Abstract
Previous work has implicated the nucleus accumbens (NAc) in the regulation of effort, defined as the amount of work an animal is willing to perform for a given reward, but little is known about the specific contributions of neuronal populations within the NAc to effort regulation. In this study, using male and female mice, we examined the contributions of direct pathway and indirect pathway neurons in the NAc core using an operant effort regulation task, in which the effort requirement is the number of lever presses needed for earning a food reward. Using optogenetics, we manipulated the activity of direct pathway spiny projection neurons (SPNs; dopamine D1-like, D1+) and indirect pathway SPNs (iSPNs; adenosine 2A receptor, A2A+). Activating dSPNs reduced lever pressing regardless of the effort requirement, as it elicited gnawing, a competing consummatory behavior. On the other hand, activating iSPNs in the NAc core (but not in the shell) reduced lever pressing in an effort-dependent manner: stimulation-induced reduction in performance was greater at higher press-to-reward ratio requirements. In contrast, optogenetically inhibiting NAc core iSPN output resulted in increased levels of effort exertion. Our results show that the indirect pathway output from the NAc core can bidirectionally regulate effort exertion.
Significance Statement
Using bidirectional optogenetic manipulation to manipulate direct and indirect pathway neurons in the nucleus accumbens core, we found that activating the direct pathway neurons reduced lever pressing regardless of the effort requirement, as it elicited competing consummatory behaviors like gnawing. On the other hand, activating the indirect pathway neurons in the NAc core reduced lever pressing in an effort-dependent manner: stimulation-inducted reduction in performance was greater at higher press-to-reward ratio requirements.
Introduction
The nucleus accumbens (NAc) is believed to play a fundamental role in regulating motivation. As a major component of the limbic basal ganglia circuit, it has long been thought to function as an interface between the neural substrates of motor control and limbic structures governing affect and homeostasis (Mogenson et al., 1980; Kelley, 2004; Yin et al., 2008; Nicola, 2010; Vancraeyenest et al., 2020). Although the NAc is not necessary for learning action–outcome contingencies, it plays a major role in regulating instrumental performance (Corbit et al., 2001). Previous studies have established that the NAc is essential for the regulation of effort exertion (Aberman and Salamone, 1999; Kelley, 2004; Ghods-Sharifi and Floresco, 2010). Effort can be distinguished from motor output per se. It is defined as the amount of work an animal is willing to perform for a given amount of reward. Instrumental or operant conditioning provides a convenient way to quantify effort by measuring the number of lever presses the animal is willing to perform for a given amount of reward. For example, when a fixed number of presses (fixed ratio or FR) is required to earn a piece of food reward, the ratio represents the effort requirement.
Previous work has suggested that dopamine (DA) depletion in the NAc does not disrupt food consumption per se but drastically reduces the number of lever presses for food rewards at high ratio requirements (Sokolowski and Salamone, 1998; Aberman and Salamone, 1999; Salamone et al., 2001). Depleted rats were still able to consume just as much home chow as normal rats, but they were much less willing to press when the ratio was high. Furthermore, rats with DA depletion in the NAc core show a reduced preference for effortful but large-reward actions (Mai et al., 2012). Such results suggest that mesolimbic DA may regulate effort exertion. It has also been shown that DA signaling in the NAc often reflects effort-related variables (Day et al., 2010; Gan et al., 2010; Ostlund et al., 2011).
Like the rest of the striatum, the NAc contains two major populations of spiny projection neurons (SPNs), which are typically classified as either direct pathway SPNs (dSPNs) or indirect pathway SPNs (iSPNs). The direct pathway neurons express D1-like DA receptors, whereas the indirect pathway neurons coexpress A2A adenosine and D2 DA receptors (Gerfen et al., 1990; Fink et al., 1992; Lu et al., 1997). Pharmacological studies have implicated the indirect pathway in regulating effort exertion. A2A agonists, which presumably increase the activity of iSPNs by activating Gs-coupled A2A receptors and increasing neuronal excitability, disrupt performance on operant tasks with high-effort requirements (Mingote et al., 2008). This finding suggests that increasing iSPN activity in the NAc core can reduce effort exertion without affecting general motor output. In contrast, genetic manipulations that reduce Gpr6 expression in iSPNs can enhance effort exertion in mice (Lobo et al., 2007). As Gpr6 normally sets the tone for Gs signaling in iSPNs, reducing its expression is expected to reduce iSPN output. However, the role of the direct pathway in effort exertion remains unclear as previous studies did not directly manipulate neural activity with high temporal and spatial resolution.
In the present study, we used optogenetics to test the contributions of the direct and indirect pathways by imposing distinct schedules of reinforcement and associated effort requirements in mice. FR schedules were used to manipulate the effort requirement without changing the reward amount or delay. To specifically manipulate the activity of direct and indirect pathway neurons, optogenetic stimulation was done in Cre driver lines (D1-Cre for the direct pathway and A2A-Cre for the indirect pathway). Our results suggest that activation of NAc core iSPNs reduces effort exertion. On the other hand, while stimulation of dSPNs also reduces lever pressing, it does so by eliciting competing consummatory behaviors like gnawing that interfere with lever pressing.
Materials and Methods
Animals
Experiments were conducted using adult mice (n = 64, 3–8 months), housed individually or in same-sex groups of no more than four. Three genotypes were used: D1-Cre mice (Drd1a-Cre, GENSAT, MMRRC_030989-UCD, n = 10, 3 males, 7 females), A2A-Cre mice (Adora2a-cre, MMRC_036158-UCD, n = 28, 14 males, 14 females), and wild-type control animals (C57BL/6J, The Jackson Laboratory, n = 26, 12 males, 14 females). Cre-positive mice were maintained in hemizygosity. Mice had free access to water, except during experimental sessions. While undergoing food restriction, mice were weighed daily and then provided with chow to maintain approximately 85% ad libitum body weights. All procedures were approved by Duke University's Institutional Animal Care and Use Committee (protocol A162-22-09).
Surgery
Before intracranial surgeries, general anesthesia was initiated via isoflurane inhalation, and mice were placed upon a stereotaxic surgical apparatus. Four separate 50 nl intracranial injections of either a Cre-dependent ChR2 vector (2 × 1013 vg/ml, pAAV5-Ef1a-DIO-hChR2(E123 T/T159C)-EYFP, Addgene_35509) or a Cre-dependent GtACR2 vector (4.2–21 × 1011 vg/ml, pAAV1-hSyn1-SIO-stGtACR2-FusionRed, Addgene_105677) were injected into each hemisphere over the course of approximately 10 min. To target the NAc core (n = 36), injections were administered targeting the following coordinates: 1 mm AP, 1.3 mm ML, and 4 mm DV; 1 mm AP, 1.3 mm ML, and 4.2 mm DV; 1.4 mm AP, 1.3 mm ML, and 4 mm DV; and 1.4 mm AP, 1.3 mm ML, and 4.2 mm DV. To target the shell region (n = 5), the following coordinates were used: 1 mm AP, 0.4 mm ML, and 3.8 mm DV; 1 mm AP, 0.4 mm ML, and 4 mm DV; 1.4 mm AP, 0.4 mm ML, and 3.8 mm DV; and 1.4 mm AP, 0.4 mm ML, and 4 mm DV. Subsequently, optic fibers measuring 100 μm in diameter were inserted at an angle of 10° from vertical (core coordinates for fiber tips: 1.2 mm AP, 1.4 mm ML, and 3.8 mm DV; shell coordinates: 1.2 mm AP, 0.8 mm ML, and 3.7 mm DV).
Lever press training
Several weeks after surgery, food restriction to 85% body weight was initiated, and mice were trained for an hour per day in 18 × 22 cm operant chambers (Med Associates). Mice were first trained on a fixed ratio 1 (FR-1) schedule of reinforcement: a 14 mg food pellet (Bio-Serv) was dispensed after each press of the single accessible lever. Timestamps were recorded for each lever press and each break of an infrared beam for detecting food port entries. After initial FR-1 training, each mouse was trained on an FR-10 schedule of reinforcement; in other words, a pellet was dispensed after 10 presses. After 2–3 days of FR-10 training, two daily optogenetic sessions tested 5 and 8 Hz stimulation frequencies. This course of training and testing was then repeated for FR-20 and FR-40 schedules of reinforcement.
Optogenetic experiments
Optogenetic experiments were conducted using microcontroller-driven (Arduino, SCR_017284) 473 nm diode-pumed solid-state lasers. For inhibition experiments (n = 17), constant illumination was delivered during six alternating 5 min epochs starting in 10 min intervals in hour-long sessions. For excitation experiments (n = 36), 10 ms pulses were delivered during the same 5 min epochs. During two daily sessions, 5 and 8 Hz stimulation frequencies were tested. This was repeated for each ratio/effort requirement. For inhibition experiments, the laser power was approximately 5 mW in the brain. For excitation experiments, the laser power was approximately 3–4 mW, but an effect was elicited in one A2A+ NAc core mouse with 1 mW.
Quantification and Statistical Analysis
Gnawing measure
Inside the operant chamber, gnawing was quantified using DeepLabCut and Matlab (MathWorks). A high-resolution video camera (Blackfly S, Teledyne forward-looking infrared) recorded mice from above. DeepLabCut identified the corners of the reward port and tracked mice's snouts over time. A custom Matlab script used the corner locations to create a zone around the bottom edge of the reward port (∼10 mm diameter), as well as the bottom half of the left and right edges. When the snout was within the zone, the behavior was classified as gnawing.
Outside the operant chamber, gnawing was assayed in a 22 × 22 cm enclosure with a transparent floor. One to four loops of black polylactide filament (1.5 mm in diameter, extending roughly 4 cm), on which mice would be able to gnaw, were fixed at each corner of the enclosure. Optogenetic manipulations were performed in an identical manner to that previously described, except that these sessions lasted 30 min, and that two sessions testing 5 and 8 Hz stimulation frequencies were administered on a single testing day. This assay was always conducted after the completion of operant testing. To quantify gnawing, a video was recorded from a camera below the enclosure, and the video data were analyzed with DeepLabCut and Matlab. DeepLabCut tracked the position of the mouth, and Matlab identified the loop locations. Behavior was classified as chewing when the mouth was within 2.5 mm of the loop.
Locomotion measure in gnawing chamber
Locomotion was quantified from video collected during stimulation in the gnawing chamber described above, using DeepLabCut and Matlab. DeepLabCut tracked the position of all four paws, and Matlab used these markers to calculate the center of support, the basis for our measure of locomotion. Locomotion speed was obtained by calculating the change in the location of the center of support across frames, computed as distance traveled per second (in 100 ms bins).
Locomotion measure in open field
Locomotion was quantified using Bonsai and Matlab (Lopes et al., 2015). Bonsai tracked the location of mice as they moved around the open field, and Matlab determined locomotion speed by calculating the change in location across frames, computed as distance traveled per second (in 100 ms bins).
Statistical analysis
All statistical tests (linear regressions, ANOVAs, t tests) were performed in GraphPad Prism (GraphPad Software). Percent change was calculated with the formula: (laser − noLaser)/abs(noLaser) * 100, where laser and noLaser represent the aggregate behavior across all light on and light off epochs, respectively.
Histology
Mice were transcardially perfused with 0.1 M phosphate buffered saline and then 4% paraformaldehyde. Brains were extracted and submerged in 4% paraformaldehyde for at least 4 d and then in 30% sucrose for at least 2 d. Sixty micrometer coronal sectioning was then performed via cryostat, and sections were mounted with an aqueous Fluoroshield DAPI mounting medium (Abcam, Ab104139) and imaged (ZEISS Axio Zoom.V16, SCR_016980).
Results
Initial operant experiments were conducted to determine whether the willingness of mice to exert effort would be impacted by optogenetically stimulating either core dSPNs or iSPNs. In order to selectively activate direct (D1+) or indirect (A2A+) pathway neurons in the NAc core, we used Cre-dependent channelrhodopsin (ChR2) and Cre driver lines that express Cre-recombinase under the promoters for the D1 or A2A receptors (Gerfen et al., 2013).
We first injected Cre-dependent ChR2 bilaterally in D1-Cre (n = 10, three males and seven females) and A2A-Cre mice (n = 9, four males and five females) and implanted optic fibers targeting the NAc core. The control group consisted of wild-type (WT) mice that received the same injections and fiber implants (n = 10, five males and five females). Histological analysis of the mouse brains after the completion of all experiments confirmed that the fibers were implanted in the NAc core (Fig. 1A, Extended Data Fig. 1-1).
Virus expression, fiber placement, and operant training in mice expressing ChR2 in the NAc core. A, Expression of Cre-dependent ChR2 and placement of optic fibers in the NAc core. Left, Fiber placement and representative histology for the direct pathway. Right, Indirect pathway. B, Acquisition data from wild-type (WT) control, D1-Cre, and A2A-Cre mice. There were no group differences in acquisition. All three groups received the same virus injection in the NAc core. Error bars indicate SEM. See also Extended Data Figure 1-1.
Figure 1-1
NAC core pathway-specific viral expression. A) Cre-dependent ChR2 expression in the direct pathway. D1 + neurons project to the ventral pallidum (VP) and substantia nigra pars reticulata (SNr). B) Cre-dependent ChR2 expression in the indirect pathway. A2A + neurons project to the VP but not SNr. Same imaging parameters for D1 + and A2A + neurons. Download Figure 1-1, TIF file.
After recovering from surgery, all mice readily acquired lever pressing during FR training. Their rates of lever pressing during training sessions are shown in Figure 1B and increased significantly for all three groups (main effect of session: F(11, 277) = 82.88, p < 0.0001).
During stimulation sessions, photostimulation immediately reduced lever pressing in both direct and indirect groups. From one epoch to the next, pressing reductions caused by stimulation were clearly observed after direct pathway activation (8 Hz stimulation data are shown in Fig. 2A). Likewise, indirect pathway stimulation also reduced lever pressing (Fig. 2B).
Time course of training and optogenetic testing for operant experiments. A, Activation of the D1+ (direct pathway) neurons in the NAc core reduced lever pressing. Blue bars indicate time of photostimulation (473 nm, ChR2, 8 Hz, 10 ms pulse width). B, Activation of the A2A+ (indirect pathway) neurons in the NAc core reduced lever pressing. Blue bars indicate time of photostimulation (473 nm, ChR2). Error bars indicate SEM.
Figure 3 shows a summary of the change in lever pressing during stimulation sessions. For FR-10 sessions, a two-way ANOVA revealed a significant interaction between genotype (WT, D1-Cre, and A2A-Cre) and frequency (5 and 8 Hz; F(2, 26) = 4.170, p = 0.0269). Post hoc analysis showed no differences between groups in the 5 Hz condition, but the D1+ (p = 0.0060) and A2A+ (p = 0.0048) groups showed a larger reduction in lever pressing than the WT group in the 8 Hz condition. For FR-20 sessions, the interaction was also significant (F(2, 26) = 8.744, p = 0.0012). Post hoc analysis showed that the reduction caused by 5 Hz stimulation in the A2A+ group was greater than in the control group (p = 0.0009). For the 8 Hz condition, both the D1+ (p < 0.0001) and the A2A+ groups (p < 0.0001) showed greater reductions in pressing compared with controls. In the FR-40 sessions, there was no interaction (F(2, 26) = 2.690, p = 0.0868), and there were significant main effects of genotype (F(2, 26) = 15.26, p < 0.0001) and frequency (F(1, 26) = 6.881, p = 0.0144). Post hoc analysis showed that reductions in pressing caused by 5 Hz were larger for the A2A+ group than the control group (p < 0.0001) and the D1+ group (p = 0.0140). Reductions caused by 8 Hz were greater in both the D1+ (p = 0.0007) and A2A+ groups (p < 0.0001) than in the control group. The reduction in the A2A+ group was also greater than the reduction in the D1+ group (p = 0.0313).
Summary of stimulation effects in D1+ and A2A+ neurons in the NAc core. ChR2 excitation of direct or indirect pathway neurons significantly reduced the rate of lever pressing. In the A2A+ group, the effect is strongest at FR-40, suggesting that the % decrease in performance is greater when the effort requirement is high. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. p values from post hoc tests following two-way ANOVAs. Error bars indicate SEM.
To determine if the reduction in lever pressing observed depends on effort requirement, we plotted the change in lever pressing as a function of the fixed ratio (the number of presses required per reward). As shown in Figure 4A, in the WT control group and the D1+ group, the reductions in pressing showed no significant relationship with the effort requirement (WT 5 Hz, R2 = 0.0157, p = 0.5094; WT 8 Hz, R2 = 0.0009, p = 0.8730; D1+ 5 Hz, R2 = 0.0168, p = 0.4943; D1+ 8 Hz, R2 = 0.0005, p = 0.9038). In contrast, in the A2A+ group, the percent change in lever pressing increases as the ratio is increased (5 Hz condition, R2 = 0.2338, p = 0.0106; 8 Hz condition, R2 = 0.1434, p = 0.0515).
Change in press rate and reward port entry as a function of fixed ratio requirement. A, Effort-dependent reduction in lever pressing was found only after activation of A2A+ neurons in the NAc core. As the ratio increases, % reduction in the press rate also increases. Regression analysis shows that the reduction in pressing caused by 5 Hz stimulation depends on increasing ratio requirement. This effect was not observed in WT controls or D1+ mice. B, Note that ratio-dependent reduction in performance was not observed in the rate of reward port entry, which requires less effort. *p < 0.05. Error bars indicate SEM.
For comparison, we also plotted the change in the rate of food port entry (Fig. 4B). When a pellet is dispensed, the mice must enter the food port to retrieve it. This behavior was measured using a beam break. Stable rates of port entry can be used as an indication that levels of motor coordination and appetite are consistent, and changes in this measure caused by changes in the ratio requirement are relatively minimal. Our results showed that changes in port entry rate did not depend on effort requirement (WT 5 Hz, R2 = 0.0181, p = 0.4790; WT 8 Hz, R2 = 0.0022, p = 0.8047; D1+ 5 Hz, R2 = 0.0452, p = 0.2593; D1+ 8 Hz, R2 = 0.0166, p = 0.4972; A2A+ 5 Hz, R2 = 0.00004, p = 0.9728; A2A+ 8 Hz, R2 = 0.0929, p = 0.1222). Thus, in all three groups, changes in the rate of food port entry showed no relationship with effort requirement (no significant effect of ratio). Regardless of the ratio schedule, mice were able to check the food port for pellets.
These results suggest that only activation of the indirect pathway (A2A+) in the NAc core had a ratio-dependent effect, resulting in a greater reduction in performance at higher ratios. But activation of the direct pathway reduced lever pressing consistently, regardless of effort requirement. In accord with this finding, we also observed significant gnawing behavior during direct pathway stimulation: when stimulated, D1+ mice often gnawed on the edges of the reward port. To quantify this behavior, we used DeepLabCut to track the position of the snout and determine when it was within a centimeter of the reward port edges (Mathis et al., 2018). We used time in this zone as a putative measure of gnawing. This analysis was done for a subset of D1+ (n = 4) and WT control (n = 5) mice whose operant behavior was recorded by a high-resolution video camera. From one epoch to the next, stimulation reversibly increased the amount of time D1+ mice spent gnawing on the reward port (8 Hz stimulation data are shown in Fig. 5A). Figure 5B shows a summary of the change in gnawing due to stimulation. For the FR-20 and FR-40 schedules, two-way ANOVAs revealed a significant main effect of genotype (FR-20, F(1, 7) = 19.63, p = 0.0030; FR-40, F(1, 7) = 19.63, p = 0.0030). Both 5 and 8 Hz stimulation caused D1+ mice to increase time spent gnawing versus WT controls (FR-20, 5 Hz, p = 0.0259; FR-20, 8 Hz, p = 0.0053; FR-40, 5 Hz, p = 0.0342; FR-40, 8 Hz, p = 0.0027). For the FR-10 schedule, there was no main effect of genotype (F(1, 7) = 3.289, p = 0.1126). However, 8 Hz stimulation did not increase increase time spent gnawing in D1-Cre mice (p = 0.1064). This demonstrates that activation of the direct pathway reduced lever pressing by introducing a competing behavior: gnawing.
Activation of the direct pathway (D1+) causes gnawing, a consummatory behavior that competes with lever pressing behavior. A, Rapid and reversible increase in gnawing behavior elicited by optogenetic activation (8 Hz) of D1+ neurons in the NAc core during operant performance. B, Optogenetic stimulation in D1+ mice causes a significant increase in gnawing. C, Direct pathway activation also causes gnawing behavior in a separate chamber specifically designed to assay gnawing with high precision. Left, Rapid and reversible increase in gnawing behavior elicited by optogenetic activation (8 Hz) of D1+ neurons in the NAc core. In contrast, activating the indirect pathway (A2A+) had no effect on gnawing. Right, % time spent gnawing at different stimulation frequencies. There was a significant difference between D1+ and WT control mice in the percentage of time spent gnawing. *p < 0.05. **p < 0.01. ****p < 0.0001. p values from post hoc tests following two-way ANOVAs. Error bars indicate SEM. See also Extended Data Figure 5-1.
Figure 5-1
Activation of A2A + neurons in the NAC core does not alter peak locomotion speed. Error bars indicate SEM. Download Figure 5-1, TIF file.
However, gnawing is best examined outside of the operant chamber, in an environment where an unobstructed view of the mouth can be obtained. To this end, most mice were also stimulated in a separate chamber, with a clear floor, where a high-resolution video camera could view the mouth from below. In agreement with our quantification of gnawing during operant behavior, activation of D1+ neurons produced gnawing in this context as well (Movie 1). Moreover, 8 Hz stimulation caused all D1+ mice to gnaw more than they did at baseline, and the increase was greater than twofold in all mice except one. Activation of A2A+ neurons, on the other hand, did not produce gnawing (genotype–frequency interaction: F(4, 38) = 3.435, p = 0.0172). Of note, D1 activation increased gnawing in a frequency-dependent manner (linear regression: R2 = 0.2488, p = 0.0081; Fig. 5C).
To further verify that activation of A2A+ neurons in the NAc core does not produce a motor deficit, we investigated whether stimulation produced changes in locomotion inside the gnawing chamber. DeepLabCut analysis revealed there was no significant change in peak locomotion speed from nonstimulated to stimulated epochs at either frequency, compared with WT controls (Extended Data Fig. 5-1). This indicates that mice can move just as quickly during activation of the indirect pathway as they do normally, arguing against a purely motor deficit.
Therefore, while activation of D1+ and A2A+ neurons in the NAc core can both reduce instrumental performance, they do so via different mechanisms. Activation of D1+ neurons elicits consummatory behaviors like gnawing that interfere with lever pressing, whereas activation of A2A+ neurons reduces lever pressing in an effort-dependent fashion. To confirm that this effect is indeed specific to the core region, we also activated A2A+ neurons in the neighboring shell region of the NAc (Fig. 6). This was done by injecting the same Cre-dependent ChR2 into the A2A-Cre (n = 8, five males and three females) and WT mouse lines (n = 8, four males and four females) used previously, but the virus and optic fibers were targeted to the shell region instead of the core. Unlike activation of A2A+ neurons in the NAc core, activation of A2A+ neurons in the shell did not significantly affect behavioral measures: neither effort-dependent nor effort-independent reductions in lever pressing nor port entry were elicited.
Activation of A2A+ neurons in the NAc shell does not affect effort exertion. A, Fiber placement for optogenetic excitation of A2A+ neurons in the NAc shell. B, Representative slice showing virus expression in the NAc shell and optic fiber targeting the shell. C, Optogenetic excitation of A2A+ neurons in the NAc shell did not significantly reduce lever pressing. Data from the NAc shell WT control group are shown for comparison. D, Changes in press rate (top) and reward port entry rate (bottom) as a function of the ratio requirement. WT control data on the left and A2A+ data on the right. Activation of A2A+ neurons in the NAc shell had no significant effect. Error bars indicate SEM.
If activation of the A2A+ neurons in the NAc core is responsible for suppressing effort exertion, inhibition of their activity could have the opposite effect of enhancing effort exertion. To test this prediction, we used the inhibitory opsin GtACR2 to reduce activity in A2A+ NAc core neurons (Mahn et al., 2018). The previously used A2A-Cre (n = 11, five males and six females) and WT mouse lines (n = 8, three males and five females) received the same optic fiber implants and GtACR2 injections. As shown in Figure 7, inhibition of the NAc core A2A+ neurons increased effort exertion as expected (main effect of inhibition: F(1, 17) = 6.330, p = 0.0222).
Optogenetic inhibition of A2A+ neurons in the NAc core increases lever pressing at higher effort requirements. A, Fiber placement for optogenetic inhibition with stGtACR2. B, Representative histology showing virus expression and optic fiber placement. C, Time course showing the effect of optogenetic inhibition of A2A+ neurons in the NAc core. D, At FR-40, inhibition of A2A+ neurons significantly increased lever pressing (*p = 0.0142; two-way ANOVA post hoc test). Error bars indicate SEM. See also Extended Data Figure 7-1.
Figure 7-1
Inhibition of A2A + neurons in the NAC core does not alter peak locomotion speed. Error bars indicate SEM. Download Figure 7-1, TIF file.
Gnawing behavior in a representative D1-Cre mouse expressing ChR2 in direct pathway neurons with 8 Hz stimulation in the NAc core. [View online]
To investigate whether inhibition of A2A+ neurons in the NAc core altered locomotion, a subset of the mice (A2A-Cre, n = 6; WT, n = 3) were also tested in the open field. The optogenetic procedure mirrored that used in the operant box, except the session lasted only 30 min. Inhibition did not alter peak locomotion speed (Extended Data Fig. 7-1).
Discussion
We showed that manipulation of neural activity in the direct and indirect pathways in the NAc core had distinct effects on motivated behavior. Lever pressing on FR schedules was significantly reduced in mice receiving either dSPN (D1+) or iSPN (A2A+) activation in the NAc core (Fig. 3). However, there was a key difference between the effects of activating dSPN and iSPNs: dSPN activation reduced pressing to a similar extent regardless of the effort requirement, whereas iSPN activation resulted in a greater reduction in performance at higher ratios (Fig. 4). An important feature of our experimental design is that it allows us to dissociate general effects on performance due to changes in the motor system or appetite from more specific changes in effort exertion. Our results showed that the reduction in effort exertion during iSPN stimulation cannot be explained by motor impairments or changes in appetite, as reward port entry or consumption was not affected in a ratio-dependent fashion.
Although the activation of both the direct pathway and the indirect pathway in the NAc core reduced lever pressing, the behavioral changes are distinct. Direct pathway stimulation elicited a dramatic increase in gnawing (Fig. 5), and this effect accounts for reduced lever pressing, as gnawing behavior interferes with lever pressing. This finding suggests that direct pathway activation in the core can result in a switch to consummatory behaviors. Our findings also suggest that, in the NAc core at least, the contribution of the direct pathway is not the opposite of the indirect pathway. Although dSPN activation also reduced lever pressing, it did so independently of the effort requirement, producing similar reductions in performance regardless of the ratio requirement, whereas iSPN activation reduced lever pressing more at higher ratio (effort) requirements.
The NAc has long been implicated in oral stereotypy (Delfs and Kelley, 1990). Previous work found that DA agonists, including D1 selective agonists, injected into the NAc could increase gnawing behavior (Tirelli and Witkin, 1995; Geter-Douglass et al., 1997). However, it is possible that the direct pathway still plays a role in effort regulation but that the dSPNs responsible for this are found in different regions, which we did not manipulate in the present study. At any rate, dSPNs in the NAc core, where we stimulated, do not appear to play a role in regulating effort exertion as measured by lever pressing.
On the other hand, iSPN activation did not elicit gnawing behavior or otherwise impact motor function; rather, it suppressed lever pressing at high ratios. When the effort required is low, as in reward port entries or consumption, iSPN activation had a limited impact on performance. The ratio-dependent decrease in lever pressing caused by indirect pathway stimulation cannot be attributed to the corresponding increase in baseline pressing. The baseline press rate increased with schedule in both the D1 and A2A+ groups, but the stimulation-induced reduction in pressing was only ratio-dependent in the A2A+ group. Furthermore, the press reduction at FR-40 was significantly greater in the A2A+ group than in the D1+ group. This difference is best explained by the interpretation that iSPNs play a critical role in regulating effort exertion.
In addition, we found that the effect on effort exertion is restricted to iSPNs in the core. Stimulation of iSPNs in the neighboring shell region did not have a significant effect (Fig. 6). Previous work has shown that DA depletion in the core, rather than the shell, was responsible for regulating performance (Sokolowski and Salamone, 1998). Thus, our results are in agreement with these observations and suggest that the NAc shell is not a critical site for effort regulation.
The lack of any effect in the control groups that also received light stimulation suggests that our long-duration stimulation did not produce artifacts related to tissue heating. In fact, we chose our stimulation parameters specifically to avoid overheating. Across all experiments, laser power never exceeded 5 mW, and we chose low stimulation frequencies for our excitation experiments to limit light exposure as much as possible.
A previous study showed that activation of either D1- or D2-SPNs in the NAc is sufficient to increase operant performance in rats (Soares-Cunha et al., 2016), whereas we found that indirect pathway activation reduces rather than increases effort exertion. However, it should be noted that our experimental designs and methods differed significantly from those in the study by Soares-Cunha and colleagues. In particular, we did not use the D2 promoter to express opsins. Because D2 receptors are expressed highly in interneurons such as the cholinergic interneurons, they may not be the best markers for iSPNs. In contrast, the expression of A2A receptors is limited to D2+/A2A+ SPNs, so we used Cre-dependent vectors in A2A-Cre mice.
Previous work has also shown that optogenetic activation of either D1- or D2-SPNs in the NAc can reinforce behavior in the absence of any naturally rewarding stimulus (Cole et al., 2018; Soares-Cunha et al., 2020). However, these studies employed 40 Hz or continuous stimulation. When stimulation was delivered at 10 Hz, a frequency very similar to the ones used here, there was no reinforcing effect of D1- or D2-SPN activation.
The indirect pathway originating in the NAc core projects to the ventral pallidum. Previous work also showed that inhibition of iSPN terminals to the ventral pallidum is sufficient to enhance performance, in accord with our optogenetic inhibition results (Nowend et al., 2001). Whereas activation of these iSPNs in the NAc core reduces effort exertion, inhibition of iSPNs can have the opposite effect and increase lever pressing, especially at FR-40 (Fig. 7). While the increase in performance is more modest, this could be a ceiling effect, as the mice are already exerting considerable effort at higher FR schedules. Moreover, during high-effort tasks like FR-40, the iSPN output may already be so low that inhibition can only boost performance slightly. Electrophysiological measures of iSPNs will be needed to test this possibility.
Our findings are also in agreement with previous work in rats showing that pharmacological manipulations of iSPNs using A2A agonists could reduce effort exertion (Mingote et al., 2008). Since A2A agonists are expected to increase iSPN activity, the effect could be similar to that of the optogenetic excitation used in our experiments. On the other hand, A2A antagonists, which like optogenetic inhibition can reduce iSPN activity, increase effort exertion, as measured by lever presses and the highest ratio achieved on a progressive ratio task (Randall et al., 2012).
Compared to previous pharmacological manipulations, a major advantage of our study is the temporal and cell-population selectivity. We are therefore able to demonstrate rapid and reversible changes in lever pressing by manipulating iSPN activity. We are also able to demonstrate bidirectional regulation of effort exertion by showing the opposite effects of excitation and inhibition on the same behavioral measure. In addition, our finding that the NAc core, but not the neighboring shell region, is critical for effort regulation is also in accord with previous work implicating the NAc core but not the NAc shell as part of a neural circuit for effort regulation (Ghods-Sharifi and Floresco, 2010).
Our results are also in agreement with previous functional magnetic resonance imaging (fMRI) studies in humans (Botvinick et al., 2009). The magnitude of NAc activation was found to vary with both reward outcome and the effort demanded to obtain individual rewards. For a fixed level of reward, the NAc was less strongly activated following high-effort demand than following low demand. This effect is correlated with preceding activation in the dorsal anterior cingulate cortex, which is thought to contribute to effort regulation. While intriguing, fMRI studies cannot reveal the specific neuronal populations involved. Recent work has shown that the activation of anterior cingulate projections to the dorsomedial striatum can reduce effort exertion (Ulloa Severino et al., 2023). As the anterior cingulate also projects to the NAc core, it is possible that its projections to the iSPNs can moderate effort exertion as well. One possibility is that cortical representations of behavioral cost are used to activate the anterior cingulate-NAc core iSPN circuit and reduce effort exertion. This hypothesis remains to be tested.
Footnotes
This work was supported by NIDA (DA040701) and NINDS (NS094754).
↵*C.T.S. and A.D.F. are co-first authors.
The authors declare no competing financial interests.
- Correspondence should be addressed to Henry H. Yin at hy43{at}duke.edu.
