriatal inputs (Hintiryan et al., 2016; Hunnicutt et al., 2016). The CTBG circuits also include points of cross–circuit interaction. For example, striatonigral projections from NAc target dopaminergic neurons in the SNc to influence modulation of dorsal striatum in an “upward spiral” configuration that can integrate limbic function with the other circuits (e.g. Haber, 2014; Humphries & Prescott, 2010). In addition, many of the cortical regions in the associative circuit do not have direct projections to spinal cord or other motor output regions (e.g. medial prefrontal cortex, orbitofrontal cortex) and thus the motor influence of this circuit is produced in part via cortico-cortical connections with M1 and premotor cortical areas (Anastasiades & Carter, 2021; Dum & Strick, 1991).

BG neuronal subtypes

The striatum contains several neuronal subtypes, most of which are GABAergic. The large majority (>90%) of the neurons in striatum are medium spiny projection neurons (MSNs) that release GABA as well as neuropeptides (Plenz and Wickens, 2017). An important feature of the CTBG circuits is the segregation of MSNs, and thus striatal output, into the direct and indirect pathways (Plenz and Wickens, 2017). The direct pathway MSNs (dMSNs) project to the SNr and inhibit tonically active GABAergic nigral neurons. The net effect of this inhibition is to remove tonic inhibitory transmission in downstream motor nuclei and relieve inhibition of thalamocortical projections with the net effect of increasing cortical output. In general, activation of the direct pathway is thought to enhance action production, although other complexities of the circuitry mean that the situation is not always so simple (Klaus et al., 2019; Yin, 2023). The indirect pathway MSNs (iMSNs) project to the GPe, where they inhibit the tonic firing of GABAergic neurons that project to the STN and SNr. This circuit has a net effect of decreasing activity of neurons in brainstem motor nuclei and cortex and is generally thought to decrease the likelihood of action production (although see Klaus et al., 2019 for information about the true complexity of circuit actions). It must also be noted that the MSNs synapse onto one another via local GABAergic connections (Plenz and Wickens, 2017).

The diversity of neurons in the striatum, and the BG in general, follows the theme in most brain regions (D.E. et al., 2017). The striatum contains several different interneuron subtypes that constitute about 5% of the total neurons in the region. The majority of these interneurons are GABAergic with several subtypes present in striatum that innervate both MSNs and other interneurons (Gerfen & Bolam, 2017). The fast-spiking interneurons (FSINs) that express the parvalbumin calcium binding protein have been identified as a target for acute and chronic alcohol actions, as will be discussed later in this review. Cholinergic interneurons (CINs) are also present and play a key role in striatal function. While the somata of these neurons are sparse, their axons have extensive ramifications that cover much of the striatum (Gerfen & Bolam, 2017). Thus, they are poised to influence cellular physiology through interactions with intrinsic neurons and afferent inputs.

There is also considerable neuronal diversity in other regions of the CTBG circuit. The GPe and SNr contain almost exclusively GABAergic neurons with some cholinergic neurons in GPe (Kita & Jaeger, 2017). Neuronal subclasses are now well defined in both regions with interesting differences in intrinsic activity, synaptic properties and afferent/efferent connectivity (Gerfen & Bolam, 2017; Kita & Jaeger, 2017; McElvain et al., 2021; Sitzia et al., 2024). Neurons in the GPe have several prominent afferent projections, with the heaviest to the STN and SNr. The SNr neurons innervate a variety of midbrain nuclei and motor nuclei in the brainstem, with interesting differences in projections of the associative (medial) and sensorimotor (lateral) subregions (McElvain et al., 2021). Notably, a subset of GPe GABAergic neurons project to striatum, providing a loop within the indirect pathway circuitry that may gate striatal output, as will be discussed later in this review (Abdi et al., 2015; Abrahao & Lovinger, 2018; Glajch et al., 2016; Kita & Jaeger, 2017; Mastro et al., 2014). As mentioned above, the neurons in both GPe and SNr exhibit tonic action potential firing. The STN contains mainly glutamatergic neurons that likewise are tonically active (Bevan, 2017). These neurons project to the GPe, generating constant, real-time feedback between these BG subregions. The STN neurons also provide strong excitatory innervation of the SNr that influences output of the BG. Cortical glutamatergic innervation of STN neurons forms the “hyperdirect” pathway that can influence SNr activity, and thus BG output, faster than either the direct or indirect pathways (Wilson, 2017).

Shaping of behavior by environmental events and experience

The CTBG circuitry is implicated in a variety of aspects of action control. These include the learning of new actions and control of the performance of established action repertoires (Yin, 2023). Before considering how the different CTBG subcircuits contribute to action learning and control, it is important to provide a brief summary of the different types of conditioning that shape mammalian actions. Learning new actions and learning when to perform or omit actions are key aspects of adaptive behavior (Thorndike, 1911). Pavlovian conditioning (PC, also referred to as stimulus-reward or stimulus-outcome, S–O, learning) is the process in which repeated pairing of an environmental event (stimulus) with a specific outcome (reward or punishment) leads the animals to produce the outcome-appropriate behavioral response when only the stimulus is presented. This process is classically known to apply to reflexes (Pavlov, 1927), but can also apply to innate actions such as approach behavior as well as learned actions (see Rescorla & Solomon, 1967; Yin, 2023 for discussion). Instrumental conditioning (IC, also known as operant conditioning) involves acquisition of new action repertoires in response to a particular outcome (usually rewarding but punishing in some cases) (Colwill & Rescorla, 1986; Dickinson, 1994; Skinner, 1938). Unlike PC, IC does not require pairing with any discrete environmental event that signals reward availability, although the environmental context is a key aspect of such learning. True instrumental behavior is defined by sensitivity to omission, that is the conditioned action will cease if the outcome is made contingent on the animal withholding its response (Davis & Bitterman, 1971). We will discuss different aspects of PC and IC as this section of the review continues, but first it is necessary to detail how CTBG circuits participate in these conditioning process.

The CTBG circuits have prominent functional roles in action learning, including PC and IC (reviewed in Yin & Knowlton, 2006; Yin, 2023). It is clear that the BG and CTBG circuits are not necessary for the simplest forms of PC, as this type of learning can be observed in animals that lack a clear cortex or BG (e.g. Aplysia californica, Walters et al., 1979). With respect to IC, early studies showed that extensive lesions of the dorsal striatum prevented such learning (Divac et al., 1967; Konorski, 1967). The accumulation of evidence over many decades indicates that the associative and sensorimotor CTBG circuits contribute to IC (Lovinger, 2010; Yin, 2023; Yin & Knowlton, 2006).

Functions of the sensorimotor circuit components

A deeper look at the three major CTBG circuits indicates how they are designed for development and performance of new actions. The cortical components of the sensorimotor circuitry include regions that receive relatively direct sensory input via the thalamus. Motor cortices that generate signals to initiate movement are also part of this circuit. The glutamatergic output of the primary motor cortex includes not only corticospinal projections that control movement initiation, but also collaterals to the DLS that, through the sensorimotor BG circuitry, provide modulatory input to brainstem motor control areas and feedback to the cortex. The cortical inputs directly activate striatal neurons. Thalamic glutamatergic inputs in this circuit are driven by environmental events to provide a fast sensory input that can activate neurons in DLS. The dopaminergic SNc inputs to DLS are activated by environmental events (including enhanced or decreased activity related to outcomes following actions), and dopamine modulates how striatal neurons respond to the cortical, thalamic and intrinsic striatal synaptic input to fine tune striatal output. Thus, one of the major roles of this circuit is to refine how cortical output controls actions based on environmental conditions (from sensory cortices and thalamus) and outcome history (from dopaminergic inputs). The dopaminergic modulation also appears to invigorate behavior, contributing to robust performance during the learning process and once actions are well-learned (Barnett et al., 2023; Yin, 2023).

Functions of the associative circuit components

The cortical aspects of the associative circuit integrate information from a variety of environmental sources with input from other cortical areas. Thus, these cortical regions process “cognitive” information such as outcome strength and valence, outcome expectancy, working memory and executive function (Goldman-Rakic, 1996; Masterman & Cummings, 1997), as opposed to more direct sensory/environmental information. The glutamatergic efferents from these cortical regions innervate striatal neurons along with thalamic inputs that contribute to processes including attention and arousal, action initiation and reinforcement, and behavioral flexibility (Cover et al., 2023; Cover & Mathur, 2021; Johnson et al., 2020; Kato et al., 2021; Nippert et al., 2024). As in the sensorimotor circuitry, the dopaminergic SNc inputs in the medial aspect of the associative circuitry provide modulation based on information about outcomes that follow behavior. However, the dopaminergic inputs to DMS appear to provide reward-related outcome information (Barnett et al., 2023; Yin, 2023). It has long been known that stimulation of SNc neurons will induce animals to produce actions that preceded such stimulation with relatively rapid onset (Wise, 2009), and subregion-specific optogenetic stimulation studies indicate that the medial SNc neurons that innervate DMS have this function (Ilango et al., 2014; Rossi et al., 2013). In contrast to the DMS, dopaminergic inputs to DLS have been implicated in signaling the salience of environmental events and enhancing movement vigor (Lerner et al., 2015; Markowitz et al., 2023; Yin, 2023). These signals will generally increase movement in response to prominent environmental events. As actions are repeated, those that are followed by dopamine increases in DLS will tend to be repeated despite no obvious evidence of reward (Markowitz et al., 2023). Thus, dopaminergic input to DLS also has a key role in action learning through a gradual reinforcement-based process (Yin, 2023). The roles of the associative and sensorimotor CTBG circuits in reward- and reinforcement-based learning will be discussed in more detail later in this review.

Circuit crosstalk

While there has been a tendency to separate different CTBG subcircuits and their functions, it is important to realize that the subcircuits communicate with one another at several levels. Feedback to the cortex in the associative circuit involves GPe, SNr and thalamocortical projections similar to those in the sensorimotor circuit (Haber, 2017). However, the cortical targets of the major thalamic projections in this circuit are the “associative” cortices that do not give rise to corticospinal projections. The outputs from SNr that are part of the associative circuit are also less likely to target brainstem motor centers, in comparison to the sensorimotor part of this nucleus (McElvain et al., 2021). Thus, the associative circuit can only influence motor output indirectly. One way this occurs is through glutamatergic afferents from the prefrontal cortex (PFC) to the premotor cortical areas (Barnett et al., 2023; Hélie et al., 2015) as well as the projections to striatum mentioned earlier. Through these connections the associative circuit output can help drive activity in the movement modulating sensorimotor circuit to influence actions.

Differential CTBG circuit roles in action learning

The two “dorsal” CTBG circuits influence action learning and production in subtly different ways. The associative circuit contributes to IC-based learning of new actions in a reward-driven Action-Outcome (A-O) manner (Barnett et al., 2023; Yin & Knowlton, 2006). In this context, the somewhat loaded term “reward” refers to the presence, magnitude and valence of the immediate outcome that follows the action (White, 1989; Yin, 2023; Yin et al., 2008). In practice, this type of learning leads to production of actions in a given context that are highly sensitive to the outcome and involves conscious tracking of ongoing outcomes and expected outcomes. This system is specialized for relatively rapid learning of new actions.

The sensorimotor circuit is specialized for learning and producing actions based on environmental context and reinforcement history (Barnett et al., 2023; Yin et al., 2023). This type of learning is most often called Stimulus-Response (S–R) learning, and we will use this term, but may be thought of more accurately as environment-response learning driven by reinforcement. In this context, the also somewhat loaded term “reinforcement” has the Skinnerian definition of an environmental consequence that increases the likelihood that the action will be repeated. Reinforcement is separable from reward, in that the former process does not depend on the immediate consequence of behavioral performance (i.e. outcome), in contrast to reward or punishment (White, 1989; Yin et al., 2006). Reinforcement does not appear to require selective attention to outcomes and shows specificity for specific movements and body parts (Yin, 2023). Empirical observations indicate that S–R learning proceeds in parallel with A-O learning, but at a slower rate (Yin, 2023), and this process has been incorporated into models of the role of different corticostriatal circuit roles in learning (Barnett et al., 2023). Reinforcement schedules and environmental conditions (e.g. stress) can increase the rate of S–R learning (DeRusso et al., 2010; Dias-Ferreira et al., 2009; Schwabe et al., 2011). Most animals develop a large repertoire of S–R based actions that are performed continuously in daily life. In silico modeling of the cellular mechanisms underlying the parallel memory processes have focused on changes in cortical input modulated by dopamine (Barnett et al., 2023; Reynolds & Wickens, 2002). The contributions of other factors, including thalamic input, remain to be incorporated into these models.

The net effect of S–R learning is to produce strongly ingrained action patterns that are produced without conscious attention to their production or outcomes (e.g. the action sequences involved in operating a bicycle in a well-trained rider). This contrasts with the conscious effort needed for A-O learning and affords the opportunity to perform well-learned actions while directing attention elsewhere. Because the two circuits can act independently, both A-O and S-R-based learning and action control are always available, allowing animals to switch between the two modes of control (Dezfouli et al., 2014). There is also evidence that humans use S–R strategies to choose between behavioral repertoires that are themselves A-O guided in nature (Cushman & Morris, 2015). This would limit behavioral repertoires, as stereotyped cognitive patterns lead individuals to default to a narrower set of previously rewarded actions. In this case, the “stimulus’ in the S–R mode is generated internally.

Separation of A-O versus S-R-driven behaviors has relied on experimental procedures based on operational definitions of the relative importance of the outcome in maintaining behavior. The most popular experimental manipulations to distinguish between A-O and S–R behavioral control center on the importance of the outcome. If changing outcome contingency alters behavior than the action is thought to be driven by an A-O process. In contrast S-R-based behavior is less sensitive to changes in outcome. Three experimental techniques are generally used to assess outcome sensitivity. Outcome devaluation, as the name implies, involves manipulations such as sensory-specific satiety or taste aversion that make the outcome less desirable (Yin, 2023; Yin & Knowlton, 2006). Another manipulation, known as contingency degradation, involves presenting rewards randomly with no contingency on action production (Yin, 2023; Yin & Knowlton, 2006). Finally, experimenters have used the omission paradigm mentioned above in which subjects are rewarded for withholding responses (Derusso et al., 2010). In all of these paradigms subjects that reduce their responding after manipulation of reward strength or contingency on action production are operationally defined as showing A-O-mediated actions. In contrast, S–R action control is operationally defined as that which continues despite the changes in reward status. While these different behavioral paradigms are often thought of as interchangeable tests of A-O-based behavior, the specifics of reward delivery, or not, in each paradigm makes it likely that different specific sets of neurons and circuits will be activated under each testing paradigm (nicely reviewed by Schreiner et al., 2020).

It is important to note that the terms “goal-directed” and “habitual” are often used to describe A-O and S–R action control. However, these terms have meanings outside of the realm of Neuroscience, and thus in many contexts it might be best to use the terms A-O and S–R. The concepts of “model-based” and “model-free” behavioral control strongly overlap with A-O and S–R control, respectively. These terms were generated largely based on modeling and work with human participants (Daw et al., 2005). The general idea is that behavior can be controlled either by an internal “model” of the relationship between actions and consequences (a cognitively demanding mode) or determined largely by the environment and past learning history (less cognitively demanding). While we often interpret experimental findings as indicating that actions are either A-O or S–R driven, the parallel function of the circuits implicated in these two processes allows for the possibility that complex behaviors could involve both types of control that change dynamically during action performance. Investigators interested in how brain circuits influence behavior should not lose sight of the importance of defining which circuit controls a given action regardless of the label we choose to use for that action. Indeed, the associative and sensorimotor circuits have roles in behaviors that are not always rigorously defined as A-O or S–R (Yin, 2023). Nonetheless these circuit influences are likely to play important roles in effects of alcohol and other misused substances.

The limbic circuit is certainly familiar to alcohol researchers, as the roles of different circuit subregions in alcohol-related behaviors have been studied extensively (Koob, 2014). This circuit is specialized for linking information about environmental events and internal states to control of actions. The cortical and cortical-like subregions that provide glutamatergic inputs to NAc, as well as target regions for NAc projections, are interconnected with brainstem areas such as the Nucleus Tractus Solitarius (NTS) and Locus coeruleus (LC) that provide the brain with information from the viscera and autonomic nervous system (Forstenpointner et al., 2022). This information helps to set the tone for affective states. The combination of this information with dopaminergic information about stimulus salience and reward sets the stage for S–O learning such that actions, and presumably affective states, become tied to discrete environmental stimuli (Asratyan, 1974; Yin, 2023). The dopaminergic modulation in NAc also plays a role in movement invigoration (Hughes, Bakhurin, et al., 2020; Panigrahi et al., 2015).

Acute alcohol effects on CTBG circuits start from the first exposure

Alcohol research has focused increasingly on brain circuit changes induced by the drug itself, drug seeking, and effects of prolonged exposure including tolerance, dependence and abstinence/withdrawal. Much is known about alcohol effects in the limbic CTBG circuitry, and the last 15–20 years have seen increased interest in effects on the dorsal circuitry and the consequences for A-O and S–R controlled behaviors (Corbit & Janak, 2016; Koob, 2014; Lovinger & Alvarez, 2017). The effects on dorsal striatum and other parts of the dorsal CTBG circuitry begin during initial exposure to alcohol. Most of the work to date has centered in the dorsal striatum as this BG subregion has a key role in modulating action selection and vigor through its influence on the rest of the circuitry (Yin, 2023). Increased understanding of the diversity of CTBG cell subtypes and synaptic influences on these subtypes indicates that more work will be needed to fully understand alcohol actions in this circuitry. Nevertheless, this body of knowledge is expanding as new tools for cell- and synapse-specific interrogation are incorporated into the research.

Alcohol effects on GABAergic transmission

Acute ethanol exposure in striatal brain slices alters GABAergic synaptic transmission, mainly via changes in GABA release. In the DMS/caudate nucleus, ethanol enhances GABA release onto MSNs, as it does in several other brain regions (Siggins et al., 2005; Wilcox, Cuzon Carlson, et al., 2014). However, inhibition of GABA release has been observed in DLS (Patton et al., 2016; Wilcox, Cuzon Carlson, et al., 2014). At synapses made by fast-spiking striatal interneurons (FSIs) onto MSNs this inhibition is long-lasting (persisting for longer than the period of EtOH application), and appears to involve an indirect effect, as it is prevented by antagonists of delta-type opiate receptors (DORs) (Patton et al., 2016). This DOR-dependent EtOH-induced long-term synaptic depression can be prevented by specific patterns of activation of FSIs, indicating that the effect of EtOH and DORs depends on prior presynaptic function and molecular signaling in presynaptic terminals (Patton et al., 2019). Notably, acute EtOH exposure also has indirect effects in DLS that could lead to net inhibition of striatal output. The striatal FSIs receive GABAergic synaptic inputs from GPe and reticular thalamus, and these inputs are potentiated in DLS during ethanol exposure via what appears to be a postsynaptic mechanism (Patton et al., 2023). How this enhanced FSI-mediated inhibition alters MSN function and alters net striatal output in vivo remains to be determined. It should be noted that while EtOH enhances the function of postsynaptic GABA_A receptors in many brain regions (see Abrahao, Salinas, & Lovinger, 2017 for review), this effect does not appear to be prominent in striatum.

Alcohol effects on glutamatergic transmission

Acute ethanol effects on striatal glutamatergic synaptic transmission are likewise complex. Ethanol inhibits the function of N-methyl-d-aspartate (NMDA)-type glutamate receptors (GluNs) in MSNs (Popp et al., 1998; Wang et al., 2007), as has been observed in neurons in many brain regions (Abrahao, Salinas, & Lovinger, 2017). NMDARs have a well-known role in synaptic plasticity, including a crucial contribution to long-term synaptic potentiation (LTP) that is implicated in learning and memory (Volianskis et al., 2015). Inhibition of NMDAR-dependent LTP was observed during acute EtOH exposure in adolescent striatal brain slices (Yin et al., 2007). However, results in adult male and female rats indicate either no effect or enhanced LTP (Avchalumov et al., 2020). The reason for this difference is not yet clear but it could be due to age, strain, species or sex differences. In the DMS, but not in ventral striatum, inhibition of NMDAR function switches to potentiation following removal of alcohol (Wang et al., 2007). This rapidly-developing effect of withdrawal involves phosphorylation of the GRIN2B GluN subunit by the Fyn tyrosine kinase, leading to an increased functional contribution of receptors containing this subunit. Indeed, subsequent studies provided evidence of an LTP-like increase in efficacy of glutamatergic synapses onto MSNs following chronic alcohol exposure and subsequent withdrawal, with particularly large effects on synapses onto dMSNs (Lagström et al., 2019; Wang et al., 2012, 2015). A recent study also showed enhanced action-related calcium responses in dMSNs, but not iMSNS, following chronic EtOH exposure (Baltz et al., 2023) (Fig. 2A). These findings indicate that either acute EtOH followed by withdrawal or chronic EtOH enhance activation of the direct pathway in the associative circuit.

Fig. 2.

Striatal direct and indirect pathways: Effects of alcohol and roles in alcohol drinking. A) Schematic diagram of the striatal projection neurons in the associative striatum (dorsomedial striatum, DMS) that make up the direct pathway (dMSN, purple) and the indirect pathway (iMSN, green) indicating their projection targets. The purple text indicates dMSN roles, while the green text indicates iMSN changes. B) Schematic diagram of the striatal projection neurons in the sensorimotor striatum (dorsolateral striatum, DLS) that make up the direct pathway (dMSN, purple) and the indirect pathway (iMSN, green) indicating their projection targets. The purple text indicates dMSN roles. (ACh – acetylcholine, CINs – cholinergic interneurons, D2Rs – D2-type dopamine receptors, EtOH - ethanol, GABA – gamma aminobutyric acid, GPe – globus pallidus external segment, LTD -long-term depression, LTP – long-term potentiation, mPFC – medial prefrontal cortex, OFC – orbitofrontal cortex, S–R – stimulus-response, SNr – substantia nigra pars reticulata).

In the lateral OFC (lOFC), acute alcohol exposure inhibits the function of presumed glutamatergic projection neurons (Nimitvilai-Roberts et al., 2021). This inhibition involves enhancement of a tonic current mediated by glycine, through a mechanism that involves dopamine activation of D1/5 receptors in astrocytes. This inhibition could contribute to suppression of associative circuit function and cognitive control during intoxication.

Alcohol effects on dopamine

Acute ethanol administration increases dopamine levels in vivo in different striatal subregions, with the largest effects in NAc, but also effects in DMS and much less effect in DLS (Boileau et al., 2003; Di Chiara & Imperato, 1985; Weiss et al., 1995). This increase appears to involve increased firing of midbrain dopaminergic neurons rather than increased release from dopaminergic axon terminals in striatum (reviewed in Morikawa & Morrisett, 2010). Notably, it is still unclear if EtOH has differential effects on different subpopulations of dopaminergic neurons in the SNc, as has been reported in VTA (Mrejeru et al., 2015). Local application via microdialysis has also been reported to increase dopamine levels in dorsal striatum, and strychnine-sensitive glycine receptors have been implicated in this effect (Clarke et al., 2015). Increases in NAc and DLS dopamine are observed following ethanol drinking and in response to environmental stimuli that predict alcohol availability (Gonzales & Weiss, 1998; Shnitko & Robinson, 2015). Increased phasic dopamine likely contributes to enhancing salience of these events, and dopamine roles in reward will increase the likelihood of alcohol-related behaviors. Alcohol-induced dopamine increases also contribute to enhanced activity/locomotion in response to low-moderate doses (Phillips & Shen, 1996). Interestingly, acute application of high EtOH concentrations to DS brain slices inhibits dopamine release (Budygin, Phillips, Robinson, et al., 2001), as has also been observed in NAc (Budygin, Phillips, Wightman, & Jones, 2001). A similar effect was seen in DS in vivo when examining electrically-stimulated dopamine release (Budygin, Phillips, Wightman, & Jones, 2001). These findings indicate a low potency effect of EtOH that may limit dopamine release via a direct effect on presynaptic terminals during strong intoxication.

While EtOH effects on dopamine release are well characterized, we know somewhat less about how the drug alters signaling through dopamine receptors. Early studies indicated that relatively high concentrations of EtOH enhance dopamine stimulated adenylyl cyclase (AC) activation in striatal tissue, a mechanism that likely involves D1 dopamine receptors (Rabin & Molinoff, 1981). This effect appeared to be due to enhanced AC activity rather than dopamine receptors per se, and this conclusion was supported by later work (Hoffman et al., 1983). These mechanisms could influence alcohol effects on striatal function and related behaviors and require further study in this context. Recent studies indicate that D2 receptors regulate alcohol-induced locomotor hyperactivity and risk avoidance behavior, as these behavioral phenotypes are altered in mice with lower levels of these receptors on iMSNs (Bocarsly et al., 2019, 2024). These mice also show greater escalation of alcohol intake and stronger alcohol preference (Bocarsly et al., 2019).

Alcohol effects on other neuromodulators

Relatively little is known about alcohol effects on extracellular levels and signaling activated by other key striatal neuromodulators. However, recent studies have provided more information about alterations in acetylcholine (ACh), a prominent small molecule striatal neuromodulator (Goldberg & Wilson, 2017). The CINs provide the bulk of the ACh, although cholinergic afferents from the pendunculopontine brainstem nucleus (PPN) also innervate striatum (Dautan et al., 2014). The CIN axons provide extensive coverage of the striatum, supplying ACh throughout the structure, with spatial inhomogeneity mainly in the “striosome” subcompartments (Goldberg & Wilson, 2017). These neurons exhibit ongoing spontaneous action potential activity, with occasional pauses in firing and rebounds following these pauses (Nosaka & Wickens, 2022). Thus, the striatum is essentially bathed in ACh most of the time, at least at low concentrations, with periodic local decreases and increases in levels. Relief from the cholinergic tone and responses to the transient ACh increases play key roles in modulation activity of striatal neurons.

There is relatively little information about how acute ethanol alters dorsal striatal ACh levels. Acute alcohol at relatively high concentrations inhibits ACh release elicited by activation of strychnine-sensitive glycine receptors and NMDA-type glutamate receptors in rat dorsal striatum (Darstein et al., 1997, 1998). It is unclear if this is due to inhibition of the receptors themselves or inhibition of some aspect of ACh release. In addition, acute EtOH inhibition of the tonic firing of CINs has been reported (Blomeley et al., 2011). Activity of the ACh synthetic enzyme choline acetyltransferase (ChAT) is increased in rat dorsal striatum following a single dose of ethanol (Ebel et al., 1979), and thus decreased release may not be due to decreased synthesis. High affinity choline uptake is accelerated rat dorsal striatum following acute alcohol injection, but this effect was not observed during acute incubation with alcohol Hunt et al. (1979), suggesting that it is not involved in the inhibition of receptor-driven ACh release. It should be noted that these reports did not indicate which dorsal striatal subregion was examined, and thus it is not yet clear if acute EtOH effects on cholinergic transmission show any spatial specificity.

Alcohol may also alter striatal cholinergic transmission via effects on ACh receptors. A wide variety of both muscarinic (mAChR) and nicotinic (nAChR) ACh receptors are expressed in striatum (Goldberg & Wilson, 2017). The mAChRs are G protein-coupled receptors (GPCRs) that alter intracellular signaling, and 5 receptor subtypes exist. The mAChR1, 3 and 5 subtypes are coupled to Gq-type G proteins. Activation of these receptors stimulates phospholipases with general effects that favor transient increases in neuronal excitability and neurotransmitter release mainly through altered ion channel function (Nestler and Duman, 1999). The 2 and 4 mAChR subtypes couple to Gi/o G proteins with net effects including inhibition of adenylyl cyclase, activation of potassium channels (most notably Gi receptor-gated, GIRK-type channels), and inhibition of voltage-gated calcium channels (Nestler and Duman, 1999). The net effects of this signaling are generally to favor transient reductions in neuronal activity and neurotransmitter release. The GPCRs can also initiate or modulate longer-lasting effects through activation of long-lasting synaptic plasticity (Lovinger et al., 2022). Among the most prominent effects of mAChRs in striatum are inhibition of ACh and glutamate release by Gi/o coupled receptors on axon terminals, and both increases and decreases in dopamine release are mediated by different receptor subtypes (Hersch et al., 1994; Shin et al., 2015, 2017; Threlfell & Cragg, 2011; Threlfell et al., 2010). Intracellular signaling, including calcium increases, produced by M1 AChRs contributes to synaptic plasticity (Augustin et al., 2018; Centonze et al., 2003; Wang et al., 2006).

The nAChRs are multimeric ligand-gated ion channels that contain 5 subunits per protein complex. Several subunits and combinations thereof are expressed in striatum including those containing the 2,4,5,6 and 7 α subunits and the β2 subunit (Abbondanza et al., 2024; Exley et al., 2012). Among the most prominent nAChR roles in striatum is the stimulation of glutamate and dopamine release via receptors located on presynaptic terminals. The nAChR subtypes on dopaminergic afferents express alpha6 along with β2 subunits (Exley et al., 2008; Jones et al., 2001; Kramer et al., 2022; Zhou et al., 2001). The α7 subunit-containing nicotinic receptors appear to mediate increases in glutamate release from cortical terminals (Mateo et al., 2017; Wonnacott et al., 2006) although a role for alpha4-containing receptors was also recently proposed (Morgenstern et al., 2022). There are also nAChRs on GABAergic striatal interneurons that depolarize the cells (Kocaturk et al., 2022).

Acetylcholine receptors have been implicated in acute EtOH effects on synaptic transmission in dorsal striatum. Antagonists of nAChRs prevent EtOH-induced decreases in synaptically-driven neuronal activation, and long-term facilitation of transmission after removal of acute EtOH also involves these receptors as well as mAChRs (Adermark, Clarke, et al., 2011). Both m and nAChRs are also implicated in intrastriatal mechanisms underlying enhanced dopamine release (Loftén et al., 2021). Additional work is clearly needed to determine the full extent of acute EtOH alterations in striatal ACh level and functions and the relevance for acute intoxication, as well as the sensitivity in different striatal subregions.

Alcohol effects on the globus pallidus

In many models of BG function, the Globus pallidus external segment (GPe) has a central role in action control (Giossi et al., 2024). GPe neurons are tonically active and thus send continuous output to brain regions including the STN, SNr, striatum and others (Kita & Jaeger, 2017). The GABAergic synapses made by these neurons are thought to produce a steady inhibition of activity of their target neurons, providing a baseline check on various components of the CTBG circuit. Most obvious is the suppression of output from the SNr via the indirect pathway, that helps to control unwanted actions (Gerfen & Bolam, 2017). However, GPe GABAergic synapses in striatum, particularly those on MSNs, also appear to have roles in suppressing actions (Gu et al., 2020; Li et al., 2006; Mallet et al., 2016; Nikolaou et al., 2013). Dynamic changes in GPe influence CTBG circuit function mainly through GABA-induced pauses in activity and subsequent rebounds in the rate of neuronal firing. This pause-rebound sequence can also affect synchrony of firing in regions such as STN and SNr that are targeted by GPe GABAergic synapses. The net effect of pauses in SNr neuron firing will be to temporarily reduce BG output, producing downstream alterations in thalamic and cortical activity.

The GPe projections to striatum have been viewed as providing feedback to MSNs. However, given the fact that GPe neurons are tonically active, it seems less likely that pallidostriatal GABAergic projections respond to initial striatopallidal input. It is more probable that pallidostriatal afferents produce tonic inhibition of MSNs and other striatal neurons that can be relieved during pauses in GPe neuron firing. Thus, one role of striatopallidal GABAergic synapses would be to temporarily pause GPe neuron firing, relieving this tonic inhibition. Subsequent burst firing of GPe neurons would produce a small temporal window of increased inhibition in striatum. The in vivo consequences of these patterns of GPe and striatal activity are poorly understood and are the subject of vigorous ongoing research in the BG field.

One factor that complicates our understanding of pallidostriatal projections is the finding that GPe neuron afferents target both MSNs and FSINs. Thus, the pallidal input could either inhibit or disinhibit striatal output depending on the pattern of the actions on these different striatal neuron subtypes. It is now clear that only a subset of pallidal neurons project to striatum, and there has been extensive focus on identifying the neuronal subtypes that give rise to these projections (Kita & Jaeger, 2017). The general consensus is that the pallidostriatal neurons show low spontaneous frequency firing, often below 10 Hz (Abdi et al., 2015). They can be identified by expression of transcription factors such as Npas1, FoxP2 and Lhx6, at least when expression of these factors is assessed by labeling based on expression early in development (Abdi et al., 2015). However, there is as yet no clear consensus on which marker best identifies GPe neurons that innervate MSNs (the so-called arkypallidal neurons) vs. FSIs (Abdi et al., 2015; Abrahao & Lovinger, 2018; Glajch et al., 2016; Mastro et al., 2014).

There is increasing interest in alcohol effects on the GPe that may influence decision making and alcohol drinking (reviewed in Hong et al., 2021). Ethanol effects on GPe neuronal activity and synaptic transmission have been measured in single neurons, brain slices and in vivo. Early studies indicated that acute EtOH exposure enhances postsynaptic responses mediated by GABA_A receptors in isolated ventral pallidal neurons (Criswell et al., 1993, 1995). Measurements of immediate early gene expression indicated increased cFos levels in GPe neurons following EtOH exposure (Hitzemann & Hitzemann, 1997; Kolodziejska-Akiyama et al., 2005; Kozell et al., 2005; Vilpoux et al., 2009). However, it remained unclear which GPe neuron subtypes were affected by EtOH. The timing of changes in GPe neuron activity following acute and chronic EtOH exposure was also not explored in detail in these studies.

Examination of acute EtOH effects on action potential firing frequency of GPe neurons revealed inhibition of firing in neurons that exhibited low baseline firing levels, with little effect on neurons that normally fired at higher frequencies (Abrahao et el., 2017a). Subsequent experiments revealed that the subclass of neurons inhibited by EtOH express the Npas1 and Lhx6 transcription factors, while neurons that expressed the parvalbumin Ca²⁺ binding protein were not inhibited. In vivo electrophysiological recordings also indicated that EtOH preferentially inhibited GPe neurons that fire at low frequencies, as opposed to high frequency firing neurons (Abrahao, Chancey, et al., 2017). The Npas1-expressing neurons give rise to arkypallidal synapses and some of the Lhx6-expressing neurons also project to striatum (Abdi et al., 2015; Glajch et al., 2016; Mastro et al., 2014). Thus, acute inhibition of GPe neurons may reduce tonic inhibition of striatal neurons, with implications for action control. However, it remains to be determined whether the EtOH effects are predominantly on neurons that innervate MSNs vs. FSIs.

Inhibition of firing in Npas1 or Lhx6-expressing GPe neurons appears to involve enhanced function of the Ca²⁺ and voltage-gated potassium channel known as BK (Abrahao, Chancey, et al., 2017). Blockade of these channels prevents EtOH inhibition of firing. Recording of single BK channels in Npas1 GPe neurons revealed that EtOH enhances the probability of channel opening without any change in channel conductance. This finding follows from earlier studies in isolated channels and cells showing that EtOH can enhance the opening of BK channels with specific subunit compositions (Dopico et al., 2016). The expression of specific BK subunits in low-frequency firing BK neurons needs to be determined for comparison with these earlier molecular biological studies (although the electrophysiology indicates that they express the channel pore-forming α subunit). The consequences of these EtOH actions in GPe for CTBG circuit function and behavior are still being explored, but one possibility is that inhibition of GABA release from pallidostriatal neurons reduces a brake on action production, as will be discussed later in this review. It will also be important to determine if effects differ in GPe subregions within the associative and sensorimotor circuits.

Chronic alcohol exposure and alcohol intake alter CTBG circuit function

When rodents are exposed to alcohol or drink alcohol for days to weeks, transmission is altered at synapses in several cortical and basal ganglia regions (Abrahao, Salinas, & Lovinger, 2017; Lovinger & Roberto, 2023). Alterations in neuronal excitability are also observed in response to such exposure (Abrahao, Salinas, & Lovinger, 2017). One general effect of this exposure is to reduce the behavioral influence of the associative circuit, while enhancing that of the sensorimotor circuit (Barker et al., 2015; Corbit et al., 2012; Gremel & Lovinger, 2017; Lovinger & Alvarez, 2017; Van Skike et al., 2019). However, this pattern is not always so straightforward, and effects within each circuit depend on the alcohol exposure regime, including dose, timing and developmental status. As mentioned in the previous section on acute alcohol actions, additional work is still needed to unravel neuronal and synaptic subtype-specific effects of chronic alcohol, but intriguing work in this area has been generated in recent years.

Chronic alcohol effects on associative cortices

Glutamatergic projection neurons in the OFC show variable changes in excitability following chronic alcohol exposure in rodents. Hyperexcitability of mouse lOFC neurons that project to DMS has been observed following chronic inhalational ethanol exposure (Nimitvilai et al., 2016; Nimitvilai-Roberts et al., 2023). The enhanced excitability of these neurons is accompanied by loss of the neuromodulatory effect of monoamines that normally dampens activity (Nimitvilai et al., 2018). The D2 dopamine receptors are one subtype affected in this way, but the general effects on the physiological influence of several Gi/o-coupled receptors suggests that the loss of modulation may result from altered intracellular signaling rather than change in receptors per se. A subset of DMS neurons that are targeted by these projections as well as BLA glutamatergic inputs also show enhanced excitability during withdrawal following chronic alcohol exposure (Nimitvilai-Roberts et al., 2025). These changes could enhance activity in the associative circuit. However, it remains to be seen how this altered activity contributes to circuit function. Furthermore, there are conflicting results regarding chronic EtOH effects on mouse lOFC neuronal excitability, as Renteria et al. (2018) reported decreased excitability following chronic inhalational ethanol exposure (CIE). Similar results were obtained monkey OFC after chronic alcohol intake (Nimitvilai et al., 2017). It is not clear what factors underlie the apparent discrepancies in these findings. Clearly species differences may play a part, but differences in time of examination following withdrawal, cortical layer targeted, or even diet may also contribute. It is notable in this context that in vivo recordings of OFC neuronal spiking in CIE-treated animals revealed increased activity associated with lever pressing, but decreased activity associated with outcomes (Cazares et al., 2021). Thus, in vivo changes in activity of OFC neurons following CIE likely depend not only on changes in neuronal excitability but also the synaptic inputs that drive these neurons in different contexts.

Altered glutamatergic transmission following chronic EtOH exposure has been observed in neurons in both cortical and striatal regions of the associative circuit. In the mPFC, enhanced function of NMDA-type glutamate receptors was observed during withdrawal following chronic EtOH exposure (Kroener et al., 2012) (Fig. 3A). This increase was accompanied by increases in expression of NR subunits and enhancement of the late phase of NMDAR-dependent LTP in mPFC glutamatergic neurons. Reduced excitability of GABAergic interneurons has also been observed in the prelimbic mPFC during withdrawal following chronic alcohol exposure of drinking (Dao et al., 2020; Hughes, Crofton, et al., 2020; Joffe et al., 2020; Thompson et al., 2023), likely contributing to overall hyperexcitability of this associative network cortical region. The firing of mPFC neurons develops patterns related to alcohol intake as rodents experience chronic drinking (Linsenbardt et al., 2019; Timme et al., 2022). Wistar rat mPFC neurons are strongly activated by presentation of a light that signals alcohol availability (Linsenbardt et al., 2019). Interestingly, this response is strongly diminished in alcohol preferring P rats, suggesting a loss of mPFC influence over alcohol intake in this high drinking species (Linsenbardt et al., 2019). However, mPFC neurons in P rats actually show enhanced activity to signals indicating that alcohol seeking is available in an aversion-resistant “compulsive” drinking paradigm (Timme et al., 2022), indicating that neuronal representations of alcohol availability in this region may depend on the strength of the drive to drink.

Fig. 3.

Alcohol-induced alterations in different corticostriatal projections related to Action-Outcome (A–O) and Stimulus-Response (S–R) control of behaviors and alcohol intake. A) Schematic diagram of corticostriatal projections from prelimbic cortex (PrL, orange), lateral orbitofrontal cortex (lOFC, green) and primary motor cortex (M1, blue), and chronic EtOH-induced alterations in neuronal activity and synaptic transmission in dorsomedial striatum (DMS) thought to contribute to A-O control of actions and drinking. B) Schematic diagram of the three corticostriatal projections and chronic EtOH-induced alterations in neuronal activity and synaptic transmission in DMS and dorsomedial striatum (DLS) thought to contribute to S–R control of actions and drinking. (INS - Insula, LTD – long-term depression, LTP – long-term potentiation).

In the OFC, LTP-like changes in glutamatergic transmission onto projection neurons were observed following chronic EtOH exposure (Nimitvilai et al., 2016). Increased expression and function of NMDA may contribute to this effect (Radke et al., 2017). Reduced expression of D4-type dopamine receptors were also observed in this region (Volkow et al., 2007). The glutamatergic projections from OFC to DMS exhibit decreased efficacy following chronic alcohol exposure (Renteria et al., 2018), with implications for cognitive performance and alcohol drinking (Nippert et al., 2024; Shields & Gremel, 2020). In vivo measurements of the function of these OFC-DMS terminals indicate that reduced responding mainly occurs during reward retrieval (Renteria et al., 2021). A loss of D1 dopamine receptor signaling in DMS MSNs coincident with an increase in retrograde endocannabinoid signaling at OFC inputs to these neurons appears to contribute to this reduction in transmission. Evidence of alterations in GABAergic and glutamatergic transmission have been observed in non-human primate OFC, but the functional consequences of these changes are not yet fully understood (Acosta et al., 2010, 2012; Hemby et al., 2006; Nimitvilai et al., 2017).

Chronic alcohol effects in striatum

Chronic EtOH exposure generally results in molecular and cellular alterations that increase excitation and reduce inhibition of the striatal projection neurons (Lovinger & Alvarez, 2017). Among these EtOH-induced adaptations are increased MSN intrinsic excitability, loss of presynaptic long-term depression (LTD) at glutamatergic cortical synapses onto MSNs, and decreased GABAergic input onto MSNs as observed in both mice and monkeys (Cuzon Carlson et al., 2011, 2018; Depoy et al., 2013, 2015; Wilcox, Cuzon Carlson, et al., 2014) (Fig. 3B). The synaptic changes generally appear to induce “disinhibition” of MSN output. However, mixed, and sex dependent, effects on glutamatergic transmission have also been observed in both DLS and DMS (Rangel-Barajas et al., 2021). As mentioned above, LTP-like enhancement of glutamatergic synapses onto MSNs was observed in DMS following chronic EtOH drinking (Lagström et al., 2019; Ma et al., 2017; T et al., 2012, 2015) (Fig. 3A). This is the sort of increase in synaptic efficacy that should enhance goal-directed actions as predicted by in silico models (Barnett et al., 2023; Reynolds & Wickens, 2002). In the sensorimotor striatum, disinhibition has been postulated to increase activity in the circuit in response to environmental events and context based on previous reinforcement history (Lovinger & Alvarez, 2017; Lovinger & Gremel, 2021) (Fig. 3B). It is worth noting that many of the changes in GABAergic synaptic function observed with chronic EtOH exposure are qualitatively similar to the effects produced by acute EtOH exposure (e.g. decreased inhibitory transmission) (Patton et al., 2016; Wilcox, Cuzon Carlson, et al., 2014). Thus, unlike synapses in other brain regions, chronic EtOH does not produce compensatory synaptic changes at striatal GABAergic synapses, but rather something akin to allostasis in which synapses reach a new and lower functional status throughout the process of exposure. However, newer evidence of EtOH effects on interneurons suggests more nuanced changes in GABAergic function following chronic EtOH exposure (Patton et al., 2023).

Glutamatergic afferents from insular cortex (INS) to DLS are also implicated in control of EtOH intake. These synapses undergo LTD involving a long-lasting decreases in glutamate release. This form of LTD can be induced by activation of mu opiate receptors, and either a single alcohol exposure or chronic exposure disrupts this plasticity (Muñoz et al., 2018). A recent study showed that the activity of INS axons in the DLS is strongly elevated around the onset of EtOH drinking bouts during the Drinking in the Dark (DID) procedure in male mice (Haggerty & Atwood, 2024) (Fig. 3B). Recordings of neuronal firing revealed increases in activity-correlated licks for both saccharin and EtOH intake, with increases in a sustained firing pattern related to quinine-resistant “compulsive” alcohol drinking that was distinct from that seen during other drinking conditions (Starski et al., 2024). These findings support the idea that insula activity and output to sensorimotor striatum help to drive alcohol seeking that may be S–R driven.

Given that cortical and thalamic neurons innervate striatal cholinergic interneurons as well as MSNs, it is important to understand EtOH effects on these synapses. Chronic EtOH also alters glutamatergic transmission onto cholinergic interneurons in the DMS (Li et al., 2024; Ma et al., 2022) (Fig. 2A). The efficacy of thalamic glutamatergic synapses onto CINs is decreased following chronic alcohol drinking (Ma et al., 2022). This contributes to a decrease in CIN-mediated inhibition of D1-MSNs. Similar decreases in glutamatergic synaptic input from orbitofrontal cortex to CINs are also observed after chronic intermittent EtOH intake (Li et al., 2024). This alteration leads to a reduction in the burst-pause firing pattern of CINs. Surprisingly, the altered CIN function due to this decrease in synaptic function had a larger impact on D2-MSNs than D1-MSNs. The mechanisms that underlie the differential effects on MSNs of changes in the two glutamatergic inputs remain to be elucidated. In addition, the conditions under which each of these effects predominate following chronic EtOH drinking is still unclear.

In vivo recordings also indicate that the activity of neurons in the dorsal striatum is associated with various phases of alcohol self-administration. Using fixed ratio (FR) and variable interval (VI) alcohol reward schedules, the latter of which generally foster S–R learning, Fanelli et al. (2013) found that neurons in rat DMS were active during alcohol delivery, while DLS neurons were generally active around lever presses. These findings indicate that both associative and sensorimotor striatum remain active in relation to alcohol seeking and drinking even after prolonged self-administration experience. It should be noted that the reward in this study was a relatively low dose of alcohol (10%). Furthermore, the rats trained on both FR and VI schedules were still sensitive to outcome devaluation, while the VI-train rats had reduced sensitivity to contingency degradation. Thus, the extent of A-O and S-R-driven alcohol seeking and taking in this study is not clear.

Chronic alcohol effects on striatal dopamine

Alterations in dopamine release within DS, and the factors that regulate this release, have also been examined recently using ex vivo brain slice electrochemical approaches. Chronic, but not acute, consumption of moderate levels of EtOH enhances dopamine release in mouse DS, and this changes to reduced release following forced abstinence (Salinas et al., 2022). In male rats, increased dopamine uptake was observed in dorsal striatum following chronic exposure (Budygin et al., 2007). Altered presynaptic regulation of release by D2 autoreceptors and nAChRs is also observed with this drinking regimen. In macaque monkey striatum dopamine release and presynaptic modulation thereof is also altered following long-term alcohol drinking, but effects depend on subregion and sex. In the associative striatum (caudate) and sensorimotor striatum (putamen) long-term drinking generally decreased dopamine release in rhesus and cynomolgus macaques, but this reduction was noticeably absent in female putamen of rhesus macaques (Salinas et al., 2021; Siciliano et al., 2015). Inhibition of release by D2 autoreceptors was reduced in males, but not females, regardless of drinking duration or abstinence status. Notably, a change in this autoreceptor function was not observed in previous studies of alcohol-drinking male cynomolgus macaque monkeys (Budygin et al., 2003; Siciliano et al., 2016). Dopamine uptake was enhanced in both subregions of female, but not male, rhesus macaque striatum, but notably this finding differed from previous studies showing reduced uptake in caudate from alcohol-drinking cynomolgus macaques (Siciliano et al., 2015, 2016). It is unclear if some of the differences relative to previous studies are due to species differences or differences in the alcohol drinking protocol.

Alcohol drinking also alters heterologous modulation of dopamine release by other neurotransmitters includind opiate peptides and ACh. Kappa opiate receptors inhibit dopamine release via an apparent presynaptic mechanism, and this modulation was enhanced in caudate from long-term drinking cynomolgus macaques (Siciliano et al., 2015). As in rodents, nAChRs that express the β2 subunit regulate dopamine release in macaque caudate and putamen (Salinas et al., 2021). This regulation was not altered following chronic drinking. The inhibitory effect of acute EtOH application to DS slices, mentioned above, is lost in caudate from chronic alcohol-drinking cynomolgus macaques (Budygin et al., 2003). Overall, these findings support the idea that dopamine release in DS is decreased following chronic alcohol drinking, at least after a short period of abstinence. This may contribute to a “hypodopaminergic” state in DS following drinking and abstinence, but the consistent finding that D2 autoreceptor function is diminished following chronic drinking suggests that this one important brake on dopamine release is also compromised in DS. Furthermore, the loss of acute EtOH inhibition could lead to higher dopamine levels during strong intoxication. However, it would be important to measure dopamine release in vivo to determine what happens in the intact striatum. The implications of these changes in DS for altered dopamine roles in behavior and drinking remain to be fully explored.

Molecular Mechanisms Implicated in Chronic Alcohol Actions in Striatum.

The following paragraphs describe the current state of knowledge about chronic alcohol effects on expression of molecules and molecular function in striatum, with an emphasis on neurons. It must be kept mind that many of the studies discussed in this section examined molecular changes in tissue without examination of cell type specificity. Indeed, the majority of these ground-breaking studies were carried out before techniques for examining molecular changes within specific cellular subtypes were generally available. Given the rapid development of techniques for cell type-specific molecular examination, there will be a need to revisit many of these findings to determine if the molecular impacts are greatest in specific neuronal subtypes (e.g. direct and indirect pathway MSNs, interneurons and glia). This level of analysis will facilitate tying molecular changes to physiology of the appropriate neurons, as well as cell-specific molecular manipulations designed to alter circuit and behavioral outcomes of chronic alcohol exposure. Thankfully, there are already nice examples of this type of cell-type specific research as can be seen from discussion of specific experiments.

Several alcohol-induced molecular changes related to altered synaptic signaling have been identified in dorsal striatum. Expression and function of striatal protein kinases and associated proteins is altered following chronic EtOH exposure or drinking with resultant changes in protein phosphorylation that modify striatal synaptic communication. Increased function of NMDA-type glutamate receptors is a well-known consequence of chronic drug exposure, presumably as a compensatory response to acute inhibition of this receptor by EtOH (reviewed in Abrahao, Salinas, & Lovinger, 2017). In the DMS, the Fyn kinase has been shown to participate in this process through phosphorylation of the NR2B receptor subunit leading to increased receptor function (Gibb et al., 2011). This process begins just after cessation of acute EtOH exposure (Wang et al., 2007), as the inhibition of NMDAR function changes to enhanced function. This process continues throughout chronic exposure and withdrawal. The increased NMDAR function induces a long-term increase in the expression and function of synaptic AMPA-type glutamate receptors in a process that closely resembles LTP (Wang et al., 2010, 2012) (Fig. 2A). The alcohol-induced changes in Fyn phosphorylation of NR2B are stronger in dMSNs compared to iMSNs (Ehinger et al., 2021). Interestingly, these molecular adaptations are not observed in the sensorimotor DLS and thus may be implicated in A-O driven processes related to alcohol intake, as will be discussed later in this review.

Chronic high level alcohol intake also increases activity of the mammalian target of rapamycin 2 (mTORC2) protein kinase in DMS (Laguesse et al., 2018). This kinase regulates actin polymerization which contributes to the structural rearrangement of dendritic spines that form the postsynaptic elements of glutamatergic synapses. Thus, the enhanced mTORC2 activity contributes to increased expression of filamentous actin (F-actin) and increased density of mature spines on MSN dendrites in DMS. The small guanosine triphosphate (GTP)-binding protein named Ras-related C3 botulinum toxin substrate protein (known as Rac1) that regulates F-actin formation is activated in DMS following prolonged EtOH intake (Hoisington et al., 2024). Thus, both phosphorylation and small GTPase protein function are implicated in EtOH effects on cytoskeletal rearrangements that contribute to chronic alcohol-induced dendritic spine changes. Notably, chronic alcohol drinking induced increases in dendrite length and complexity were also observed (Wang et al., 2015) but did not depend on mTORC2 activity (Laguesse et al., 2018). These alterations in dendrite and spine morphology likely contribute to the increased glutamatergic transmission observed on MSNs, with particular importance for increased excitation of dMSNs (Wang et al., 2015) (Fig. 2A).

Regulation of protein phosphorylation by striatal phosphatases is also altered in DMS, but not DLS, following chronic EtOH exposure. This change in regulation appears to work in concert with enhanced Fyn kinase-mediated phosphorylation to alter MSN synaptic transmission and physiology. For example, increased phosphorylation of the STriatal-Enriched protein tyrosine Phosphatase 61 (STEP₆₁) is observed following chronic consumption (Darcq et al., 2014). This phosphorylation decreases phosphatase activity, facilitating the Fyn-mediated phosphorylation of NR2B described above. Similar changes in the function of the protein tyrosine phosphatase α (PTPα) also appear to contribute to this process (Ben Hamida et al., 2013). Thus, EtOH exposure sets into motion multiple molecular signaling changes that contribute to enhanced cortical drive of DMS MSNs.

Peptidergic growth factors have also been implicated in chronic EtOH effects on striatum. Brain derived neurotrophic factor (BDNF) is produced by cortical afferents to striatum, and this production is reduced following chronic EtOH drinking (Logrip et al., 2009, 2015). BDNF signals through two receptors, the tyrosine kinase tropomyosin receptor kinase B (TrKB, the high affinity receptor associated with most of the trophic BDNF actions) (Patapoutian & Reichardt, 2001), and the P75 protein, a lower affinity BDNF receptor (Kraemer et al., 2014). These receptors activate multiple molecular cascades that alter protein expression affecting a variety of neuronal functions. Altered BDNF expression was observed in striatum following chronic alcohol exposure (McGough et al., 2004), and signaling through TrkB is implicated in neuronal changes in DLS during early phases of alcohol exposure (Logrip et al., 2015). However, more prolonged and escalating EtOH intake enriches synaptic levels of P75, but not TrkB, in DLS (Darcq et al., 2016). This receptor shift in sensorimotor striatum appears to alter how BDNF contributes to S-R-driven alcohol drinking, which will be discussed later in this review.

Recent studies have also used “omic” approaches to identify changes in the molecular makeup of cells in associative and sensorimotor circuits and signaling related to chronic alcohol exposure or consumption. Assessment of protein expression in the DLS, DMS and NAc was performed following acute or protracted forced abstinence in mice allowed to drink in a two-bottle choice paradigm (Duffus et al., 2024). The largest changes were observed in proteins implicated in processes including neurodegeneration, metabolism, cellular organization, protein translation and molecular transport. Unexpectedly, proteins that are diagnostic of inflammation were not greatly altered despite the fact that inflammatory mechanisms have been widely postulated to underlie many of the deleterious effects of prolonged alcohol exposure. This set of protein and pathway changes may be unique to striatum, and it should also be noted that the overall alcohol consumption levels in this model were relatively modest. However, the findings point to interesting changes in the structure and function of neuronal elements, including synaptic alterations that might contribute to maintenance of drinking or cognitive impairment following drinking and abstinence. In the OFC, expression of RNA encoding the presynaptic protein adaptor complex 2, as well as expression of the protein itself correlates with alcohol intake in mice and Alcohol use Disorder (AUD) in humans (Mulholland et al., 2023). With this information in hand it will now be important to determine which specific protein changes have the biggest consequences for cellular function.

Comparisons of lines with differential alcohol drinking and preference profiles may identify molecular networks that predispose the animals to striatal phenotypes that contribute to enhanced drinking. To this end, Grecco and coworkers examined RNA and protein expression differences as well as phosphoprotein differences in dorsal striatum (without subregion dissection) of selectively bred High Alcohol Preferring (HAP) and Low Alcohol Preferring (LAP) mice. All mice were alcohol-naïve in this case (Grecco et al., 2021). The molecular analysis in this study pointed to differences in striatal cellular organization, cytoskeleton and synaptic transmission that may contribute to differential risk for high alcohol intake in the HAP mice. Decreased levels of the cannabinoid type 1 receptor (CB1) were observed in dorsal striatum of HAP3 vs. LAP3 mice (Millie et al., 2020), perhaps leading to more facile loss of presynaptic depression in the high drinkers.

Chronic alcohol effects on human striatal and cellular morphology

Studies of human brain morphology and cellular status provide important information about chronic alcohol effects that contribute to altered CTBG circuit function and related behavioral changes. Several stuch studies indicate AUD-related changes in the putamen. The volume of the putamen is decreased in individuals with AUD relative to healthy controls as measured with in vivo brain imaging (Tomasi et al., 2021). Decreased left putamen volume, along with reduced volume of some cortical brain regions, is also predictive of AUD (Hahn et al., 2022). This is part of a general trend for decreased gray matter associated with AUD. A recent postmortem study of human putamen suggests that glial cell function is altered in individuals with AUD relative to controls (Rasool et al., 2024). The morphology of microglia and astrocytes was altered in ways that indicate activation of these glial cell types. These findings and the consequences of such changes deserve more attention as glial changes can signal neuroinflammation and can contribute to changes in neuronal function, as reviewed by Hong et al. (2021).

Chronic alcohol and SNr function

Recent studies have also begun to examine EtOH effects in the SNr, the primary output nucleus of the BG. Enhanced GABAergic transmission from DLS to lateral SNr was observed during forced abstinence following four weeks of CIE (Sitzia et al., 2024) (Fig. 2B). In contrast there is no such change in the DMS-medial SNr synapses. No changes were observed in glutamatergic synaptic inputs from STN to SNr following the same chronic EtOH + forced abstinence protocol in either lateral or medial reticulata areas.

CTBG circuit roles in alcohol effects on behavior and alcohol drinking

Studies in laboratory animal models

Examination of chronic alcohol effects on learning and task performance in rodents has generally revealed increasing reliance on the sensorimotor circuit, especially following alcohol intake or exposure (Barker et al., 2015; Corbit et al., 2012; Gremel & Lovinger, 2017; Lovinger & Alvarez, 2017; Van Skike et al., 2019). This change in circuit control results in changes in action performance strategy including increased use of egocentric and S–R strategies in maze tasks and IC (aka operant) tasks, respectively. Interestingly, even exposure to a context in which alcohol was previously given can engender S–R driven behavior (Ostlund et al., 2010). Decreased behavioral flexibility has also been observed following alcohol exposure, and changes in associative circuitry function are implicated in this drug-induced cognitive impairment (Cheng, Magnard, Langdon, Lee, & Janak, 2025; reviewed in Nippert et al., 2024). Increased performance of well-learned actions has also been observed (Depoy et al., 2013, 2015), even when the impact of reward status was not specifically examined.

Chronic alcohol drinking or exposure has also been shown to engender drinking that is more resistant to devaluation (i.e. habitual alcohol intake) (Corbit et al., 2012; Barker et al., 2020; Dickinson et al., 2002; Giuliano et al., 2021; Mangieri et al., 2012; O’Tousa and Grahame, 2014; Renteria et al., 2020; but see Samson et al., 2004; Shillinglaw et al., 2014). This has been seen with reward devaluation procedures that involve both satiety and conditioned taste aversion (Barker et al., 2020; Corbit et al., 2012; Dickinson et al., 2002; Mangieri et al., 2012). This S–R driven alcohol drinking is associated with increases in some parameters associated with burst-like licking behavior (Renteria et al., 2020). The development of S–R control of alcohol seeking and drinking may depend on the level of alcohol exposure, as the associative circuit continues to influence alcohol intake following chronic low-level intake (Lu et al., 2019; Ma et al., 2018; Roltsch Hellard et al., 2019; Wang et al., 2015). However, it is unlikely that this is the only factor that contributes to development of S-R-based strategies, as this type of control of seeking has been observed with prolonged drinking at relatively low levels (Corbit et al., 2012). Interactions between exposure level and duration may play a role. It has been hypothesized that the development of S-R-based seeking is driven by the changes in excitability and synaptic transmission in the sensorimotor circuit that were discussed earlier in this review. However, changes in associative circuitry may disrupt cognitive control over alcohol intake that could also contribute to greater reliance on external stimuli and environmental context in the determination of alcohol seeking and drinking (Baltz et al., 2023; Renteria et al., 2018).

Chronic alcohol effects in associative circuitry

Recent studies have examined the effect of manipulating different molecules, neurons and circuits within the CTBG circuit to determine their roles in alcohol effects on cognition and their contribution to alcohol seeking and intake. Decreased neuronal population activity related to behavioral control, and increased activity related to alcohol availability are observed in mPFC neurons during aversion-resistant “compulsive” drinking in P rats, and enhancing this activity reduces such drinking (Timme et al., 2022). Inducing LTP at mPFC synapses onto dMSNs in DMS produces a long-lasting enhancement of alcohol drinking (Ma et al., 2018) (Fig. 2A and . 3A). This finding supports the idea that enhanced excitatory transmission at this synapse contributes to higher alcohol intake. Inducing long-term depression (LTD) at these synapses has the opposite effect, inducing a sustained decrease in intake (Ma et al., 2018) (Fig. 2A). However, the effect of inducing LTD appears to depend on the MSN that is targeted by the afferents, as induction of this form of plasticity at mPFC synapses in DMS can also increase EtOH seeking and drinking via a mechanism involving D2 receptors and possibly involving synapses on iMSNs (Roltsch Hellard et al., 2019) (Fig. 2B). In contrast to potentiation of mPFC inputs to DMS, inducing LTP at inputs from OFC reduced EtOH seeking and drinking (Cheng et al., 2021) (Fig. 2A). This finding fits with the idea that these OFC inputs are depressed following chronic alcohol exposure and restoring baseline efficacy levels can re-establish top-down control over alcohol intake as well as other behaviors (Renteria et al., 2018) (Fig. 3B). It is unclear if the drinking observed in these studies was driven by A-O or S–R processes. It must also be noted that the mPFC and OFC project to different regions of DMS (Fig. 3), and thus the pattern of control of drinking likely differs depending on the exact associative subcircuit that is examined.

Additional evidence indicates that alterations in OFC activity are associated with relapse to alcohol seeking. The activity of rat OFC neurons was enhanced during EtOH seeking in an operant task following extensive intermittent alcohol drinking, and the increases were greater in animals that were in the high drinking category (Hernandez & Moorman, 2020) (Fig. 2A). Inhibition of OFC neurons using pharmacological or chemogenetic approaches in rats decreased cue-induced reinstatement of EtOH self-administration, although notably this inhibition had no effect on free-choice drinking or lever-pressing for alcohol (Arinze & Moorman, 2020; Hernandez et al., 2020). Similar results were obtained with context-induced renewal of seeking when pharmacological inactivation of OFC was performed, and the alcohol context activated neurons in this region (Bianchi et al., 2018). These results may reflect a lack of control of the OFC and connected associative circuitry over alcohol seeking and intake in experienced drinkers, and a more specific role in relapse induced by environmental conditions conditioned to signal alcohol reward. Perhaps aspects of the associative circuit may be re-activated when particular environmental events occur, leading to reinvigoration of A-O-driven seeking and intake.

Chronic alcohol exposure followed by protracted forced abstinence led to enhanced activity of neurons in the premotor (M2) cortex that project to the DMS measured during performance of a well learned food-reinforced instrumental task that required long-lasting lever presses (Schreiner et al., 2023). This increased activity was associated with a decrease in the ability of mice to use recent experience for action control. Reducing the activity of the DMS-projecting M2 neurons during task performance normalized the reliance on recent performance. These findings indicate that hyperactivity in this corticostriatal component of the associative circuit may produce unclear signals related to ongoing outcomes, that ultimately biases animals toward using a more S-R-based pressing strategy.

Glutamatergic transmission in DMS has also been postulated to play a role in limiting EtOH intake, perhaps by supporting continued associative circuit control and limiting sensorimotor control. In this context Bauer and coworkers recently examined the effect on EtOH intake and locomotion of blocking AMPA-type glutamate receptors in DMS using the DID procedure in a mouse strain bred for high alcohol intake (the Crossed High Alcohol Preferring, cHAP strain) (Bauer et al., 2024). They found increased intake after antagonist injection, with no accompanying change in locomotion. These data are intriguing, but additional studies are warranted to determine which glutamatergic synapses contribute to this effect. The cHAP mice may have a bias toward sensorimotor circuit control of EtOH intake, but interestingly AMPA antagonist injection into DLS had no effect on intake in these mice (Bauer et al., 2024). Adenosine signaling through A2A receptors in DMS also appears to play a role in maintaining A-O control of EtOH intake (Nam et al., 2013).

The alcohol-induced alterations in glutamatergic input to striatal CINS, discussed previously, also influence cognitive function and alcohol intake. Reduced function of thalamic glutamatergic inputs to DMS CINs appears to be involved in alcohol impairment of cognitive flexibility (Ma et al., 2022) (Fig. 2A). Interestingly, impairment of OFC synapses onto DMS CINs after chronic alcohol drinking was also implicated in cognitive flexibility deficits (Li et al., 2024). This is a bit surprising given the difference in the net alcohol effects on d- and iMSNs of changes in the two glutamatergic inputs to CINs discussed above. Additional work will be required to sort out the circuit effects underlying CIN roles in alcohol-induced cognitive impairment. Enhancing the OFC input to DMS CINs following chronic alcohol drinking reduced alcohol seeking and drinking in free choice and operant paradigms (Li et al., 2024) (Fig. 2A). Thus, CINs may be another cellular target for therapies designed to control alcohol intake.

Changes in the associative circuitry have also been implicated in the transition from goal-directed to habitual alcohol intake. Enhancement of action-related activity at the terminals of OFC-DMS projections was observed following chronic alcohol exposure, while decreased activity was seen during reward retrieval in these same mice (Renteria et al., 2021) (Fig. 3B). These presynaptic changes were associated with loss of effect of devaluation of lever pressing for food. Decreased dopaminergic signaling and increased retrograde endocannabinoid/CB1 receptor signaling appeared to contribute to the changes in presynaptic activity and loss of devaluation effects. These findings indicate that chronic EtOH exposure engages endocannabinoid signaling system in DMS that suppresses glutamate release at OFC terminals and promotes S-R-based behavior (Gremel et al., 2016). Decreased endocannabinoid-mediated synaptic depression in DLS has also been observed following chronic alcohol drinking and exposure, suggesting similar EtOH effects on this type of synaptic plasticity occurs in both the associative and sensorimotor circuits (Adermark, Jonsson, et al., 2011; Depoy et al., 2013, 2015) (Fig. 3B).

A recent study also found that alcohol enhances activity of medial OFC interneurons, and this appears to reduce binge drinking (Gimenez-Gomez et al., 2024). Suppressing the activity of these neurons leads to enhanced drinking, and projections to the mediodorsal thalamus are involved in the actions of this pathway.

Chronic alcohol effects in sensorimotor circuitry

Studies of alcohol seeking in rodents have implicated cellular and molecular substrates in the DLS in development and maintenance of S–R based seeking. Dopamine increases in DLS are associated with operant lever-pressing for alcohol delivery on a variable interval schedule (Shnitko & Robinson, 2015). Corbit and coworkers showed that blocking D2 dopamine or AMPA-type glutamate receptors resulted in devaluation of operant responses for EtOH under conditions where these responses would normally be reward insensitive (Corbit et al., 2014). In Sprague Dawley rats, injection of a dopamine receptor antagonist into anterior DLS reduced alcohol drinking in an adjunctive water/alcohol intake protocol, but only after prolonged drinking experience that resulted in excessive intake (Marti-Prats et al., 2023). Boehm and coworkers have also found evidence for a role of striatal glutamatergic transmission in alcohol drinking. For example, they found that blocking AMPA receptors in DLS reduced binge-like EtOH intake in C57Bl6J mice that have a relatively short history of drinking experience (Bauer, McVey, Germano, et al., 2022). As mentioned previously, the glutamatergic input from INS to DLS neurons is implicated in controlling EtOH intake. Indeed, stimulation of this pathway decreases drinking in the DID paradigm (Haggerty et al., 2022). These findings indicate that ongoing glutamatergic and dopaminergic transmission in DLS are necessary to maintain sensorimotor circuit control of actions involved in alcohol seeking. In humans, stronger activation of putamen during instrumental task performance was observed in those with AUD (Sjoerds et al., 2013).

Opiate receptor antagonists are widely used for AUD treatment (O’Malley and O’Connor, 2011)). However, there is relatively little information about whether these compounds are more effective for drinking driven by A-O or S–R processes. In this context, data indicating that naltrexone treatment was less effective in reducing alcohol intake in rats that self-administer on a schedule that promotes S–R based intake is interesting (Hay et al., 2013). This finding suggests that endogenous opiate signaling may be important for maintaining A-O driven intake.

Intracellular Molecular Mechanisms Implicated in alcohol-induced behavioral changes

The growth factors and intracellular signaling pathways discussed earlier in this review play important roles in striatal control of alcohol seeking and drinking. Elevating BDNF in DLS prevents excessive drinking (Jeanblanc et al., 2009, 2013), while reducing DLS expression of the growth factor promotes greater intake (Jeanblanc et al., 2009; Logrip et al., 2015). Activation of TrkB leading to MAP kinase activation appears to underlie these effects (Jeanblanc et al., 2013). The activation of this pathway leads to increased signaling through the dynorphin/kappa opiate receptor signaling system in dorsal striatum that contributes to decreased EtOH intake (Logrip et al., 2008). However, after prolonged EtOH drinking when BDNF signaling switches to p75 the growth factor may begin to promote excessive alcohol intake, indicated by the observation that reducing p75 expression or inhibiting the receptor decreases binge-like alcohol drinking (Darcq et al., 2016). These findings indicate that BDNF signaling through TrkB may initially serve to limit alcohol intake, but prolonged intake lessens this effect and signaling that promotes drinking is enhanced. It will be important to determine the cellular and circuit mechanisms in striatum underlying this transition.

Protein kinases expressed in striatal MSNs are also implicated in control of drinking. The Leucine-rich repeat kinase 2 (Lrrk2) is expressed by striatal MSNs and implicated in Parkinson’s Disease (Cookson et al., 2010). It also contributes to signaling by D1 dopamine receptors in dMSNs, presumably through its ability to regulate receptor trafficking and intracellular cAMP signaling (Parisiadou et al., 2014; Rassu et al., 2017). Recent studies indicated that Lrrk2 expression level is associated with alcohol intake in mice and humans (da Silva e Silva et al., 2016; Martins de Carvalho et al., 2020). Deletion of Lrrk2 from dMSNs enhanced alcohol drinking and increased locomotor activation in response to injected EtOH (da Silva et al., 2024). Alcohol drinking in dMSN Lrrk2 KO mice also became more resistant to increased task demand or pairing with aversive outcomes, suggesting that loss of signaling by this kinase leads to more “compulsive” drinking. While it is not clear which striatal subregions contribute to this effect, enhanced neuronal excitability was observed in DMS from the dMSN Lrrk2 KO mice. These findings support the idea that enhanced activation of DMS dMSNs enhances alcohol intake and other alcohol-related behaviors (Wang et al., 2015).

As discussed previously, the Fyn tyrosine kinase is implicated in alcohol-induced striatal synaptic plasticity, and signaling through this kinase in dMSNs in the DMS also regulates alcohol drinking (Ehinger et al., 2021; Ron & Berger, 2018). Reducing the expression of the PTPα phosphatase that regulates Fyn phosphorylation of NR2B reduces alcohol consumption, especially at high EtOH concentrations (Ben Hamida et al., 2012), presumably by preventing alcohol-induced Fyn actions. In contrast, reducing expression of the STEP₆₁ phosphatase that counteracts Fyn effects leads to increased alcohol intake (Darcq et al., 2014; Legastelois et al., 2015). The STEP₆₁ knockout mice also become less sensitive to the reduction in alcohol intake normally produced by quinine adulteration (Legastelois et al., 2015). These findings support the idea that the Fyn kinase-mediated induction of plasticity at glutamatergic synapses in DMS promotes increased alcohol intake (Ron & Berger, 2018).

Other striatal intracellular signaling pathways that mediate effects of chronic EtOH also appear to play roles in alcohol-related behaviors. The Rac-1 small GTPase protein appears to be involved in learning A-O associations for alcohol, but not sucrose (Hoisington et al., 2024). Additional studies in this area will be needed to reveal the many striatal signaling pathways involved in chronic EtOH effects on behavior, as well as the physiological mechanisms that are altered by changes in these signaling pathways.

The infralimbic prefrontal cortex glutamatergic afferents to the NAc have also been implicated in enhancement of S–R strategies, possibly through dopamine and glutamatergic signaling (Barker et al., 2015; Trantham-Davidson et al., 2017; Tu et al., 2007; Woodward, 2000; Yin et al., 2007). Chronic EtOH exposure decreases expression of the metabotropic glutamate receptor type 2 (mGlu2) in rat infralimbic cortex projection neurons (Meinhardt et al., 2013). Interestingly, mGlu2 mRNA is also decreased in a homologous brain region in humans with AUD. This is associated with escalated alcohol intake. Replacement of these receptors in rat reduces intake to levels observed prior to chronic EtOH exposure. Increased glutamate release from infralimbic inputs to NAc due to loss of mGlu2 autoreceptor regulation at these corticostriatal terminals appears to contribute to this effect. A recent report suggests that rats with persistent and punishment-resistant alcohol seeking behavior also show altered mGlu2 function in the prelimbic PFC, but surprisingly this was accompanied by altered transmission within this cortical region but not to the NAc (Domi et al., 2024). It should be noted that mGlu2 knockout mice show escalated alcohol intake, and loss of mGlu2 expression in the alcohol-preferring P rat line contributes to enhanced drinking in this line (Zhou et al., 2013). It remains to be determined if mGlu2 changes, at the IL-NAc synapses or elsewhere, contribute to S–R based drinking strategies.

Surprising recent findings and other CTBG circuit roles in behavior

A few surprising recent findings will likely force a reevaluation of simple models of associative and sensorimotor striatal roles in alcohol drinking. As mentioned above, Boehm and coworkers examined binge-like drinking in C57 black 6/J mice and the cHAP mice. They found that altering DLS function and glutamatergic transmission reduces binge-like alcohol drinking when animals first start drinking but has less effect after prolonged drinking experience (Bauer et al. 2022a, 2022b). The DLS may have a particular role in “front-loading” behavior, i.e. rapid initiation of drinking bouts, and also appears to contribute to quinine-insensitive drinking. The lack of effect of intra-DLS AMPAR antagonist injections in cHAP mice, mentioned above, also indicate that this glutamatergic synapses in DLS may not have a role in strongly compulsive drinking (Bauer et al., 2024). These findings suggest that the DLS regulation of drinking begins fairly early in exposure, at least when binge drinking is involved. This finding is not too surprising if one considers that the sensorimotor circuit contributes to learning of new behaviors in parallel with the associative circuit and that EtOH alters neuronal and synaptic function in DLS during initial acute exposure, as discussed earlier in this review. It will be important to determine which aspects of drinking behavior involve the two parallel circuits. Perhaps DLS disinhibition during binges leads to rapid drinking onset and resistance to aversive factors that would normally disrupt drinking, while other parts of the circuit, or even other circuits, may contribute to compulsive alcohol intake after long-term drinking experience.

Experiments in mice also indicate a link between sensitization to alcohol effects and activation of the sensorimotor striatum (González-Marín et al., 2020). Sensitization refers to the process through which a response (usually behavioral) to an environmental event or substance becomes larger after multiple exposures. In this study, female swiss webster mice showed differing degrees of sensitized locomotion after exposure to EtOH. Stronger sensitization was associated with faster acquisition of operant EtOH self-administration and higher break points in a progressive ratio self-administration paradigm (possibly indicating stronger motivation for alcohol). This increase was accompanied by altered alcohol sensitivity of cFos expression in DLS. Although only correlative, these findings suggest that locomotor sensitization and incentive sensitization develop in parallel, and may set the stage for stable neuronal activity in DLS that foster S–R related alcohol seeking.

Another surprising new finding concerning sensorimotor circuitry involvement in alcohol drinking comes from work in rhesus macaques performed as part of the work of the Integrative Neuroscience Consortium on Alcoholism Stress (INIAstress) (Grant et al., 2022). Investigators used the Designer Receptor Exclusively Activated by Designer Drug (DREADD) approach to modulate the activity of neurons (mostly of which are MSNs) in the monkey putamen. The hM4Di DREADD that generally reduces neuronal function and synaptic transmission was expressed in putamen and ethanol intake was examined during the initial phases of learning to drink. Activation of the receptor enhanced the rate at which animals consumed alcohol, which generally led to higher blood alcohol levels. This finding indicates that the function of putamen MSNs may moderate drinking levels during early phases of drug use. These findings are somewhat unexpected given that putamen was thought to be involved only after long-term chronic drinking. However, the sensorimotor circuitry is known to contribute to production of ongoing actions, and thus may work to stabilize drinking behavior based on past experiences. Disrupting the function of this circuitry could enhance the role of other circuits in behavioral control, in this case promoting faster than normal intake perhaps driven by reward, novelty or stress relief.

GPe and striatonigral pathway roles in alcohol-related behavioral changes

There has been less work examining the role of GPe in alcohol seeking and drinking in comparison to striatum. However, an intriguing recent report indicated that genetic ablation of arkypallidal projections from lateral GPe to DLS facilitated shifts from A-O to S–R based instrumental actions reinforced by either sucrose of ethanol (Baker et al., 2023). Thus, the inhibition of DLS output provided by GPe GABAergic synapses may serve as a break on use of the sensorimotor circuit for seeking of substances including alcohol.

The impact of EtOH inhibition of GPe neurons on CTBG circuit function and behavior is still largely unknown. Experiments in humans and rodents indicate roles for GPe and the arkypallidal neurons in response inhibition (Gu et al., 2020; Li et al., 2006; Mallet et al., 2016; Nikolaou et al., 2013). In a “stop-signal” task participants (human or rodent) are taught to withhold a response when a particular environmental stimulus (e.g. a light or tone) is presented just prior to normal response production. Ethanol reduces the stimulus-induced action suppression in human volunteers even at low concentrations (McPhee & Hendershot, 2023; Nikolaou et al., 2013; Shinozaki et al., 2024). This ability is also impaired in those with, and at risk for, AUD (Groman et al., 2009; Wilcox, Dekonenko, et al., 2014). These findings indicate that alcohol contributes to impairment of the ability to stop performance of unwanted actions. This is certainly an important phenotype of excessive drinking and AUD. Individuals will make risky decisions, including continuing to drink when impaired, under the influence of alcohol. These problems may be exacerbated in AUD.

The behavioral consequences of the enhanced direct pathway output specific to the sensorimotor striatum are just starting to be examined. Sitzia and coworkers examined how changes in this pathway may contribute to circuit control of instrumental behavior (Sitzia et al., 2024), following up on the aforementioned earlier studies indicating an enhanced role for sensorimotor circuits in control of instrumental/operant performance following chronic EtOH exposure (CIE). Instrumental lever-pressing behavior was examined in mice during withdrawal following chronic intermittent EtOH vapor exposure, with air-alone exposed mice serving as controls. The mice were trained on a series of schedules ranging from fixed ratio 1 to random ratio 20 (RR20). Chronic EtOH exposure did not alter performance in any phase of the task. To determine if the sensorimotor striatonigral direct pathway neurons contributed to task performance, the hM4Di DREADD was expressed in DLS several weeks prior to CIE (Sitzia et al., 2024). The DREADD agonist clozapine-N-oxide (CNO) was injected peripherally daily during instrumental training and performance. Injection of CNO did not alter instrumental performance in any phase of the task in mice given chronic air exposure. In contrast, CNO injection reduced lever pressing and rewards earned under the RR20 schedule in mice exposed to CIE and subsequent withdrawal. This finding indicates that chronic alcohol exposure enhances the role of the direct pathway projection in sensorimotor striatum in instrumental task performance, particularly under a task schedule where performance is normally controlled by the associative circuit (Fig. 2B). Enhanced striatonigral GABAergic transmission may contribute to this stronger sensorimotor circuit role, but other changes in the circuit within striatum may also contribute.

CTBG circuit roles in effects of alcohol drinking in humans

Studies in humans have examined correlations between alcohol-related behaviors and measurements of brain activity (usually with fMRI) or structure (various in vivo imaging and postmortem approaches). Song et al. (2024) observed altered cortical connectivity to both NAc and putamen in AUD participants, measured with fMRI. Resting activity in the putamen and associated cortical regions is negatively correlated with AUD severity (Gerhardt et al., 2022). Hyperactivity in caudate and putamen in response to stress-eliciting stimuli is also observed in men with AUD (Smith et al., 2023). Non-treatment-seeking individuals with AUD who scored high on scales for relief and habit-based drinking showed greater activation in DS in response to presentation of alcohol-related images (Burnette et al., 2021), and stronger activation was also seen in heavy drinking college students (Dager et al., 2013). In addition, activity in the DS elicited by presentation of such images was strongly correlated with pain catastrophizing, a phenotype that is associated with several aspects of AUD (Nieto et al., 2022). Individuals with AUD show stronger implicit associations among alcohol-related items relative to healthy controls, and this categorization was associated with elevated activity in the Putamen and Insula (Ames et al., 2014). Such associations may indicate a role for the sensorimotor circuit in unconscious preference for certain environmental stimuli that trigger alcohol-related behaviors.

In vivo structural imaging studies indicate prolonged reduction in the volume of OFC in individuals with AUD, and lower volume is associated with risk of relapse (Cardenas et al., 2011; Demiracka et al., 2011; Durazzo et al., 2011; Laakso et al., 2002; Wang et al., 2016). There is evidence that normal volumes are restored after abstinence (Cardenas et al., 2011; O’Neill et al., 2001). Thus, decreased volume is associated with AUD symptoms and time course. White matter loss appears to account for some of the loss in OFC volume, as indicated by both in vivo imaging and postmortem studies (Harris et al., 2008; Miguel-Hidalgo et al., 2006), although decreases in neurons/gray matter are also observed (Miguel-Hidalgo et al., 2006). Markers of neuroinflammation have also been observed in postmortem AUD patient OFC, indicating that such processes may contribute to cellular toxicity and structural changes (Crews et al., 2013; Qin & Crews, 2012; Vetreno et al., 2013).

Reduced OFC function is also associated with AUD in humans. Persistent decreases in blood flow and glucose utilization in OFC have been reported even well into abstinence (Catafau et al., 1999; Volkow et al., 1994, 1997). Molecular and pharmacological studies have also provided evidence of altered GABAergic inhibitory transmission and neuromodulation in human AUD OFC (Hommer et al., 1997; Jin et al., 2012; Mitchell et al., 2012; Volkow et al., 1993, 1997, 2007).

Behavioral changes consistent with reduced OFC function are also observed in individuals with AUD. These include impaired performance in reversal tasks and tasks that require inhibition of actions (e.g. go/no go tasks) (Fillmore & Rush, 2006; Fortier et al., 2008, 2009; Kamarajan et al., 2005; Smith et al., 2014; Stavro et al., 2013; Vanes et al., 2014). Cognitive tests such as the Iowa Gambling Test also reveal reduced performance by individuals with AUD (Bechara et al., 2001; Brevers et al., 2014; Cantrell et al., 2008; Dom et al., 2006; Fein et al., 2004; Le Berre et al., 2014; Miranda et al., 2009; Noël et al., 2007). These impairments resemble effects of OFC damage (reviewed in Shields & Gremel, 2020). In studies explicitly examining the relationship between OFC function and behavior, correlations between the two measures break down in AUD individuals (Boettiger et al., 2007; Duka et al., 2011; Forbes et al., 2014). Enhanced OFC response to alcohol versus non-alcohol related environmental cues is also associated with AUD (Ernst et al., 2013; Hermann et al., 2006; Myrick et al., 2008; Reinhard et al., 2015; Wrase et al., 2002), although this has not always been a consistent finding likely due to different experimental conditions and participant ages (George et al., 2001; Ihssen et al., 2010; Lingford-Hughes et al., 2006; Schneider et al., 2001; Seo et al., 2013; Tapert et al., 2004). Increased OFC cue-reactivity has been associated with behavioral effects including increased craving and increased signs of interest in alcohol-related pictures (Claus et al., 2011; Ernst et al., 2013), as well as increased risk of heavy drinking and relapse (Dager et al., 2014; Myrick et al., 2004, 2008; Reinhard et al., 2015; Schacht et al., 2014). These findings suggest that long-term drinking and AUD may shift the assessment of environmental events toward those related to alcohol, although admittedly these types of studies in humans are correlative.

In a gambling-type task, individuals with AUD showed enhanced activation in dorsal striatum when expecting large rewards, and this response was strongest in the Caudate nucleus (van Holst et al., 2014). This finding may indicate that reward-based associative circuit activation is intact and accessible under certain environmental conditions in AUD. Altered low-frequency fluctuations in the fMRI BOLD signal were also observed in males with AUD (Dai et al., 2023). Le et al. (2024) observed altered putamen function and connectivity related to pain as a drinking motive. This finding may suggest that the sensorimotor circuit is involved in associations between aversive environmental events and alcohol. Positive alcohol expectancy is associated with decreased functional connectivity between the insular cortex and putamen (Le et al., 2022). Treatment with naltrexone, an opiate receptor antagonist with efficacy in AUD treatment, altered connectivity in networks that included caudate and putamen striatal subregions. These changes were correlated with performance on a reward-driven task (Spencer et al., 2023). Nalmafene, a naltrexone-like antagonist, also altered regional networks involving caudate and putamen in an AUD sample (Grundinger et al., 2022). Altering the way that these components of the CTBG circuit respond to environmental rewards may thus be an avenue to reduce AUD symptoms. There is also evidence that alcohol intake is influenced more strongly by impulsivity than either A-O (model-based) or S–R (model-free) processes (Nebe et al., 2018). It should be noted that this study was performed in social drinkers, and thus it is unclear if different results would be obtained in those with AUD. Overall, the role of model-free/S–R learning and striatal synaptic transmission/modulation in the development of AUD requires more study (as reviewed in Huys et al., 2016; McKim et al., 2016).

It is also important to consider individual differences in the likelihood of adopting S–R versus A-O strategies in response to alcohol exposure. Such differences have been observed in rodent studies (Barker et al., 2014), where some responses to conditioned stimuli appear to predict development of S-R-based alcohol seeking. However, the extent to which these differences contribute to human alcohol seeking is less clear.

Future directions

Despite the considerable progress made in understanding how the different CTBG circuits are affected by alcohol, and in turn contribute to alcohol-related behaviors, there is still much to be learned. As emphasized previously, the cellular and circuit physiological consequences of many of the alcohol-induced alterations in neurotransmitters, receptors and intracellular signaling are not well understood. This work is important as it should provide candidate molecular targets for therapeutics designed to treat aspects of AUD.

Several components of the associative and sensorimotor circuits have also not been explored in the context of alcohol actions and alcohol-related behaviors. The hyperdirect pathway from cortex to STN provides a fast route for cortical inhibition of BG output. Depending on the cortical source, this circuit can de-emphasize specific actions. Disruption of hyperdirect pathway function could be a key component of loss of inhibitory control in the context of alcohol seeking and drinking. Conversely, strengthening of this pathway may suppress behaviors that might be more adaptive in the face of alcohol-related environmental cues and contexts. Thus, examination of this circuitry is worthy of additional study. Indeed, little is known about alcohol effects in the STN itself, although work from the Baunez group suggests roles in motivation for EtOH and escalation of EtOH intake, perhaps by increasing preference over other rewards (Lardeux & Baunez, 2008; Pelloux & Baunez, 2017). The STN and associated brain regions are activated by voluntary alcohol consumption (Dudek et al., 2015), and STN is one of the brain regions activated during withdrawal following repeated EtOH exposures (Chen et al., 2008). A recent imaging study in humans suggest that network connectivity to STN predicts binge drinking, perhaps associated with reduced response inhibition (Morris et al., 2016). Determining the molecular and cellular substrates of alcohol in this BG region will be important, along with determining how STN activity works within the hyperdirect pathway and the larger CTBG network to alter alcohol-related behaviors, such as altered response inhibition and drinking (reviewed in Bunse et al., 2014).

Cortical regions outside of the prefrontal cortex have received relatively little attention in studies of alcohol effects and AUD. In particular, little is known about the effects in primary motor cortex (M1) despite the prominent role of this cortical region in the sensorimotor circuit, although recent studies are providing new information in this area. A recent study by Che et al. (2024) recorded from M1 neurons in anesthetized mice following 4–5 weeks of forced alcohol exposure, and observed decreased firing, predominantly in deep layer presumed interneurons. This change could lead to a change in excitatory/inhibitory balance in favor of greater excitation in M1. A similar chronic EtOH exposure paradigm activated astrocytes and decreased dendritic spines in M1, and these effects were reduced by persistent running on a treadmill (Liu et al., 2024), although dendritic spine changes were not consistently detected in M1 following adult alcohol exposure in rodents (Amodeo et al., 2021). A chronic moderate binge drinking protocol led to decreased glucose utilization in M1 as measured in vivo (Rapp et al., 2022). A single low dose of alcohol increased gamma oscillations in human motor cortex during finger movement (Campbell et al., 2014). A similar dose also decreased connectivity between the supplementary motor cortex and M1 (Luchtmann et al., 2013). These findings indicate that M1 function is likely impaired during acute alcohol exposure/intoxication. Postmortem measurement of RNA transcripts in motor cortex of human AUD brains suggests effects on myelination, neurodegeneration and cellular trafficking (Mayfield et al., 2002), and decreased expression of NMDA receptor-coding RNA (Ridge et al., 2008). Changes in GABA receptor subunit expression in human AUD postmortem and rodent M1 following chronic EtOH exposure were also observed, and the rodent changes generally resembled changes in other cortical regions (Buckley et al., 2006; Grobin et al., 2000). In humans cortical M1 stimulation with transcranial magnetic stimulation indicates enhanced excitability/disinhibition in individuals with prolonged alcohol exposure or AUD. Examination of acute alcohol effects in other cortical regions outside prefrontal cortex has been limited to date (Harrison et al., 2017). It will be important not only to understand how both acute and chronic EtOH exposure affect different cortical regions, but also their synaptic connections and functional integration with circuits that influence a variety of behaviors.

Effects of EtOH on aspects of thalamostriatal function have been discussed in this review. However, much remains to be gleaned about thalamic participation in different CTBG circuits and how that participation is altered by acute and chronic alcohol exposure. The ultimate hope is that more in-depth research in both humans and non-human organisms will provide an integrated picture of how the different CTBG circuits respond to alcohol and how their control of behavior is influenced by alcohol exposure and contributes to AUD. This will require concerted efforts in cell and molecular biology and physiological analysis of specific cell types, along with measurement and manipulation of circuit function in concert with measurements of behaviors and alcohol-related cognitive functions such as craving. The ultimate hope is that a better understanding of how disordered CTBG circuit function contributes to AUD will lead to new and efficacious treatments, both pharmacological and non-pharmacological, for this devastating disorder.

Finally, it should be evident that the discussion of alcohol actions in CTBG circuits and the roles of different cell types in this review was highly neuron-centric. This is mainly because far less is known about alcohol interactions with other cell types in these brain regions, although the role of glia in alcohol actions is a growing research focus (Erickson et al., 2021; Harder et al., 2024; Hong et al., 2021). For example, Hong et al. (2020) provided evidence that enhancing glial expression of an adenosine transporter in DMS promoted a shift toward more S–R control of reward seeking, including for EtOH. The roles of glia, hemodynamics and even influences of non-neural organs on CTBG function and alcohol actions is a much-needed area for future research.