Introduction

Alterations to executive function as well as continued and dysregulated alcohol use are hallmarks of Alcohol Use Disorder (AUD) and are thought to result in part from effects of chronic alcohol on cortical and striatal brain circuits. For example, goal-directed action control is controlled by the dorsal medial region of the striatum (DMS) which has long been shown to be altered by chronic alcohol. Prior investigations have suggested activity within DMS is differently changed following chronic alcohol and that such changes are responsible for a loss of behavioral control. However, whether and how chronic alcohol alters activity in various DMS neuron populations during action control is not clear.

The DMS is a heterogeneous structure, largely composed of two GABAergic output pathways often referred to as the direct and indirect pathways. The direct pathway contains GABAergic spiny neurons (SPNs) that can be differentiated through their monosynaptic connection with the Substantia Nigra Pars Reticulata as well as their expression of the dopamine type-1 receptor (D1-SPNs). In contrast, the GABAergic SPNs of the indirect pathway make a monosynaptic connection with the Globus Pallidus, and express the dopamine type-2 receptor (D2- SPNs). Modulating these output pathways are various interneuron populations, including fast-spiking interneurons that are often identified and targeted by their expression of Parvalbumin (PV interneurons). These interneurons provide a strong inhibitory clamp of SPN activity. The striatum as a whole is driven by glutamatergic inputs largely from cortex and thalamus. In particular, the DMS receives input from associative cortex and thalamus, and is thus poised to integrate and modulate information important for goal-directed control. Indeed, numerous findings across species show that intact DMS is critical for goal-directed behavior.

Chronic alcohol has been reported to disrupt goal-directed control over both general reward-related behavior as well as a loss of goal-directed control over alcohol self-administration. This has been associated with specific changes in corticostriatal transmission, as well as with general changes to transmission within DMS. Generally, chronic alcohol has been associated with reduced GABAergic transmission in DMS, and increased glutamatergic transmission, although recent findings have suggested this may depend on the specific synapses interrogated. Pertinent to the present investigation, a prior examination of chronic alcohol drinking reported potentiated glutamatergic transmission onto the direct pathway, and potentiated GABAergic transmission onto the indirect pathway. Mimicking or countering these changes in transmission in vivo increased or decreased alcohol drinking, respectively. Furthermore, both acute and chronic alcohol has been shown to affect transmission from and onto PV interneurons and striatal PV interneurons support compulsive alcohol consumption. As it is hypothesized that alcohol self-administration recruits striatal function that eventually becomes less dependent on DMS, it suggests chronic alcohol may lead to an imbalance between DMS activity that results in dysfunctional goal-directed control.

Here we examine action-related calcium activity of DMS direct and indirect pathways as well as local PV interneurons following chronic alcohol exposure. We employed a reinforcement schedule that biases goal-directed action control in mice. Prior alcohol exposure has been found to disrupt goal-directed control under this schedule. While we used bulk calcium fiber photometry from genetically or projection-defined populations as a proxy for activity, it should be noted that within striatum, bulk calcium activity is not linearly related to the number of action potentials. However, bulk calcium measurements do provide insight into dendritic as well as somatic calcium activity that likely reflects incoming transmission and is related to plasticity processes. In line with prior hypotheses, we find that prior chronic alcohol induced long-lasting increases in the calcium activity of the direct pathway during decision-making, but reduced calcium activity of the indirect pathway and PV fast-spiking interneurons. Our findings support a hypothesis of DMS population activity imbalance resulting in reduced contribution of DMS to behavioral control.

Methods and materials

Subjects

In the present studies, C57Bl/6 J mice (Jackson Labs), Pvalb^tm1(cre)Arbr (PV Cre+) (Jackson Labs), B6.FVB(Cg)-Tg(Drd1-cre)EY266Gsat/Mmucd (D1-Cre+) (MMMRC Repository), and B6.FVB(Cg)-Tg(Adora2a-cre)KG139Gsat/Mmucd (A2a-Cre) (MMMRC Repository) mice were used for experiments. Adult (>8 weeks) male and female mice were housed in groups of 2–4, with mouse chow (Labdiet 5015) and water ad libitum, and were kept on a 14 hr light/10 hr dark cycle. While both males and female mice were included, experiments were not powered to identify sex difference, influence of circulating hormones in males or females, or effect of genotype on the calcium modulation observed. All experiments were conducted during the light phase of the light cycle. All experiments were approved by the University of California San Diego Institutional Animal Care and Use Committee and were carried out in accordance with the National Institutes of Health (NIH) “Principles of Laboratory Care”. Investigators were not blind to the experimental groups. The Animal Care and Use Committee of the University of California, San Diego approved all experiments and experiments were conducted according to the National Institutes of Health (NIH) “Principles of Laboratory Care” guidelines.

Surgical procedures and histological verification

All mice were anesthetized with isoflurane (1–2%) and were given stereotaxically guided injections into the DMS (coordinates from Bregma in mm: anterior [A], +0.5; medial [M], ±1.5; ventral [V]: −3.25) via 500 nl volume Hamilton syringes (Reno, NV). To localize GCaMP6s expression to direct pathway neurons we took two approaches. In one, C57Bl/6 J mice were injected with 300nl AAV-Ef1a-Cre-WPRE in the SNr (coordinates from Bregma in mm: anterior [A], −3.4; medial [M], ±1.5; ventral [V]: −4.3) and injected with 300 nL AAV-CAG-Flex-GCaMP6s in the DMS (N = 2 M 1F Air, 2 M 2F CIE). In the second, D1 Cre+ mice were injected with 300 nL AAV-CAG-Flex-GCaMP6s in the DMS (N = 3 M Air, 5 M CIE) . To localize to parvalbumin positive and the indirect pathway, PV-Cre+, and A2a-Cre+ mice received unilateral injections of 300 nL AAV-CAG-Flex-GCaMP6s (pAAV.CAG.Flex.GCaMP6s.WPRE.SV40 targeted to DMS, respectively. Post injection, an optic fiber was implanted immediately above the injection site (V: −3.15). After completion of behavioral procedures, mice were deeply anesthetized, brains extracted and placed into 4% Paraformaldehyde for 24–48 hrs and then transferred to 1X PBS. A subset of D1+ Cre and A2a-Cre+ mice were perfused with Calcium Chloride to allow for imaging of GCaMP6s expression in the absence of immunohistochemical procedures. Spread of GCaMP6s expression was not assessed in PV Cre+ mice due to the sparsity of PV Cre+ neurons in DMS. Optic fiber placement was qualified by examining tracts in 100-μm-thick brain slices under a macro fluorescence microscope (Olympus MVX10). Only mice with 1) signal detection and 2) fiber placement within DMS were included for analyses.

Chronic intermittent ethanol exposure and repeated withdrawal procedure

At least one week post surgical procedures, mice were exposed to ethanol vapor or air and repeated withdrawal procedures. All mice underwent four rounds, with each round consisting of 16 hr of vapor exposure followed by an 8 hr withdrawal, repeated for four consecutive days followed by 3 days of no vapor exposure. 190 proof ethanol (Koptec, Pennsylvania) was volatilized by bubbling air through a flask containing 95% ethanol at a rate of 2–3 L/min. The resulting ethanol vapor was combined with a separate air stream to give a total flow rate of approximately 10 L/min, which was delivered to the mice in their home cages in Plexiglas chambers (Plas Labs Inc). Blood ethanol concentrations (BEC) were collected at the end of each round from sentinel mice (mean BEC across the four rounds of vapor = 36.7 ± 3.7 mM). No pyrazole or loading ethanol injections were given prior to placement in vapor chambers.

Behavioral training procedures

To avoid effects of acute withdrawal, food restriction began three days after the last day of vapor procedures, to reduce body weight to ∼85–90% of baseline, with this restriction maintained throughout training and testing. Behavioral training procedures began five days post last vapor procedure. Mice were placed in sound attenuating operant boxes (Med-Associates) and were trained to press a single lever (left or right) for a food reinforcer (Purified Food Pellet, 20 mg, BioServ #F0071) delivered in a food receptacle. Under a Random Time (RT) schedule, mice were first trained to make a head entry to retrieve the outcome, which was delivered on average every 60 s (RT60) in the absence of levers. Next, mice were trained on a continuous reinforcement schedule (CRF) in which each lever press produced a single outcome. The maximum number of outcomes they could potentially earn across three daily CRF sessions were 5, 15, and 30 outcomes, respectively. Following CRF training, mice were then trained under a random ratio (RR) schedule of reinforcement that biases goal-directed control in mice. Mice received two days of training in RR10 (on average the 10th lever press produces the outcome), followed by 4–5 days under RR20. A subset of mice (A2a Cre+ n = 6, D1 Cre+ n = 5, PV Cre+ n = 10) stayed at RR10 training for the entire duration of RR training due to overall low levels of lever pressing. Sessions ended after 30 outcomes were earned or after 60 min had elapsed.

Fiber photometry procedures and analyses

Bifurcated optic fibers (400 uM, Thorlabs) (to allow simultaneous recording of two mice) were attached for the last 4–6 days of instrumental training, beginning on the first or second day of random ratio training. A blue LED (470 nm, Thorlabs) was used for excitation of dorsal striatum. Regions of interest for each optic fiber were selected by Bonsai software and fluorescence emissions were focused through a 10 × objective (Olympus) onto a CMOS camera (FLIR Systems). Fluorescence intensity and analog signals for lever press, head entries, and outcome delivery were acquired simultaneously, thresholded, and timestamped for later analyses using Bonsai software. After each session, Bonsai software saved photometry signals and behavioral timestamps within comma-separated value files (.csv) that were then imported into Matlab (Mathworks Inc., Natick, MA) for subsequent analysis using custom scripts. Raw fluorescence intensity signals underwent running median (5th order) and low pass (high cutoff frequency of 1 Hz) filtering to reduce noise and electrical artifacts. To correct for photobleaching in which a signal captured from fluorophores degrades by continuous light exposure during the session, we high pass filtered the signal with a low cutoff frequency of 0.001 Hz. We then performed a quality check on the filtered fluorescence intensity signal for low expression and fiber decoupling. Briefly, sessions that did not exceed a 15 second moving window calculation of the signal's 97.5 percentile by a minimum 1% fluorescence change or did not pass a visual inspection for within-session fiber-ferrule decoupling artifacts were excluded from further analyses.

Statistical analyses

All analyses were two-tailed and statistical significance was defined as an α of p < 0.05. Statistical analysis was performed using GraphPad Prism 8.3.0 (GraphPad Software). Behavioral training data, including lever presses, response rate, and rewards earned were analyzed using two-way repeated measures ANOVAs with Treatment Group (Air vs. CIE) and Day as factors.

For each trace of fiber photometry data, peri‑event changes in fluorescence intensity were calculated via z-score normalization of each corresponding behavioral epoch (i.e. −2 s to 5 s around the lever press at time 0) to a preceding baseline period (ie: for lever press from −5 to −2 prior to lever press). For behavioral epochs, we included analyses centered around each lever press, the first lever press of a bout of lever pressing, as well as the first head entry made following reward delivery. The latter includes a reward-related period that may include perception, consumption, and evaluation. We then analyzed this data two ways. First, z-scored fluorescence traces were combined across all mice within a group to preserve the variance seen within a subject. Comparisons between Air and CIE mice were made running permutation tests (1000 shuffles) that required at least 5 consecutive samples to be different from one another. Second, we analyzed the data per animal in a single session to account for the potential differences between subjects. Comparisons between Air and CIE mice were made using permutation testing. Population Ca²⁺ activity traces were then smoothed with MATLAB's Savitzky–Golay smoothdata method using a 20 sample sliding window for visual display purposes only.

Results

Chronic ethanol effects on direct pathway calcium activity during actions

Post surgical and vapor procedures (Fig. 1A-C), Air (final n = 6, 5 M 1F) and CIE (final n = 9, 7 M 2F) mice expressing GCaMP6s in the direct pathway similarly acquired lever-press related behaviors. Two-way ANOVAs performed on lever press (Fig. 1D), lever press rate (Fig. 1E), and rewards earned (Fig. 1F) all showed a main effect of Day (Fs’ > 6.23, ps’ < 0.01), but no main effect of Treatment Group nor interaction. Thus, as previously reported, prior chronic alcohol exposure did not lead to gross alterations in acquisition of lever-press related behaviors.

Fig. 1.

_{Chronic alcohol increases calcium activity of the direct pathway. A. Schematic of experimental timeline. B. Surgical targeting of GCaMP6s to the direct pathway and implantation of optic fiber, and C. example of optic fiber placement within DMS in animal perfused with 10 mM CaCl2 to show GCaMP6s expression. Lever presses (D), Lever press rate (E), and rewards earned (F) during training. For D-F, plotted are means and error bars representing ± SEM. Average (SEM) calcium activity traces by trial from Air and CIE mice in relation to lever pressing (G), onset of lever press bouts (H), and first head entry following reward delivery (I). Average (SEM) calcium activity traces by per mouse session averages during lever presses (J), onset of lever press bouts (K), and first head entry following reward delivery (L). Differences between Treatment groups was evaluated using permutation testing, with dashed line indicating p < 0.05.}

We next examined calcium activity of direct pathway populations (Fig. 1G-L). While direct pathway SPNs showed modulation around lever-press related behaviors, we found largely increased calcium activity modulation in CIE mice compared to Air mice. Permutation testing revealed significantly greater calcium modulation from baseline modulation around a lever press (Fig. 1G) (Air trial n = 9288, CIE trial n = 8179) (ps’ < 0.05). However, grouping lever presses by bout and examining calcium activity centered around the first lever press in a bout showed a reduced increase in calcium modulation in CIE mice compared to Air mice prior to the onset of a lever press bout (Fig. 1H) (Air trial n = 921, CIE trial n = 882) (ps’ < 0.05). Increased calcium activity modulation was also present during the first head entry following reward delivery, with permutation testing showing a large significant difference in calcium activity recruited between Air and CIE mice during this reward-related epoch (Fig. 1I) (Air trial n = 711, CIE trial n = 727) (ps’ < 0.05). This pattern of findings was also seen when we analyzed activity using per mouse averages, except there were no differences at the onset of a lever-press bout (Fig.J-L). Thus, prior CIE exposure largely led to a long-lasting increase in lever-press-related calcium activity modulation in the direct pathway during decision-making.

Chronic ethanol effects on indirect pathway calcium activity during actions

Air (final n = 4, 1 M 3F) and CIE (final n = 6, 2 M 4F) mice expressing GCaMP6s targeted to the indirect pathway (Fig. 2A-C) also acquired lever press behavior to similar degrees. Once again, there was only a main effect of Day across when examining lever presses (Fig. 2D), lever press rate (Fig. 2E), and rewards earned (Fig. 2F) (Fs’ > 6.61, ps’ < 0.05), with no main effects of Treatment Group or interaction.

Fig. 2.

Chronic alcohol decreases calcium activity of the indirect pathway. A. Schematic of experimental timeline. B. Surgical targeting of GCaMP6s to the indirect pathway and implantation of optic fiber. C. Example of optic fiber placement within DMS in animal perfused with 10 mM CaCl₂ to show GCaMP6s expression. Lever presses (D), Lever press rate (E), and rewards earned (F) during training. For D-F, plotted are means and error bars representing ± SEM. Average (SEM) calcium activity traces by trial from Air and CIE mice in relation to lever pressing (G), onset of lever press bouts (H), and first head entry following reward delivery (I). Average (SEM) calcium activity traces by per mouse session averages during lever presses (J), onset of lever press bouts (K), and first head entry following reward delivery (L). Differences between Treatment groups was evaluated using permutation testing, with dashed line indicating p < 0.05.

Unlike calcium activity in the direct pathway, calcium activity modulation in the indirect pathway was largely reduced in CIE mice when compared to Air mice. While indirect pathway calcium activity increased prior to lever pressing, permutation testing showed that indirect pathway calcium activity modulation was largely reduced in CIE compared to Air mice around a lever press (Fig. 2G) (Air trial n = 3802, CIE trial n = 4543) as well as at the onset of a lever press bout (Fig. 2H) (Air trial n = 396, CIE trial n = 551), (ps’ < 0.05). While of a lesser magnitude, reward-related indirect pathway calcium activity modulation was overall slightly greater in Air compared to CIE mice (Fig. 2I) (Air trial n = 283, CIE trial n = 347) (ps’ < 0.05). A similar pattern was seen when data was analyzed using session averages per mouse, except no significant difference in modulation were observed during the first head entry after reward (Fig.J-L). Thus, prior chronic alcohol exposure largely decreased calcium activity in the indirect pathway population.

Chronic ethanol effects on striatal interneuron calcium activity during actions

We also examined whether chronic alcohol altered calcium activity modulation in a population of striatal interneurons, the fast-spiking PV interneurons (Fig. 3A-C). Once again, acquisition behavior was similar between Air (final n = 5, 2F 3 M) and CIE (final n = 7, 3F 4 M) exposed mice. Across lever press rate (Fig. 3E) and rewards earned (Fig. 3F), there was a main effect of Day (Fs’ > 2.65, ps’ < 0.05), but no effect of Treatment Group or interaction and no differences in lever presses made (Fig. 3D).

Fig. 3.

Chronic alcohol decreases calcium activity of local interneurons. A. Schematic of experimental timeline. B. Surgical targeting of GCaMP6s to PV interneurons and implantation of optic fiber. C. Fiber placement in DMS. Lever presses (D), Lever press rate (E), and rewards earned (F) during training. For D-F, plotted are means and error bars representing ± SEM. Average (SEM) calcium activity traces by trial from Air and CIE mice in relation to lever pressing (G), onset of lever press bouts (H), and first head entry following reward delivery (I). Average (SEM) calcium activity traces per mouse session average during lever presses (J), onset of lever press bouts (K), and first head entry following reward delivery (L). Differences between Treatment groups was evaluated using permutation testing, with dashed line indicating p < 0.05.

Calcium activity of PV Cre+ interneurons showed reduced modulation in CIE compared to AIR mice during lever pressing (Fig. 3G) (Air trial n = 3446, CIE trial n = 7024), and at the onset of lever press bouts (Fig. 3H) (Air trial n = 426, CIE trial n = 700). However, compared to Air mice, CIE mice showed increased calcium activity modulation of PV Cre+ interneurons during reward-related epochs (Fig. 3I) (Air trial n = 356, CIE trial n = 616) (ps’ <0.05). Analyses on session averages per mouse also found the same pattern of calcium modulation (Fig. J-L). This suggests that prior chronic alcohol exposure led to reduced action-related recruitment of calcium activity in PV interneurons, but increased calcium modulation during reward-related epochs in DMS during action control.

Discussion

Here we show that prior chronic alcohol exposure leads to a long-lasting imbalance in the activity of DMS populations during action-related behaviors. This imbalance was reflected in increased modulation of calcium activity in direct pathway SPNs, while largely decreased calcium activity modulation was observed in indirect pathway SPNs as well as in PV DMS interneurons. Such changes are in line with prior findings suggesting that chronic alcohol results in potentiated recruitment of direct pathway SPNs and increased inhibition of indirect pathway neurons. Here, we show for the first time that such chronic-alcohol induced changes are reflected in the endogenous calcium activity of DMS pathways during action-related behaviors. These findings provide evidence that chronic alcohol induces an imbalance in DMS activity, and such an imbalance may lead to less functional control by DMS over behavior.

While intact DMS function has been widely implicated as necessary for goal-directed control, the proper balance of activity between direct and indirect pathways has recently been more appreciated. Indeed, both direct and indirect pathways are engaged during lever pressing and can support optogenetic-intracranial self-stimulation. Effective control of behavior by DMS is thought to engage cortical and thalamic recruitment of balanced activity of direct and indirect pathways. Recent examples show that chronic alcohol-induced changes in strength of select cortical inputs can lead to loss of goal-directed control. Interestingly, this was a reduction of orbitofrontal cortex transmission onto the direct, but not indirect, pathway in DMS. Thus, one interpretation has been reduced cortical drive (albeit from one area) into DMS leads to a loss of DMS recruitment and goal-directed control.

However, other works examining other prefrontal cortex projections have shown potentiated inputs onto D1 SPNs of the direct pathway in DMS, as well as decreased thalamic input to striatum. In the present findings, we saw increased calcium activity modulation in D1 SPNs, but decreased modulation in D2 SPNs and a local interneuron population. Intriguingly, we also saw reduced modulation of PV interneurons in the DMS. Recent work ablated this population in the dorsal lateral striatum and found reduced organized ethanol consumption with ethanol consumption showing increased sensitivity to the addition of quinine. As interneurons can provide a strong clamp on SPN firing and are recruited in part through cortical input, our data suggest less modulation of PV interneurons may reflect less engagement of this population to shape SPN activity patterns during actions. These findings along with prior works showing alcohol-induced changes to SPNs themselves suggests that at minimum, dysfunctional or a loss of DMS control over behavior could result from a gain or loss in strength of specific cortical or thalamic inputs conveying relevant information, and/or through functional changes in the DMS neurons themselves. Further work examining how chronic alcohol disrupts the ability of converging inputs to recruit appropriate DMS activity is clearly warranted.

Care should be taken in interpreting the present results as bulk population calcium in striatal SPNs is likely not reflective of action potential firing. Given the extent of SPN arborization, the bulk signals collected in the present data likely reflect a combination of threshold and subthreshold calcium activity in the soma, dendritic arbors, and axon collaterals within striatum. We did not examine sensitivity to outcome devaluation in the present studies and therefore cannot say whether the current subjects were more reliant on habitual or goal-directed control, although prior works have shown chronic alcohol-exposure disrupts goal-directed control for non-alcohol reward. However, our data do suggest that chronic alcohol resulted in altered calcium modulation in identified SPN populations during action performance. As prior findings showed DMS inactivation was ineffective at altering behavioral control following chronic alcohol, our data suggest that alteration to activity of DMS populations may underlie deficits in using DMS for behavioral control. We further hypothesize that this imbalance in recruitment likely contributes to the reduced contribution of DMS to goal-directed control following chronic alcohol.

Introduction

Alcohol Use Disorder (AUD) is characterized by changes in executive function and ongoing, uncontrolled alcohol consumption. These changes are believed to stem, in part, from the long-term effects of alcohol on brain circuits in the cortex and striatum. The dorsal medial striatum (DMS) is crucial for goal-directed actions and is known to be affected by chronic alcohol exposure. Previous studies indicate that chronic alcohol alters DMS activity, potentially leading to a loss of behavioral control. However, the specific ways in which chronic alcohol affects activity in different DMS neuron populations during action control remain poorly understood.

The DMS is a complex area containing two main output pathways that use GABA, known as the direct and indirect pathways. The direct pathway consists of GABAergic spiny neurons (SPNs) that connect directly to the Substantia Nigra Pars Reticulata and express dopamine type-1 receptors (D1-SPNs). In contrast, the indirect pathway's GABAergic SPNs connect directly to the Globus Pallidus and express dopamine type-2 receptors (D2-SPNs). Various interneuron populations, such as fast-spiking parvalbumin (PV) interneurons, regulate these pathways by strongly inhibiting SPN activity. The striatum receives primary input from the cortex and thalamus via glutamatergic signals. The DMS specifically processes input from the associative cortex and thalamus, allowing it to integrate information vital for goal-directed control. Research confirms that an intact DMS is essential for goal-directed behavior across many species.

Chronic alcohol consumption is known to impair goal-directed control, affecting both general reward-seeking and the ability to control alcohol self-administration. These impairments are linked to specific alterations in corticostriatal transmission and broader changes in transmission within the DMS. Typically, chronic alcohol is associated with reduced GABAergic transmission and increased glutamatergic transmission in the DMS, though this can vary depending on the specific neural connections studied. Relevant to this investigation, previous research on chronic alcohol drinking has reported strengthened glutamatergic transmission to the direct pathway and increased GABAergic transmission to the indirect pathway. Replicating or reversing these changes in transmission in living organisms either increased or decreased alcohol consumption. Additionally, both short-term and long-term alcohol exposure have been shown to impact transmission involving PV interneurons, which contribute to compulsive alcohol consumption in the striatum. The theory that alcohol self-administration gradually becomes less dependent on DMS function suggests that chronic alcohol may cause an imbalance in DMS activity, leading to impaired goal-directed control.

This study investigates the action-related calcium activity in the direct and indirect pathways of the DMS, as well as in local PV interneurons, after chronic alcohol exposure. A reinforcement schedule was used in mice to promote goal-directed action control, as previous alcohol exposure has been shown to disrupt this type of control under similar conditions. Bulk calcium fiber photometry was employed to estimate neural activity from genetically or projection-defined populations. It is important to acknowledge that in the striatum, bulk calcium activity does not have a direct linear relationship with the number of action potentials. Nevertheless, these measurements offer insights into both dendritic and somatic calcium activity, which likely reflect incoming neural transmission and processes related to brain plasticity. Consistent with earlier hypotheses, the findings indicate that chronic alcohol exposure led to lasting increases in direct pathway calcium activity during decision-making, but decreased calcium activity in the indirect pathway and PV fast-spiking interneurons. These results support the idea that chronic alcohol causes an imbalance in DMS population activity, thereby reducing the DMS's contribution to behavioral control.

Methods and materials

Subjects

The study utilized C57Bl/6J, Pvalbtm1(cre)Arbr (PV Cre+), B6.FVB(Cg)-Tg(Drd1-cre)EY266Gsat/Mmucd (D1-Cre+), and B6.FVB(Cg)-Tg(Adora2a-cre)KG139Gsat/Mmucd (A2a-Cre) mice. Adult male and female mice, over eight weeks old, were housed in groups of two to four, with unlimited access to mouse chow and water. A 14-hour light/10-hour dark cycle was maintained. Although both sexes were included, the study was not designed to detect sex differences, the influence of hormones in males or females, or the effect of genotype on calcium modulation. All experiments took place during the light phase of the light cycle. Procedures were approved by the University of California San Diego Institutional Animal Care and Use Committee and adhered to the National Institutes of Health (NIH) “Principles of Laboratory Care.” Researchers were aware of the experimental groups.

Surgical procedures and histological verification

Mice underwent anesthesia with isoflurane and received stereotaxic injections into the DMS. Coordinates for injection were +0.5 mm anterior, ±1.5 mm medial, and -3.25 mm ventral from Bregma, using 500 nL Hamilton syringes. To target GCaMP6s expression to direct pathway neurons, two methods were employed: C57Bl/6J mice received 300 nL AAV-Ef1a-Cre-WPRE injections into the SNr (coordinates: -3.4 mm anterior, ±1.5 mm medial, -4.3 mm ventral) followed by 300 nL AAV-CAG-Flex-GCaMP6s in the DMS. Alternatively, D1 Cre+ mice received 300 nL AAV-CAG-Flex-GCaMP6s injections directly into the DMS. For parvalbumin-positive neurons and the indirect pathway, PV-Cre+ and A2a-Cre+ mice received unilateral 300 nL injections of AAV-CAG-Flex-GCaMP6s (pAAV.CAG.Flex.GCaMP6s.WPRE.SV40) into the DMS. An optic fiber was implanted directly above the injection site (-3.15 mm ventral). Following behavioral procedures, brains were extracted after deep anesthesia, fixed in 4% paraformaldehyde for 24-48 hours, and then transferred to 1X PBS. Some D1+ Cre and A2a-Cre+ mice were perfused with calcium chloride to visualize GCaMP6s expression without immunohistochemistry. The spread of GCaMP6s expression was not evaluated in PV Cre+ mice due to the low density of PV Cre+ neurons in the DMS. Optic fiber placement was verified by examining tracts in 100-µm-thick brain slices using a macro fluorescence microscope. Only data from mice with detectable signals and correct fiber placement within the DMS were included in the analysis.

Chronic intermittent ethanol exposure and repeated withdrawal procedure

At least one week after surgery, mice were exposed to either ethanol vapor or air, followed by repeated withdrawal periods. The procedure involved four rounds, each comprising 16 hours of vapor exposure followed by an 8-hour withdrawal. This was repeated for four consecutive days, then followed by three days without vapor exposure. Ethanol (190 proof) was vaporized by bubbling air through a flask containing 95% ethanol at a rate of 2-3 L/min. This vapor was then combined with another air stream to achieve a total flow rate of approximately 10 L/min, delivered to the mice in Plexiglas home cages. Blood ethanol concentrations (BEC) were measured from sentinel mice at the end of each round, with an average BEC of 36.7 ± 3.7 mM across the four rounds. No pyrazole or initial ethanol injections were administered before placing mice in the vapor chambers.

Behavioral training procedures

To prevent the effects of acute withdrawal, food restriction began three days after the final vapor procedure, aiming to reduce body weight to approximately 85-90% of baseline. This restriction was maintained throughout all training and testing. Behavioral training commenced five days after the last vapor exposure. Mice were placed in sound-attenuating operant boxes and trained to press a single lever (left or right) to receive a 20 mg food pellet delivered into a food receptacle. Initially, mice were trained on a Random Time (RT) schedule, where they learned to make a head entry to retrieve an outcome delivered, on average, every 60 seconds (RT60) without levers present. Subsequently, mice were trained on a continuous reinforcement (CRF) schedule, where each lever press yielded one outcome. The maximum number of outcomes obtainable in three daily CRF sessions was 5, 15, and 30, respectively. After CRF training, mice transitioned to a random ratio (RR) reinforcement schedule, which promotes goal-directed control. Mice underwent two days of RR10 training (where, on average, every tenth lever press produced an outcome), followed by 4-5 days of RR20 training. A subgroup of mice remained at RR10 training for the entire RR duration due to consistently low lever-pressing rates. Sessions concluded after 30 outcomes were earned or after 60 minutes had passed.

Fiber photometry procedures and analyses

Bifurcated optic fibers (400 µM) were connected during the last 4-6 days of instrumental training, starting on the first or second day of random ratio training, enabling simultaneous recording from two mice. A blue LED (470 nm) excited the dorsal striatum. Bonsai software selected regions of interest for each optic fiber, and fluorescence emissions were focused through a 10x objective onto a CMOS camera. Fluorescence intensity and analog signals for lever presses, head entries, and outcome delivery were concurrently acquired, thresholded, and timestamped using Bonsai for later analysis. After each session, Bonsai saved photometry signals and behavioral timestamps in comma-separated value files, which were then imported into Matlab for analysis using custom scripts. Raw fluorescence intensity signals were subjected to a 5th-order running median and a low-pass filter (high cutoff frequency of 1 Hz) to reduce noise and electrical artifacts. To correct for photobleaching, where fluorophore signals degrade with continuous light exposure, the signal was high-pass filtered with a low cutoff frequency of 0.001 Hz. A quality check was performed on the filtered fluorescence intensity signal to identify low expression or fiber decoupling. Sessions were excluded from further analysis if a 15-second moving window calculation of the signal's 97.5 percentile did not show at least a 1% fluorescence change, or if visual inspection revealed fiber-ferrule decoupling artifacts within the session.

Statistical analyses

All statistical analyses were two-tailed, with significance set at p < 0.05. GraphPad Prism 8.3.0 was used for statistical computations. Behavioral training data, including lever presses, response rate, and rewards earned, were analyzed using two-way repeated measures ANOVAs, with Treatment Group (Air vs. CIE) and Day as factors. For fiber photometry data, peri-event changes in fluorescence intensity were calculated by z-score normalizing each behavioral epoch (e.g., -2 seconds to 5 seconds around a lever press at time 0) to a preceding baseline period (e.g., -5 to -2 seconds prior to the lever press). Behavioral epochs analyzed included those centered around each lever press, the first lever press of a sequence, and the initial head entry after reward delivery. The latter epoch encompassed reward perception, consumption, and evaluation. Data was analyzed using two approaches. First, z-scored fluorescence traces were combined across all mice within a group to retain subject variability. Permutation tests (1000 shuffles) were used for comparisons between Air and CIE mice, requiring at least five consecutive samples to differ significantly. Second, data was analyzed per animal per session to account for inter-subject variability, with permutation testing used for comparisons between Air and CIE mice. For visualization, population Ca2+ activity traces were smoothed using MATLAB’s Savitzky–Golay smoothdata method with a 20-sample sliding window.

Results

Chronic ethanol effects on direct pathway calcium activity during actions

After surgical and vapor procedures, mice exposed to air (n=6; 5 males, 1 female) and those exposed to chronic intermittent ethanol (CIE) (n=9; 7 males, 2 females) showed similar acquisition of lever-press behaviors. Both groups expressed GCaMP6s in the direct pathway. Two-way ANOVAs on lever presses, lever press rate, and rewards earned revealed a main effect of Day, indicating that performance changed over time. However, there was no main effect of Treatment Group or interaction, suggesting that prior chronic alcohol exposure did not cause major changes in the initial learning of lever-pressing behaviors.

Analysis then focused on the calcium activity of direct pathway populations. Direct pathway SPNs exhibited changes in activity related to lever-pressing. Notably, CIE-exposed mice showed generally increased calcium activity modulation compared to air-exposed mice. Permutation testing indicated significantly greater calcium modulation from baseline around individual lever presses in CIE mice. However, when lever presses were grouped by bout, and activity was examined around the first press of a bout, CIE mice displayed a reduced increase in calcium modulation compared to air-exposed mice before the start of a lever press bout. Elevated calcium activity modulation was also observed during the first head entry after reward delivery, with permutation testing showing a substantial difference in recruited calcium activity between the groups during this reward-related period. This overall pattern was consistent when analyzing activity using per-mouse averages, though no differences were observed at the onset of a lever-press bout in this analysis. Therefore, previous CIE exposure resulted in a persistent increase in lever-press-related calcium activity modulation within the direct pathway during decision-making.

Chronic ethanol effects on indirect pathway calcium activity during actions

Air-exposed mice (n=4; 1 male, 3 females) and CIE-exposed mice (n=6; 2 males, 4 females), with GCaMP6s targeted to the indirect pathway, acquired lever-press behavior to similar extents. Similar to the direct pathway findings, analysis of lever presses, lever press rate, and rewards earned showed only a main effect of Day, indicating changes in performance over time. No main effects of Treatment Group or interactions were observed.

In contrast to the direct pathway, calcium activity modulation in the indirect pathway was generally reduced in CIE-exposed mice compared to air-exposed mice. Although indirect pathway calcium activity increased before lever pressing, permutation testing revealed a significant reduction in indirect pathway calcium activity modulation in CIE mice around individual lever presses and at the onset of a lever press bout. Reward-related indirect pathway calcium activity modulation was also slightly greater in air-exposed mice compared to CIE mice, albeit to a lesser extent. A similar pattern was observed when analyzing session averages per mouse, except that no significant modulation differences were noted during the first head entry after reward. Therefore, previous chronic alcohol exposure primarily decreased calcium activity in the indirect pathway population.

Chronic ethanol effects on striatal interneuron calcium activity during actions

The study also investigated whether chronic alcohol altered calcium activity modulation in fast-spiking PV interneurons, a population of striatal interneurons. Acquisition behavior remained similar between air-exposed mice (n=5; 2 females, 3 males) and CIE-exposed mice (n=7; 3 females, 4 males). Analysis of lever press rate and rewards earned showed a main effect of Day, but no effect of Treatment Group or interaction, and no differences in the total number of lever presses.

Calcium activity of PV Cre+ interneurons exhibited reduced modulation in CIE-exposed mice compared to air-exposed mice during individual lever presses and at the onset of lever press bouts. In contrast, CIE-exposed mice showed increased calcium activity modulation of PV Cre+ interneurons during reward-related periods when compared to air-exposed mice. Analyses based on session averages per mouse yielded the same pattern of calcium modulation. These findings suggest that prior chronic alcohol exposure resulted in decreased action-related recruitment of calcium activity in PV interneurons, but increased calcium modulation during reward-related periods within the DMS during action control.

Discussion

This study demonstrates that prior chronic alcohol exposure causes a lasting imbalance in the activity of DMS populations during behaviors involving action. This imbalance manifested as increased calcium activity modulation in direct pathway SPNs, alongside a general decrease in calcium activity modulation in indirect pathway SPNs and PV DMS interneurons. These alterations align with previous research indicating that chronic alcohol leads to enhanced recruitment of direct pathway SPNs and increased inhibition of indirect pathway neurons. The current findings are the first to show these alcohol-induced changes reflected in the intrinsic calcium activity of DMS pathways during action-related behaviors. The results provide evidence that chronic alcohol creates an imbalance in DMS activity, which may impair the DMS's functional control over behavior.

While the intact function of the DMS is widely recognized as essential for goal-directed control, the importance of a proper balance between direct and indirect pathway activity has recently gained more recognition. Both direct and indirect pathways are active during lever pressing and can support optogenetic-intracranial self-stimulation. The DMS is believed to effectively control behavior by recruiting balanced activity in both direct and indirect pathways through cortical and thalamic inputs. Recent studies indicate that chronic alcohol-induced changes in the strength of specific cortical inputs can result in impaired goal-directed control. Notably, this included a reduction in orbitofrontal cortex transmission to the direct pathway, but not the indirect pathway, within the DMS. One interpretation suggests that reduced cortical input (even from a single area) into the DMS contributes to a decrease in DMS engagement and a loss of goal-directed control.

Other research, however, examining different prefrontal cortex projections, has reported strengthened inputs onto D1 SPNs of the direct pathway in the DMS, alongside reduced thalamic input to the striatum. In the current study, increased calcium activity modulation was observed in D1 SPNs, while decreased modulation was seen in D2 SPNs and a local interneuron population. Specifically, reduced modulation of PV interneurons in the DMS was an intriguing finding. Previous research found that ablating this PV interneuron population in the dorsal lateral striatum led to less organized ethanol consumption, with increased sensitivity to quinine. Since interneurons can powerfully inhibit SPN firing and are partly recruited by cortical input, the current data suggest that less modulation of PV interneurons might indicate reduced engagement of this population in shaping SPN activity patterns during actions. These findings, combined with earlier work showing alcohol-induced changes to SPNs themselves, propose that impaired or lost DMS control over behavior could result from either an increase or decrease in the strength of specific cortical or thalamic inputs carrying relevant information, or from functional changes within the DMS neurons themselves. Further research is needed to understand how chronic alcohol disrupts the ability of converging inputs to properly engage DMS activity.

Caution is advised when interpreting these results, as bulk population calcium signals in striatal SPNs may not directly reflect action potential firing. Due to the extensive branching of SPNs, the bulk signals collected likely represent a combination of threshold and subthreshold calcium activity within the soma, dendritic arbors, and axon collaterals in the striatum. The study did not assess sensitivity to outcome devaluation, precluding conclusions about whether the subjects relied more on habitual or goal-directed control. However, previous studies have shown that chronic alcohol exposure impairs goal-directed control for non-alcohol rewards. Despite these limitations, the data indicate that chronic alcohol led to altered calcium modulation in specific SPN populations during action performance. Given prior findings that DMS inactivation did not alter behavioral control after chronic alcohol exposure, the current data suggest that changes in DMS population activity may underpin difficulties in utilizing the DMS for behavioral control. It is hypothesized that this imbalance in recruitment likely contributes to the diminished role of the DMS in goal-directed control following chronic alcohol.

Introduction

Alcohol Use Disorder (AUD) is often characterized by changes in how the brain makes decisions and by continued, uncontrolled alcohol consumption. These changes are believed to stem from the long-term effects of alcohol on specific brain circuits, especially those in the cortex and striatum. For example, the dorsal medial striatum (DMS) plays a key role in actions aimed at achieving a specific goal. Research has shown that chronic alcohol exposure can alter the activity within the DMS, potentially leading to a loss of behavioral control. However, it remains unclear how chronic alcohol specifically changes the activity of different types of neurons in the DMS during goal-directed actions.

The DMS is a complex brain area containing various types of neurons. It primarily consists of two main output pathways that use the neurotransmitter GABA: the direct and indirect pathways. Neurons in the direct pathway (D1-SPNs) can be identified by their connection to the Substantia Nigra Pars Reticulata and their expression of the dopamine type-1 receptor. In contrast, neurons in the indirect pathway (D2-SPNs) connect to the Globus Pallidus and express the dopamine type-2 receptor. Both pathways are modulated by interneurons, such as parvalbumin (PV) interneurons, which strongly inhibit the activity of the main striatal neurons. The striatum, including the DMS, receives signals primarily from the cortex and thalamus. Because the DMS receives input from brain regions involved in association, it is well-positioned to process and adjust information crucial for goal-directed control. Studies across various species confirm that a healthy DMS is essential for these behaviors.

Chronic alcohol exposure has been shown to impair goal-directed control, affecting both general reward-seeking behaviors and the ability to control alcohol intake. These disruptions are linked to specific changes in the communication between the cortex and striatum, as well as general alterations in transmission within the DMS itself. Chronic alcohol generally reduces GABAergic transmission and increases glutamatergic transmission in the DMS, though this can vary depending on the specific brain connections examined. Previous research relevant to this study found that chronic alcohol drinking enhanced glutamatergic transmission to the direct pathway and increased GABAergic transmission to the indirect pathway. Reversing or mimicking these transmission changes in living animals was shown to decrease or increase alcohol consumption, respectively. Both acute and chronic alcohol have also been observed to affect communication involving PV interneurons, and striatal PV interneurons are known to support compulsive alcohol consumption. Since alcohol self-administration is thought to initially involve striatal function that later becomes less dependent on the DMS, it suggests that chronic alcohol may create an imbalance in DMS activity, leading to dysfunctional goal-directed control.

This study investigated the calcium activity related to actions in DMS direct and indirect pathways, as well as in local PV interneurons, following chronic alcohol exposure. A reinforcement schedule was used that encourages goal-directed action control in mice, as prior alcohol exposure has been shown to disrupt this type of control under similar conditions. While bulk calcium fiber photometry was used to measure activity in genetically or projection-defined neuron populations, it is important to note that in the striatum, bulk calcium activity does not directly reflect the number of action potentials fired. However, these bulk calcium measurements do provide insights into calcium activity in both dendrites and cell bodies, which likely reflects incoming signals and is linked to processes of brain plasticity. Consistent with previous theories, the study found that prior chronic alcohol exposure led to long-lasting increases in the calcium activity of the direct pathway during decision-making. Conversely, it reduced the calcium activity of the indirect pathway and PV fast-spiking interneurons. These findings support the idea that chronic alcohol causes an imbalance in DMS neuron population activity, which may result in the DMS contributing less effectively to behavioral control.

Methods and materials

Study Participants and Procedures

Adult C57Bl/6 J, Pvalbtm1(cre)Arbr (PV Cre+), B6.FVB(Cg)-Tg(Drd1-cre)EY266Gsat/Mmucd (D1-Cre+), and B6.FVB(Cg)-Tg(Adora2a-cre)KG139Gsat/Mmucd (A2a-Cre) mice were used in these experiments. Both male and female mice, over eight weeks old, were housed in groups and provided with food and water freely, following a 14-hour light/10-hour dark cycle. Although both sexes were included, the study was not designed to identify sex differences, the influence of hormones, or the effect of genetic variations on calcium modulation. All experiments were conducted during the light phase. All procedures were approved by the University of California San Diego Institutional Animal Care and Use Committee and followed the National Institutes of Health (NIH) guidelines for laboratory animal care. Researchers were aware of the experimental groups.

All mice underwent surgical procedures involving anesthesia and stereotaxic injections into the DMS. To specifically observe GCaMP6s expression in direct pathway neurons, two methods were used: injecting a Cre-expressing virus into the Substantia Nigra Pars Reticulata of C57Bl/6 J mice, along with a Flex-GCaMP6s virus into the DMS; or injecting Flex-GCaMP6s directly into the DMS of D1 Cre+ mice. For targeting parvalbumin-positive and indirect pathway neurons, PV-Cre+ and A2a-Cre+ mice, respectively, received unilateral injections of Flex-GCaMP6s into the DMS. Following injections, an optic fiber was implanted just above the injection site. After behavioral studies, brains were collected and preserved to verify GCaMP6s expression and optic fiber placement. Only data from mice with detectable signal and correct fiber placement within the DMS were included in the analysis.

Chronic Alcohol Exposure

At least one week after surgery, mice were exposed to either ethanol vapor or air, followed by repeated withdrawal periods. The procedure involved four rounds, each consisting of 16 hours of vapor exposure followed by an 8-hour withdrawal period, repeated for four consecutive days, with three days off vapor exposure afterwards. Ethanol vapor was generated by bubbling air through alcohol and delivered to mice in their home cages within Plexiglas chambers. Blood ethanol concentrations were measured from control mice at the end of each round, averaging approximately 36.7 mM. No loading ethanol injections were given.

Behavioral Training

To avoid immediate withdrawal effects, food restriction began three days after the last vapor exposure, aiming to reduce body weight to about 85–90% of baseline, and was maintained throughout training and testing. Behavioral training started five days after the last vapor procedure. Mice were placed in operant boxes and trained to press a single lever for a food pellet. Initially, mice learned to make a head entry to retrieve food delivered randomly every 60 seconds without levers present. Next, they were trained on a continuous reinforcement (CRF) schedule, where each lever press produced one food pellet, with increasing daily outcome limits. Following CRF, mice were trained under a random ratio (RR) schedule (RR10 then RR20), which encourages goal-directed control. Some mice with low lever pressing rates remained at RR10. Sessions ended after 30 food pellets were earned or after 60 minutes.

Data Collection and Analysis

Bifurcated optic fibers were attached to mice for the final 4–6 days of instrumental training, starting on the first or second day of random ratio training. A blue LED was used to excite the dorsal striatum, and fluorescence emissions were captured by a CMOS camera. Fluorescence intensity and behavioral signals (lever presses, head entries, reward delivery) were recorded simultaneously. Raw fluorescence signals were filtered to reduce noise and corrected for photobleaching. A quality check was performed to exclude sessions with low signal or fiber detachment. Data was then imported into Matlab for analysis using custom scripts.

For each fiber photometry trace, changes in fluorescence intensity were calculated by normalizing to a preceding baseline period around each behavioral event (e.g., lever press, bout onset, reward delivery). This data was analyzed in two ways: first, z-scored fluorescence traces were combined across all mice within a group to capture within-subject variance; second, data was analyzed per animal for each session to account for individual differences. Permutation tests were used to compare Air and CIE groups, with statistical significance defined as a p-value less than 0.05. Calcium activity traces were smoothed for visual presentation only.

Results

Chronic ethanol effects on direct pathway calcium activity during actions

Following surgical and vapor procedures, both Air and CIE mice with GCaMP6s expression in the direct pathway learned lever-pressing behaviors similarly. Statistical analyses of lever presses, lever press rate, and rewards earned showed that performance improved over days, but there were no significant differences between the Air and CIE groups. This indicates that prior chronic alcohol exposure did not substantially alter the learning of lever-press behaviors.

When examining the calcium activity of direct pathway neurons, increased activity modulation was largely observed in CIE mice compared to Air mice during lever-press related behaviors. Permutation testing revealed significantly greater calcium modulation around a lever press in CIE mice. However, when looking at calcium activity centered around the first lever press in a bout, CIE mice showed a reduced increase in calcium modulation compared to Air mice before the bout began. Increased calcium activity modulation was also present during the first head entry after reward delivery, with a significant difference in recruited calcium activity between Air and CIE mice during this reward-related period. A similar pattern was observed when data was analyzed using averages per mouse session, except for the lack of difference at the onset of a lever-press bout. Therefore, prior chronic alcohol exposure primarily resulted in a long-lasting increase in lever-press-related calcium activity modulation in the direct pathway during decision-making.

Chronic ethanol effects on indirect pathway calcium activity during actions

Air and CIE mice expressing GCaMP6s targeted to the indirect pathway also acquired lever press behavior to similar extents. As with the direct pathway group, performance improved over days for lever presses, lever press rate, and rewards earned, but there were no significant differences between treatment groups.

In contrast to the direct pathway, calcium activity modulation in the indirect pathway was largely reduced in CIE mice compared to Air mice. While indirect pathway calcium activity did increase before lever pressing, permutation testing showed that indirect pathway calcium activity modulation was significantly reduced in CIE mice around a lever press and at the onset of a lever press bout. Although to a lesser degree, reward-related indirect pathway calcium activity modulation was generally slightly higher in Air mice compared to CIE mice. A similar pattern was observed when data was analyzed using session averages per mouse, with the exception of no significant difference during the first head entry after reward. Thus, prior chronic alcohol exposure largely decreased calcium activity in the indirect pathway neuron population.

Chronic ethanol effects on striatal interneuron calcium activity during actions

The study also investigated whether chronic alcohol altered calcium activity modulation in fast-spiking PV interneurons in the striatum. Similar to the other groups, acquisition behavior was comparable between Air and CIE-exposed mice. Lever press rate and rewards earned showed improvement over days, but no significant differences were found between treatment groups or in the total number of lever presses made.

Calcium activity of PV Cre+ interneurons showed reduced modulation in CIE mice compared to Air mice during lever pressing and at the onset of lever press bouts. However, compared to Air mice, CIE mice displayed increased calcium activity modulation of PV Cre+ interneurons during reward-related periods. Analysis of session averages per mouse revealed the same pattern of calcium modulation. This suggests that prior chronic alcohol exposure led to reduced action-related recruitment of calcium activity in PV interneurons, but increased calcium modulation during reward-related times in the DMS during action control.

Discussion

This study demonstrates that prior chronic alcohol exposure leads to a long-lasting imbalance in the activity of different neuron populations within the dorsal medial striatum (DMS) during behaviors related to action. This imbalance was characterized by increased calcium activity modulation in direct pathway neurons, while calcium activity modulation was largely decreased in indirect pathway neurons and in PV interneurons within the DMS. These findings align with previous research suggesting that chronic alcohol results in enhanced activation of direct pathway neurons and increased inhibition of indirect pathway neurons. For the first time, this study shows that such chronic alcohol-induced changes are reflected in the natural calcium activity of DMS pathways during behaviors involving action. These results provide evidence that chronic alcohol creates an imbalance in DMS activity, which may reduce the functional control that the DMS exerts over behavior.

While intact DMS function is widely known to be essential for goal-directed control, the proper balance of activity between the direct and indirect pathways has recently received more attention. Both direct and indirect pathways are active during lever pressing and can support self-stimulation. Effective behavioral control by the DMS is thought to involve the coordinated and balanced activity of direct and indirect pathways, influenced by inputs from the cortex and thalamus. Recent examples show that chronic alcohol-induced changes in the strength of certain cortical inputs can lead to a loss of goal-directed control. Specifically, a reduction in the transmission from the orbitofrontal cortex to the direct, but not indirect, pathway in the DMS has been noted. One interpretation of this is that reduced cortical input into the DMS leads to less DMS engagement and a loss of goal-directed control.

However, other studies examining different prefrontal cortex projections have shown enhanced inputs to direct pathway (D1) neurons in the DMS, as well as decreased thalamic input to the striatum. In the present study, increased calcium activity modulation was observed in D1 neurons, but decreased modulation was seen in D2 neurons and a local interneuron population. Interestingly, there was also reduced modulation of PV interneurons in the DMS. Recent work demonstrated that ablating this neuron population in the dorsal lateral striatum led to reduced organized ethanol consumption, which became more sensitive to the addition of quinine. Since interneurons can strongly regulate the firing of striatal output neurons and are partly activated by cortical input, the data suggests that less modulation of PV interneurons might mean this population is less engaged in shaping the activity patterns of striatal output neurons during actions. These findings, combined with previous work showing alcohol-induced changes to the striatal output neurons themselves, suggest that dysfunction or a loss of DMS control over behavior could result from either an increase or decrease in the strength of specific cortical or thalamic inputs carrying relevant information, and/or through functional changes within the DMS neurons themselves. Further research is needed to understand how chronic alcohol disrupts the ability of converging inputs to activate appropriate DMS activity.

Care should be taken when interpreting these results, as bulk population calcium signals in striatal neurons may not directly reflect action potential firing. Given the extensive branching of these neurons, the bulk signals collected in this data likely represent a combination of calcium activity at and below the firing threshold in the cell body, dendrites, and axon collaterals within the striatum. The study did not examine sensitivity to outcome devaluation, so it cannot be stated whether the subjects in this study relied more on habitual or goal-directed control, although prior research has shown that chronic alcohol exposure impairs goal-directed control for non-alcohol rewards. Nevertheless, the data does suggest that chronic alcohol resulted in altered calcium modulation in identified striatal neuron populations during the performance of actions. Since previous findings indicated that DMS inactivation was ineffective at altering behavioral control after chronic alcohol exposure, this data suggests that changes to the activity of DMS populations may underlie the deficits in using the DMS for behavioral control. It is hypothesized that this imbalance in activity contributes to the reduced involvement of the DMS in goal-directed control following chronic alcohol exposure.

Introduction

Changes in brain function and the ongoing, uncontrolled use of alcohol are key signs of Alcohol Use Disorder (AUD). These changes are believed to stem, in part, from the long-term effects of alcohol on brain circuits in the cortex and striatum. For example, the dorsal medial region of the striatum (DMS) controls goal-directed actions, and chronic alcohol has long been known to alter this area. Previous studies suggested that activity within the DMS changes differently after long-term alcohol exposure, leading to a loss of behavioral control. However, how chronic alcohol specifically affects the activity of various neuron groups within the DMS during action control remains unclear.

The DMS is a complex brain area containing two main types of output neurons, often called the direct and indirect pathways. The direct pathway contains spiny neurons (SPNs) that connect directly to the Substantia Nigra Pars Reticulata and respond to dopamine type-1 receptors (D1-SPNs). In contrast, SPNs of the indirect pathway connect directly to the Globus Pallidus and respond to dopamine type-2 receptors (D2-SPNs). Different types of interneurons, including fast-spiking neurons identified by their Parvalbumin (PV interneurons) expression, help regulate these output pathways by strongly inhibiting SPN activity. The entire striatum receives signals, mostly from the cortex and thalamus. The DMS specifically gets input from areas related to thought and feeling, making it well-suited to process information important for goal-directed control. Research across many species confirms that a healthy DMS is crucial for goal-directed behavior.

Chronic alcohol use has been shown to disrupt the ability to control goal-directed behaviors, affecting both general reward-seeking and the self-administration of alcohol. This disruption has been linked to specific changes in how the cortex and striatum communicate, as well as general changes in signal transmission within the DMS itself. Generally, long-term alcohol use has been associated with reduced inhibitory signals and increased excitatory signals in the DMS, though recent findings suggest this can vary depending on the specific connections studied. Importantly for the current investigation, a previous study of chronic alcohol consumption reported stronger excitatory signals to the direct pathway and stronger inhibitory signals to the indirect pathway. Replicating or counteracting these signal changes in living organisms either increased or decreased alcohol consumption. Furthermore, both short-term and long-term alcohol use have been shown to affect signals to and from PV interneurons, and striatal PV interneurons contribute to compulsive alcohol consumption. It is thought that alcohol self-administration initially involves striatal function but eventually becomes less dependent on the DMS, suggesting that chronic alcohol may cause an imbalance in DMS activity, leading to faulty goal-directed control.

This study examines the calcium activity of DMS direct and indirect pathways, as well as local PV interneurons, during actions after long-term alcohol exposure. A reinforcement schedule designed to encourage goal-directed action control was used in mice. Previous alcohol exposure has been shown to disrupt goal-directed control under this schedule. While bulk calcium fiber photometry was used to measure activity in specific neuron populations, it is important to note that within the striatum, bulk calcium activity does not directly reflect the number of action potentials (nerve impulses). However, bulk calcium measurements do provide insights into calcium activity in both dendrites and cell bodies, which likely reflects incoming signals and is related to how the brain changes and adapts. Consistent with previous theories, this study found that prior chronic alcohol exposure led to lasting increases in calcium activity in the direct pathway during decision-making, but reduced calcium activity in the indirect pathway and PV fast-spiking interneurons. These findings support the idea that an imbalance in DMS population activity contributes to the DMS having less influence over behavioral control.

Methods and Materials

For these studies, specific strains of C57Bl/6 J mice and genetically modified mice (PV Cre+, D1-Cre+, A2a-Cre) were used. These were adult male and female mice, housed in groups, with access to food and water. They were kept on a regular light/dark cycle, and all experiments took place during the light phase. Animal care and experimental procedures were approved and followed national guidelines.

All mice underwent surgery to inject a specific virus (GCaMP6s) into the DMS or related brain areas, which made certain neurons light up when active. This allowed researchers to observe the activity of direct pathway neurons, indirect pathway neurons, or parvalbumin (PV) interneurons. An optical fiber was then implanted directly above the injection site to record light signals from these active neurons. After behavioral procedures, the mice's brains were examined to confirm the correct placement of the optical fiber. Only data from mice with a clear signal and correct fiber placement within the DMS were used for analysis.

At least one week after surgery, mice were exposed to either ethanol vapor (alcohol) or plain air, followed by repeated withdrawal periods. This involved four rounds of 16 hours of vapor exposure followed by 8 hours of withdrawal, repeated for four days, then three days without vapor exposure. This method aimed to simulate chronic intermittent ethanol exposure (CIE). Blood alcohol levels were monitored in sentinel mice throughout the process.

To prepare for behavioral training, food intake was restricted about three days after the last vapor exposure to reduce body weight. Behavioral training began five days after the last vapor procedure. Mice were trained in special operant boxes to press a lever for a food reward. They first learned to retrieve rewards delivered at random times. Then, they moved to a schedule where every lever press yielded a reward. Finally, they were trained on a random ratio (RR) schedule, where an average number of presses (e.g., 10 or 20) was required for a reward, which encourages goal-directed behavior. Sessions ended after a certain number of rewards or 60 minutes.

During the last few days of behavioral training, optical fibers were connected to record light signals from the active neurons. A blue light was used to excite the neurons in the dorsal striatum, and their fluorescent emissions were captured by a camera. Fluorescence intensity and behavioral events (lever presses, head entries, reward delivery) were recorded simultaneously. Raw signals were filtered to reduce noise and corrected for signal degradation. Sessions with low signal or fiber detachment were excluded. Data analysis involved calculating changes in fluorescence intensity around specific behavioral events by comparing them to a baseline period. This "z-score normalization" allowed for precise measurement of neuron activity during actions. Statistical analysis compared activity between alcohol-exposed and air-exposed groups using specific tests to determine significant differences.

Results

After surgery and alcohol or air exposure, mice in both the air-exposed (Air) and chronic intermittent ethanol-exposed (CIE) groups, whose direct pathway neurons were monitored, learned the lever-pressing behavior similarly. Statistical analyses confirmed that over time, all groups improved in lever presses, lever-press rate, and rewards earned, but there were no significant differences between the Air and CIE groups in these general behavioral measures. This showed that prior chronic alcohol exposure did not cause major changes in how the mice learned lever-pressing actions.

When examining the calcium activity of direct pathway neurons, these neurons showed changes in activity around lever-pressing behaviors. However, CIE mice generally displayed increased calcium activity compared to Air mice. Specifically, there was significantly greater calcium activity in CIE mice around each lever press. While CIE mice showed a reduced increase in calcium activity before a series of lever presses, they exhibited a large and significant increase in calcium activity during the first head entry after a reward was delivered. This pattern was also seen when activity was analyzed per mouse session average, except for the lack of differences at the start of a lever-press series. Therefore, previous CIE exposure led to a lasting increase in the calcium activity of the direct pathway related to lever pressing and decision-making.

Mice with monitored indirect pathway neurons, like the direct pathway group, also learned the lever-press behavior to similar extents, with improvements over time but no significant differences between Air and CIE groups in overall behavior.

In contrast to the direct pathway, calcium activity in the indirect pathway was largely reduced in CIE mice compared to Air mice. While indirect pathway calcium activity naturally increased before lever pressing, this increase was significantly smaller in CIE mice both around individual lever presses and at the start of a lever-press series. Reward-related indirect pathway calcium activity was also slightly, but significantly, greater in Air mice compared to CIE mice, though the difference was less pronounced. A similar pattern was observed when data were analyzed by session averages per mouse, with the exception of the reward-related activity. Thus, prior chronic alcohol exposure generally decreased calcium activity in the indirect pathway.

The study also investigated whether chronic alcohol altered calcium activity in fast-spiking PV interneurons within the striatum. Again, both Air and CIE exposed mice showed similar learning of the lever-press task, with improvements over time but no differences in general behavior between the groups.

Calcium activity in PV interneurons showed reduced changes in CIE mice compared to Air mice during lever pressing and at the beginning of lever-press series. However, interestingly, CIE mice displayed increased calcium activity in PV interneurons during reward-related periods compared to Air mice. Analyzing session averages per mouse revealed the same pattern of calcium activity. These findings suggest that prior chronic alcohol exposure led to reduced recruitment of calcium activity in PV interneurons during actions, but increased calcium activity during reward-related events in the DMS during behavioral control.

Discussion

This study demonstrates that previous chronic alcohol exposure leads to a lasting imbalance in the activity of different neuron populations within the DMS during action-related behaviors. This imbalance was evident as increased calcium activity in direct pathway neurons, while calcium activity was largely decreased in indirect pathway neurons and PV interneurons within the DMS. These changes align with earlier findings suggesting that chronic alcohol increases the activity of direct pathway neurons and enhances inhibition of indirect pathway neurons. For the first time, these findings show that such long-term alcohol-induced changes are reflected in the natural calcium activity of DMS pathways during behaviors related to action. This evidence supports the idea that chronic alcohol causes an imbalance in DMS activity, which may lead to the DMS having less effective control over behavior.

While a healthy DMS is widely recognized as essential for goal-directed control, the proper balance of activity between the direct and indirect pathways has recently gained more attention. Both direct and indirect pathways are active during lever pressing and can support self-stimulation of the brain. Effective behavioral control by the DMS is thought to involve the cortex and thalamus coordinating balanced activity in these direct and indirect pathways. Recent research shows that chronic alcohol-induced changes in the strength of specific cortical inputs can lead to a loss of goal-directed control. For instance, one study found a reduction in signals from the orbitofrontal cortex to the direct, but not indirect, pathway in the DMS. This led to the interpretation that reduced cortical input (even from one area) to the DMS could result in less DMS involvement and a loss of goal-directed control.

However, other studies examining different areas of the prefrontal cortex have shown increased inputs to direct pathway neurons in the DMS, along with decreased input from the thalamus to the striatum. In the current study, increased calcium activity was observed in direct pathway neurons, but decreased activity in indirect pathway neurons and a local interneuron population. Intriguingly, there was also reduced activity in PV interneurons within the DMS during actions. Recent work has shown that removing this neuron population in a different part of the striatum led to less organized ethanol consumption, with increased sensitivity to bitterness. Since interneurons can strongly regulate the firing of other neurons and are partly activated by cortical input, the data suggests that less activity in PV interneurons might mean this group is less involved in shaping how other neurons in the DMS respond during actions. These findings, along with prior research showing alcohol-induced changes to the neurons themselves, suggest that dysfunction or loss of DMS control over behavior could result from changes in the strength of specific cortical or thalamic inputs carrying relevant information, or from functional changes within the DMS neurons themselves. Further research is needed to understand how chronic alcohol disrupts the ability of combined inputs to generate appropriate DMS activity.

Care should be taken when interpreting these results, as general population calcium activity in striatal neurons may not directly reflect the firing of individual nerve impulses. Given the complex branching of these neurons, the bulk signals collected in this study likely represent a combination of calcium activity in the cell body, branches, and connections within the striatum, both at and below the threshold for firing. This study did not examine how sensitive the mice were to changes in reward value, so it cannot be said whether the subjects were relying more on automatic habits or goal-directed control, although previous work has shown that chronic alcohol exposure disrupts goal-directed control for non-alcohol rewards. However, the data does suggest that chronic alcohol led to altered calcium activity in identified neuron populations in the DMS during action performance. Since prior findings showed that inactivating the DMS was ineffective at changing behavioral control after chronic alcohol, the data suggest that alterations to the activity of DMS populations may underlie difficulties in using the DMS for behavioral control. It is further hypothesized that this imbalance in neuron recruitment likely contributes to the reduced role of the DMS in goal-directed control after chronic alcohol exposure.

Introduction

When people drink too much alcohol for a long time, it can lead to Alcohol Use Disorder (AUD). This problem often causes changes in how a person thinks and controls their actions. Scientists believe this happens because long-term alcohol use harms certain parts of the brain. One important brain area is called the DMS. It helps control actions that are done on purpose. Past studies have shown that long-term alcohol use changes how the DMS works, leading to a loss of control over actions. But it is not yet clear exactly how alcohol changes the activity of different cells in the DMS when a person is trying to control their actions.

The DMS brain area is made up of different kinds of cells. It has two main groups of cells, often called the "direct pathway" and the "indirect pathway." These cells send signals out of the DMS. There are also other helper cells called interneurons, like "PV interneurons," that help control the main cells. Signals come into the DMS from other parts of the brain, like the cortex and thalamus. The DMS uses all these signals to help guide actions and make choices. Many studies show that a healthy DMS is very important for acting on purpose.

Long-term alcohol use is known to make it harder for people to control their actions, especially when it comes to seeking rewards or alcohol itself. This happens because alcohol changes how brain cells in the DMS send signals to each other. For example, some studies found that alcohol makes signals stronger in the direct pathway and stronger in the indirect pathway. If these changes were copied in animals, it made them drink more alcohol. If these changes were stopped, it made them drink less. Alcohol also affects the PV interneurons, and these cells play a role in heavy alcohol use. It is thought that over time, alcohol use changes how the DMS works, making it less able to guide actions properly.

In this study, scientists looked at how activity in the DMS brain area changes after long-term alcohol exposure. They studied the direct and indirect pathways, along with the PV interneurons. They used a special way to train mice that helped them see how alcohol affects goal-directed actions. The scientists measured "calcium activity," which shows how active brain cells are. They found that long-term alcohol exposure caused lasting changes: the direct pathway cells became more active during decisions, but the indirect pathway cells and PV interneurons became less active. These findings suggest that long-term alcohol use creates an "imbalance" in the DMS. This imbalance may make the DMS less able to help control actions.

Methods and materials

Scientists used different types of mice for these studies, including both male and female mice over 8 weeks old. The mice lived in groups and had plenty of food and water. All experiments happened during the daytime.

To study the brain cells, the mice had a small surgery. Scientists injected a special substance (GCaMP6s) into a brain area called the DMS. This substance makes brain cells light up when they are active. Then, a tiny glass fiber was placed above the injection site to measure the light signals. After the experiments, the scientists checked the brains to make sure the injections and fiber placements were correct.

After surgery, some mice were given alcohol vapor, while others breathed regular air. This was done in cycles: 16 hours of alcohol vapor, then 8 hours without, for four days in a row. This was repeated four times. This process is like how people might drink alcohol heavily and then stop, causing withdrawal.

Next, the mice were trained to press a lever for food. This started a few days after the alcohol or air exposure ended. First, they learned to get food from a dish. Then, they learned that pressing a lever gave them food. Slowly, the task got harder, making them press the lever more times to get food. This training helped scientists study how the mice controlled their actions.

While the mice were pressing the lever, scientists used special equipment to measure the light signals from their brain cells. This "fiber photometry" showed how active the different brain cells were during their actions. The signals were cleaned up to remove any extra noise. Scientists then used math and computer programs to compare the brain cell activity between the mice exposed to alcohol and those that breathed air. They looked at how activity changed when the mice pressed the lever, started a set of presses, or received their food reward.

Results

After the surgery and alcohol exposure, both groups of mice (those exposed to alcohol and those exposed to air) learned to press the lever for food in a similar way. The alcohol exposure did not cause big changes in how they learned this new task or how many times they pressed the lever.

Next, scientists looked at the activity of the "direct pathway" brain cells in the DMS. They found that in mice exposed to alcohol, these cells became much more active when the mice were making decisions, like pressing the lever. This was also true when the mice received their food reward. This means that long-term alcohol exposure made the direct pathway cells more active during actions.

Then, the scientists looked at the "indirect pathway" brain cells. Unlike the direct pathway, these cells showed much less activity in mice exposed to alcohol compared to mice that breathed air. This was true when the mice pressed the lever and when they started a set of lever presses. There was also a small decrease in activity when they received their reward. So, long-term alcohol exposure made the indirect pathway cells less active.

Finally, scientists studied the "PV interneurons" in the DMS. These helper cells also showed less activity in alcohol-exposed mice during lever pressing and when starting a set of presses. However, during the reward time (when they got food), these cells actually showed more activity in the alcohol-exposed mice. This suggests that long-term alcohol exposure changed how PV interneurons work during actions and when getting rewards.

Discussion

This study shows that drinking alcohol for a long time causes a lasting problem in how different groups of brain cells in the DMS work during actions. The "direct pathway" cells became more active, while the "indirect pathway" cells and the "PV interneurons" became less active. This means the DMS is out of balance. These findings match what other studies have suggested: long-term alcohol use makes some brain cells more active and others less active. This study is the first to show these changes in real-time activity when animals are performing actions. These results suggest that long-term alcohol use makes the DMS work in a way that leads to less control over actions.

A healthy DMS is known to be vital for making choices on purpose. Lately, scientists have learned that the direct and indirect pathways in the DMS need to work together in a balanced way. Both pathways are active when an animal presses a lever. For the DMS to guide behavior well, signals from other brain areas, like the cortex and thalamus, need to create a good balance in these pathways. Some studies have shown that long-term alcohol use changes how strong these signals are, which can lead to a loss of controlled actions. For example, signals from one part of the cortex to the direct pathway might get weaker after alcohol use. This could mean that the DMS is not getting enough strong signals to guide actions properly.

Other studies have found different changes, such as stronger signals from some parts of the brain to the direct pathway, and weaker signals from the thalamus. In this study, the direct pathway cells showed more activity, but the indirect pathway cells and the PV interneurons showed less. It was interesting that the PV interneurons in the DMS showed less activity. Other research has shown that removing these PV interneurons in another part of the brain reduced heavy alcohol drinking. Since interneurons help control how other brain cells fire, our results suggest that the PV interneurons might not be working as well to guide the activity of the main brain cells during actions. All these findings together suggest that long-term alcohol use might harm the DMS's ability to control behavior. This could happen because signals coming into the DMS change, or because the DMS cells themselves change. More research is needed to understand how alcohol affects the DMS's ability to use incoming signals properly.

It is important to remember that measuring "calcium activity" shows how active brain cells are, but it does not directly show how many electrical signals they send. This measurement gives a general idea of brain cell activity. In this study, scientists did not test if the mice were acting out of habit or making choices on purpose. However, other studies have shown that long-term alcohol exposure makes it harder for animals to make choices on purpose for non-alcohol rewards. Our results show that alcohol changed the activity of certain brain cells in the DMS during actions. Other research has found that turning off the DMS did not help control behavior after long-term alcohol use. This means the changes in DMS cell activity seen in our study might explain why the DMS becomes less useful for guiding actions after long-term alcohol use. Scientists believe this imbalance in brain cell activity makes the DMS less able to help control actions that are done on purpose after long-term alcohol use.