Abstract
When rats are given discrete choices between social interactions with a peer and opioid or psychostimulant drugs, they choose social interaction, even after extensive drug self-administration experience. Studies show that like drug and nondrug food reinforcers, social interaction is an operant reinforcer and induces dopamine release. However, these studies were conducted with same-sex peers. We examined if peer sex influences operant social interaction and the role of estrous cycle and striatal dopamine in same- versus opposite-sex social interaction. We trained male and female rats (n = 13 responders/12 peers) to lever-press (fixed-ratio 1 [FR1] schedule) for 15 s access to a same- or opposite-sex peer for 16 d (8 d/sex) while tracking females’ estrous cycle. Next, we transfected GRAB-DA2m and implanted optic fibers into nucleus accumbens (NAc) core and dorsomedial striatum (DMS). We then retrained the rats for 15 s social interaction (FR1 schedule) for 16 d (8 d/sex) and recorded striatal dopamine during operant responding for a peer for 8 d (4 d/sex). Finally, we assessed economic demand by manipulating FR requirements for a peer (10 d/sex). In male, but not female rats, operant responding was higher for the opposite-sex peer. Female's estrous cycle fluctuations had no effect on operant social interaction. Striatal dopamine signals for operant social interaction were dependent on the peer's sex and striatal region (NAc core vs DMS). Results indicate that estrous cycle fluctuations did not influence operant social interaction and that NAc core and DMS dopamine activity reflect sex-dependent features of volitional social interaction.
Significance Statement
Recent studies have shown that when rats are given a mutually exclusive choice between social interaction with a same-sex peer and opioid or psychostimulant drugs, they choose social interaction. In the present study, we examined if the peer sex influences operant social interaction and the role of the estrous cycle and striatal dopamine in same- versus opposite-sex social interaction. Responding was higher for the opposite-sex peer in male but not female rats, and estrous cycle had no effect on operant social interaction of either sex. Same-sex versus opposite-sex operant social interaction was associated with different dopamine responses in NAc core and DMS. Our study shows that peer sex influences operant responding for social interaction and associated neuronal responses.
Introduction
Preclinical studies on mechanisms of addiction have yet to result in new treatments (Heilig et al., 2016). To address this translational gap, we recently proposed a reverse-translational approach with the goal of developing animal models that mimic successful treatments, and use those models to identify new mechanisms and medications, and improve translation (Venniro et al., 2020). One such treatment is the community reinforcement approach which relies on operant principles by reinforcing abstinence with opportunities for social engagement like support groups and positive work environments (Azrin, 1976; Stitzer et al., 2011). Similar principles underlie our recently developed operant model of choice between drugs and rewarding social interaction in rats (Venniro et al., 2018). In this model, the opportunity to choose a social-reward eliminated or significantly decreased self-administration of opioid and psychostimulant drugs, even in rats that met “addiction” criteria (Venniro et al., 2018, 2021, 2022; Venniro and Shaham, 2020).
One limitation of these studies, and other studies on operant social interaction and choice between drugs and social interaction, is the exclusive use of same-sex peers (Achterberg et al., 2016; Vanderhooft et al., 2019; Hackenberg et al., 2021; Baldwin et al., 2022; Chow et al., 2022; Marcus et al., 2022; Schulingkamp et al., 2023; Smith et al., 2023). This limits translation of the social choice voluntary abstinence model because, in humans using drugs, positive social interactions occur with both same- and opposite-sex peers.
The effect of peer sex on operant social interaction in rats is largely unknown. Angermeier (1960) reported that the peer's sex had no effect on operant social interaction in male rats. However, the study did not include experiments where female rats were trained to lever-press for same- or opposite-sex peer. This study's data are also limited because a single male and female peer was used. Evans et al. (1994) reported that female rats lever-press for access to social interaction with castrated male rats. However, that study did not examine operant responding in male rats for female peers nor did it examine operant responding for same-sex peers.
Like drug and nondrug rewards, social interactions induce striatal dopamine release (Gunaydin et al., 2014; Dai et al., 2022; Solié et al., 2022). These and other studies have focused on dopamine release during innate (unconditioned) social interaction (Louilot et al., 1986; Gunaydin et al., 2014; Dai et al., 2022). One study examined dopamine activity during operant social interaction, but the social peers were unfamiliar same-sex juvenile mice (Solié et al., 2022). Thus, striatal dopamine dynamic during operant social interaction with same- versus opposite-sex familiar same-age peers are yet to be characterized.
We examined the effect of same- versus opposite-sex peer pairings on operant social interaction and economic demand for social interaction. We also examined the role of estrous cycle and striatal dopamine in operant responding for same- versus opposite-sex peers. Studies examining the estrous cycle and unconditioned (innate) social interaction have shown that female rats engage with male rats more during the proestrus/estrus phases of the cycle (Eliasson and Meyerson, 1975; Perez et al., 2019). Thus, we determined if the estrous cycle phase would affect operant responding for social interaction.
Striatal dopamine plays a critical role in drug self-administration and relapse-related behaviors in rat and monkey models (Wise, 2004; Pierce and Kumaresan, 2006; Nader et al., 2008; Bossert et al., 2013) and rewarding social interaction in rats (Manduca et al., 2016; Vanderschuren et al., 2016). To measure dopamine transients during operant social interaction, we used fiber photometry with the dopamine sensor GRABDA (Sun et al., 2018, 2020) and recorded dopamine from nucleus accumbens (NAc) core and dorsomedial striatum (DMS). We chose these regions because previous studies show that they play critical roles in reward learning, decision-making, and social behavior (Roesch et al., 2007; Vanderschuren et al., 2016; Langdon et al., 2018; Collins and Saunders, 2020).
Materials and Methods
Subjects
We used 12 male and 13 female Sprague Dawley rats (estimated age at arrival: ∼7–8 weeks PND; body weight at the time of arrival: males, 201–225 g; females, 176–200 g; Charles River). All the rats arrived pair-housed and were acclimated together for 1 week, after which we single-housed the rats for the rest of the experiment. We maintained the rats under a reverse 12:12 h light/dark cycle (lights off at 8 A.M.) with free access to food (Teklad Rat Diet, Envigo) and water in their home cage. We kept the rats separated for 2 weeks prior to the start of the experiment and handled them daily the week leading up to the experiment. We performed the experiment in accordance with the NIH Guide for the Care and Use of Laboratory under protocols approved by the NIDA IRP Animal Care and Use Committee.
Apparatus
We conducted the experiment in operant chambers with an attached partner chamber (ENV-00CT-SOCIAL; Med Associates). A guillotine door (ENV-10B2-SOC) and a perforated metal divider separated the two chambers, preventing the rats from crossing into the other chamber while allowing for full face-to-face and forepaw contact. We outfitted the front panel of the operant chamber with an illuminated nosepoke (ENV-114BM) with two retractable levers (ENV-112CM) mounted on either side. A white cue light (ENV-221M) was positioned above each lever, and at the top of the panel, a red house-light (ENV-221RD) was situated above the nosepoke port. Two Sonalert tones (2.9 kHz, ENV-223AM and 4.5 kHz, ENV-223HAM were positioned at the top corner on the back panel of the operant chamber, next to the guillotine door. The partner chamber contained no manipulanda or stimuli except the guillotine door which connected the chambers (see Fig. 1A). Med-PC IV controlled the operant chambers. To synchronize behavioral data with neural data, we used an independent control panel (SG-716B) that connected to a 28 V to TTL adapter (SG-231) connected to a FP3002 system (Neurophotometrics).
Operant chamber and experimental timeline. A, Social self-administration chambers. The main chamber contains all manipulandum, and the partner chamber holds the peer rat. Note that the main and partner chambers are separated by a guillotine door and a perforated metal gate. When the door is down (no access) the rats cannot engage in social interaction, whereas when it is up (access), the rats can engage with each other. B, Timeline of the experiment.
Experimental phases
We divided the rats into “responder rats” (n = 13; seven females) and “peer rats” (n = 12; six females). The responder rats were the experimental subjects and assigned to the operant chamber, granting them access to the manipulanda. Peer rats were the reinforcing stimulus and assigned to the partner chamber. We paired each responder rat with a male and female peer rat, and these specific rats served as the reinforcing stimuli for that responder for the entire experiment. Note, we assigned two responder rats, one of each sex, to the same male and female peers but we only conducted one pairing condition per training session. To also note, prior to surgery, we removed one female rat from the study due health problems. We replaced this rat with another female rat from the same cohort and trained her with the same peer as the removed rat. See Figure 1B for the experimental timeline.
Effect of estrous on operant responding for a same- or opposite-sex peer
We first trained the rats on a door shaping procedure which lasted 40 min. We allowed the rats to habituate to the chamber (i.e., responder rats to operant chamber and peer rats to partner chamber) for the first 5 min, followed by 25 min of a fixed-time (FT) 60 s schedule where the guillotine door lifted open for 15 s, allowing the rats to engage in partial-physical contact (i.e., social interaction) with a same- or opposite-sex peer (20 trials), and, for the remaining 10 min, we allowed the rats to habituate to chamber again.
Next, we trained the rats on a fixed ratio (FR) 1 schedule for social interaction for four sessions. A red houselight illuminated at the start of each session and remained on until the session concluded. During each trial, we signaled the start of the trial by extending a single lever (counterbalanced across subjects), where pressing the lever a single time resulted in the retraction of the lever and the lifting of the guillotine door for 15 s, allowing access for social interaction with a same- or opposite-sex peer. Each trial was separated by a 20 s intertrial interval (ITI), and each session lasted for 45 min. Next, we gave the rats a 2 d break in experimentation, at which point we switched the peer rat with another peer of the other sex. We then trained the responder rats on an FR1 for social interaction for four consecutive sessions with the new peer of the other sex on a different lever. We repeated this pattern of training until a total of 8 sessions with each peer sex was achieved (i.e., 16 sessions total). Peer-sex order was counterbalanced across subjects.
Approximately 90 min after each session, we performed vaginal lavage on all the female rats to determine which phase of the estrous cycle they were in. We collected the samples onto slides, stained with crystal violet, and imaged with a light microscope (McLean et al., 2012; Cora et al., 2015). We also held male responder- and peer-rats in similar positions as the female rats during this time to simulate similar levels of potential stress.
To determine the phase of the estrous cycle in female rats, we categorized each of the samples into proestrus, estrus, metestrus, or diestrus. The phase of the cycle was based on the relative proportion of leukocytes, cornified epithelial cells, and nucleated epithelial cells observed (Cora et al., 2015). We blinded experimenters to the samples. Two experimenters reached a consensus to determine the phase of each sample. In instances where there was no consensus, a third experimenter, blinded to prior results, categorized the samples (see Fig. 2C).
Operant responding for a same- versus opposite-sex peer and the role of estrous cycle. A, Number of social rewards earned by male (left) and female (right) rats when paired with a male or female peer. B, The last 3 d average of training for rewards earned (left) and the latency (in seconds) to emit a lever press (right). MM, male responder–male peer; MF, male responder–female peer; FM, female partner–male peer; FF, female partner–female peer. C, Representative cytology of cells across the estrous cycle. D, The effect of estrous cycle of the female peer on operant responding in male (left) and female (middle) rats, and the effect of estrous cycle of the female responder rat when paired with either a male or female peer (right). P is proestrus, E is estrus, M is metestrus, and D is diestrus. #Significant difference between male versus female responder; p < 0.05. *Post hoc significant difference between male responder with a male versus female peer, p < 0.05. n = 6 males, 7 females.
Characterization of striatal dopamine during operant responding for a same- or opposite-sex peer
Using the same cohort of rats, 3 d after the last FR1 schedule training session, we performed bilateral intracranial surgery for fiber photometry on responder rats. We unilaterally injected GRABDA (pAAV-hsyn-GRAB_DA2m from Addgene #140553; (Sun et al., 2020)) into the NAc core (AP +1.7 mm, ML + or −1.7 mm, and DV −6.3 and −6.2 mm from the brain surface) and the DMS (AP −0.4 mm, ML + or −2.6 mm, and DV −3.7 and −3.6 mm from the brain surface). The coordinates chosen for DMS, which are posterior to the NAc core, were based on prior studies (Stalnaker et al., 2012; Costa et al., 2023); the reason for targeting this more posterior subregion is that there is evidence for encoding of associative relationships in local neuronal activity in this region (Stalnaker et al., 2012). See Figure 2C for placements. We counterbalanced the injections into the NAc core and DMS across the hemispheres, injecting a total 1.4 μl of GRABDA virus at 0.7 μl per each DV site at 0.2 μl/min via an infusion pump. We then implanted optic fibers (200 μm diameter, 1.25 mm ferrule; Neurophotometrics) in each region in the corresponding location of the most dorsal (i.e., NAc core DV −6.2; DMS DV −3.6) viral infusion and secured the optic fibers and a custom 3D-printed protective headcap to the skull with dental cement and jeweler screws. We injected the rats with ketoprofen (2.5 mg/kg, s.c., Covetrus) after surgery and the following 3 d to relieve pain and inflammation; we also injected the rats for a week following surgery with gentamicin (5 mg/kg, s.c., APP Pharmaceuticals) to prevent infection.
After ∼2 weeks of recovery, we retrained the rats (n = 6 males and 6 females) with the same peers on the same FR1 schedule for 15 s of access for social interaction with a same- or opposite-sex peer. We ran the rats in the same manner as described above where rats lever-pressed for a peer of a given sex for four consecutive sessions, followed by a 2 d break, and then four consecutive sessions of training with the other sex peer. We ran the rats untethered for eight sessions (4 d/peer-pairing condition), followed by eight sessions (4 d/peer-pairing condition) tethered to habituate the rats to the photometry system, and, finally, for 8 sessions (4 d/peer-pairing condition) where we recorded dopamine transients during operant responding for social interaction. During the recording phase, we removed 1 male rat (n = 5 males and 6 females) from the study due to health problems. Peer-sex order was counterbalanced across subjects.
We recorded dopamine signals using a FP3002 system (Neurophotometrics) which was operated by Bonsai (Lopes et al., 2015 open-source program). We used a branching fiber-optic patch cord (200 μm diameter, 0.37 NA, Doric Lenses) that allowed us to connect both implanted-fibers in the rats simultaneously in up to 4 rats. We connected optic fibers together via shrink-wrapped protected brass sleeves (Thorlabs) and secured them using custom 3D-printed headcaps attached to a custom-made swivel-tether system that allowed for relatively free movement of the rats. We pulsed a 470 nm (active signal) and 415 nm (isobestic reference) light at 40 Hz (20 Hz acquisition rate for each channel for recordings). Excitation light intensity was measured at ∼45–60 µW at the tip of the patch cord. Thus, dopamine signals were recorded throughout the entirety of each session, which was capped at 45 min to avoid any issues with photobleaching.
We processed the signals using custom scripts in MATLAB (MathWorks) to characterize dopamine transients during operant responding for social interaction. We first applied a fourth-order median filter to the raw fluorescence signals from the 470 (active) and 415 (reference) channels. Next, we performed a second-order polynomial fit to the reference channel data, which was then subtracted from the active channel to correct for dopamine-independent variations in fluorescence. We z-scored the resulting signal for each trial, using the 2.5 s before each trial onset as a baseline. We also determined the area under the curve (total dopamine signal amplitude deviation from 0) and the maximum peak amplitude (highest dopamine signal amplitude deviating from 0) at lever extension, leading up to the lever press, and at lever press. We were able to determine dopamine signals for lever extension and lever press given there was enough time between the events (see Fig. 3A,B); for instances where there was overlap (<4 s) we excluded these data from analysis. For lever extension, we used the moment of lever extension to 0.5 s after as our analysis window. For time leading up to lever press, we used a 1 s window leading up to the lever press as our analysis window and at lever press, we used the moment of lever press to 0.5 s after as our analysis window.
Demand assessment of same- versus opposite-sex peer
Finally, 4 d after the last recording session, we assessed the rats on an economic demand procedure for a same- or opposite-sex peer. We trained the rats on FR1, 2, 4, 8, and 16 reinforcement schedules for 15 s of access for social interaction (two sessions per FR schedule) for 10 consecutive sessions with a given peer sex. Next, we gave the rats a 3 d break from experimentation before resuming training with the opposite peer sex for 10 consecutive sessions at each FR schedule for two sessions. The order of the FR was presented using a Latin square design; we used the same order across both peer-sex pairings for each rat. During this phase of the experiment, we also recorded dopamine transients for the different response requirements. However, due to a technical error we were unable able to synchronize the neural data reliably and accurately with the behavioral data. Thus, we only report the behavioral data of the economic demand assessment.
Statistical analysis
We analyzed the data using linear mixed-effects (LME) model in JMP (version17) where we set α to 0.05. In all behavioral analyses, we used subjects (the rats) as a random factor. We followed up on significant interactions, when allowed, using Tukey's HSD post hoc test. We describe the factors for each statistical analysis in the “Results” section. We only report significant effects critical for data interpretation and indicate results of post hoc analyses in the figures. We provide a full reporting of statistical results in Table 1. For all correlations described in Tables 2 and 3, we used Spearman's ρ; for all correlations described in Tables 4 and 5, we used Pearson's r. In all neural analyses, we used trials (the trials across all subjects for all sessions) as a random factor. Economic demand assessment was analyzed in R using nonlinear mixed effects (NLME) modeling from the “nlme” package (Pinheiro et al., 2017). We applied the exponentiated form of the demand equation (Koffarnus et al., 2015):
Correlations between latency (in seconds) to lever press and area under the curve (AUC)
Correlations between latency (in seconds) to lever press and maximum peak amplitude
Correlations between Q0 and area under the curve (AUC) and maximum peak amplitude
Correlations between α and area under the curve (AUC) and maximum peak amplitude
Results
Effect of peer sex and estrous cycle on operant responding for access to social interaction
Peer sex
Both male and female rats learned to lever press for 15 s access for a same- or opposite-sex peer for social interaction (Fig. 2A). We first analyzed the data with responder sex (nominal; male vs female) as a between-subject factor and peer sex (nominal; male vs female) and session (continuous; 1–8) as within-subjects factors. This analysis showed significant effects of sex [F(1,11) = 5.6, p = 0.038] and session [F(1,11) = 7.8, p = 0.018] but not peer sex (p = 0.13). Since the training cycled between same- versus opposite-sex peers, we analyzed the data of the mean of the last 3 d of training with each peer sex (Fig. 2B) to determine if statistically significant sex differences would emerge after acquisition of the operant response (“maintenance phase”). We used responder sex as the between-subject factor and peer sex as the within-subject factor and the analysis showed significant effects of peer sex [F(1,11) = 6.0, p = 0.033] and responder sex × peer sex [F(1,11) = 5.9, p = 0.034]. Post hoc analysis showed that male rats paired with a female peer lever pressed more for access when compared to male–male, female–male, and female–female peer pairings. Using the same model, the analysis of latency (in seconds) to press, showed a significant effect of peer sex [F(1,11) = 5.5, p = 0.039].
Estrous cycle
To determine if the role of estrous cycle affects operant responding for social interaction in male and female rats, we first analyzed rewards earned based on the female peer's phase (Fig. 2D). We analyzed the data with responder sex as the between-subject factor and peer phase (nominal; proestrus, estrus, metestrus, and diestrus) as the within-subject factor; the analysis showed a significant effect of responder sex [F(1,11.31) = 6.9, p = 0.023]. When we analyzed the effect of estrous cycle of the female responder on operant responding for a male or female peer with responder phase and peer sex as the within-subject factors, the analysis showed no significant effects (p-values >0.1).
Together, these results show that male rats completed more trials for access to a peer than female rats. Additionally, once the rats were trained, the male rats paired with a female peer completed more trials for access to a peer than when paired with a male peer. Finally, unexpectedly, our data indicate that the estrous cycle has no significant effect on operant responding for a peer, regardless of the peer's sex or responder's estrous phase.
Characterization of striatal dopamine during operant responding for same- or opposite-sex peer
Behavioral data during retraining
Both male and female rats continued to lever-press for a same- or opposite-sex peer after receiving bilateral GRABDA injections and fiber optic implants (Fig. 3A,B). We analyzed the retraining data prior to recording with responder sex (nominal; male vs female) as a between-subject factor and peer sex (nominal; male vs female) and session (continuous; 1–8) as within-subjects factors. This analysis showed significant effects of peer sex [F(1,8.9) = 6.5, p = 0.031], session [F(1,9.4) = 46.5, p < 0.0001], and responder sex × peer sex [F(1,8.9) = 8.2, p = 0.019]. Post hoc analysis showed that male rats paired with a female peer completed more trials for access when paired with a male peer.
Operant responding for a same- versus opposite-sex peer and characterization of striatal dopamine responses. The number of social rewards earned during self-administration by male (A) and female (B) rats when paired with a male or female peer following GRABDA injections and optic fiber implants. BL represents baseline prior to surgery (3 d mean), and the dotted line past session 8 represents start of recorded sessions. C, Representative section and optic fiber placements in the nucleus accumbens core (NAc core) and dorsal medial striatum (DMS); green and orange dots reflect male and female rats, respectively. E, Representative schematic of lever extension (left) and lever press (right). Dopamine responses at lever extension and leading up to lever press in the NAc core (E,F) and DMS (G,H) in male and female rats. n = 5 males, 6 females. The dotted line represents the moment of lever extension or lever press. Calculated AUC (I–L) and max peak amplitude (M–P) measures for all recorded trials in NAc core and DMS at lever extension and leading up to lever press. MM, male responder-male peer; MF, male responder-female peer; FM, female partner-male peer; FF, female partner-female peer. #Significant difference between male versus female responder, p < 0.05. *Post hoc significant difference between male responder with a male versus female peer, p < 0.05. n = 296–505 trials per condition. See Figure 4 for corresponding latency (in seconds) to emit a lever press and dopamine responses at lever press in the NAc core and DMS.
Using the same model, we also analyzed latency to lever press (Fig. 4A,B). The analysis showed significant effects of peer sex [F(1,9.3) = 8, p = 0.017], session [F(1,10.3) = 14.6, p = 0.003], and responder sex × peer sex [F(1,9.3) = 6.4, p = 0.032]. Post hoc analysis showed that for male rats the latency to lever press for a female peer was shorter than for a male peer.
Latency (s) to press and characterization of striatal dopamine at lever press. The latency to press for social interaction during self-administration by male (A) and female (B) rats when paired with a male or female peer following GRABDA injections and optic fiber implants. BL represents baseline prior to surgery (3 d mean), and the dotted line past session 8 represents start of recorded sessions. Calculated AUC (C,D) and max peak amplitude (E,F) measures for all recorded trials in NAc core and DMS at lever press. MM, male responder–male peer; MF, male responder–female peer; FM, female partner–male peer; FF, female partner-female peer. n = 296–505 trials per condition.
Behavioral data during recording
Rats paired with a female peer completed more trials for social interaction relative to a male peer. We analyzed the behavioral data during fiber photometry recording with responder sex (nominal; male vs female) as a between-subject factor and peer sex (nominal; male vs female) and session (continuous; 1–4) as within-subjects factors. The analysis showed significant effects of peer sex [F(1,9) = 7.2, p = 0.025] and session [F(1,9) = 13.0, p = 0.006]. The latency analysis, which used the same model, showed a significant effect of session [F(1,9) = 7.1, p = 0.026].
Fiber photometry data
We examined dopamine transients in the NAc core and DMS at lever extension and the time leading up to lever press (Fig. 3E–H). We note that the male and female peers were assigned to different levers (e.g., right lever = male-peer, left lever = female peer; counterbalanced across subjects), which, combined with the additional counterbalancing of the optic fiber implant locations, was designed to avoid any lateralization effects of the recorded signals, especially in DMS (Parker et al., 2016). To analyze dopamine transients associated with a male or female peer, we determined the area under the curve (AUC, Fig. 3I–L) and the maximum peak (max peak, Fig. 3M–P) for each trial and collapsed the data across sessions. We analyzed the neural data with responder sex (nominal; male vs female) as a between-subject factor and peer sex (nominal; male vs female) as a within-subject factor. Note that Figures 3I–P and 4C–F are derived from Figure 3E–H.
Lever extension
AUC
Independent of the peer sex, in both the NAc core and DMS, the male rats had higher AUC values than the female rats (Fig. 3I,K). The NAc core analysis showed significant effects of responder sex [F(1,79.2) = 6.5, p = 0.013] and responder sex × peer sex interaction [F(1,66.7) = 8.14, p = 0.006]. In the DMS, the analysis showed significant effects of responder sex [F(1,82.9) = 10.4, p = 0.002] and peer sex [F(1,66.4) = 5.9, p = 0.018].
Max peak. Independent of the peer sex, in both the NAc core and DMS, the males had higher max peak values than the females (Fig. 3J,L). In the NAc core, the males had higher max peak values for a lever extending that was associated with a female peer. The NAc core analysis showed significant effects of responder sex [F(1,82.2) = 5.9, p = 0.017], peer sex [F(1,47.3) = 10.1, p = 0.026], and responder sex × peer sex interaction [F(1,47.3) = 36.5, p < 0.001]. The DMS analysis also showed a significant effect of responder sex [F(1,81.7) = 8.5, p = 0.005].
Pre-press
AUC
During the 1 s period leading up to the lever press, there were no significant main or interaction effects in the analysis of either NAc core or DMS in male and female rats (Fig. 3M,O).
Max peak. During the 1 s period leading up to the lever press, there were no significant main or interaction effects in the analysis of either NAc core or DMS in male and female rats (Fig. 3N,P).
Lever press
AUC
During the moment of press to 0.5 s after, females, but not males, showed higher dopamine signal AUC values in the NAc core at lever press for a female peer (Fig. 4C). Independent of responder sex, in the DMS, AUC values were higher at lever press for a female peer (Fig. 4D). The NAc core analysis showed significant effects of responder sex [F(1,63.4) = 29.9, p < 0.001], peer sex [F(1,62.4) = 6.0, p = 0.017), and responder sex × peer sex [F(1,62.4) = 8.5, p = 0.005]. The DMS analysis showed a significant effect of peer sex [F(1,54.2) = 7.5, p = 0.008].
Max peak. In the NAc core, and to a lesser degree in the DMS, females had higher max peak values at lever press with a female peer (Fig. 4E,F). The NAc core analysis showed significant effects of responder sex [F(1,68) = 9.3, p = 0.003], peer sex [F(1,67.1) = 7.8, p = 0.007], and responder sex × peer sex [F(1,67.1) = 6.5, p = 0.013]. The DMS analysis showed significant effects of responder sex [F(1,56.2) = 15.6, p < 0.001] and responder sex × peer sex [F(1,53.0) = 10.5, p = 0.002].
In conclusion, at lever extension (moment of extension to 0.5 s after), males showed higher AUC and max peak values in both the NAc core and DMS. In addition, males showed higher AUC and max peak values in NAc core when lever presses were rewarded with a female peer. There were no differences in either AUC or max peak in both the NAc core and DMS leading up to the lever press (pre-press, 1 s period leading up to lever press). In contrast, at lever press (moment of press to 0.5 s after), females showed higher AUC and max peak values than males in both the NAc core and DMS. In addition, females showed higher AUC and max peak values in NAc core at lever press when paired with a female peer.
Time-based analysis of activity during the recording sessions
During the initial analysis of the neural data, we observed a change in transient patterns across the trials where we primarily observed correlations between latency (in seconds) to press and AUC and max peak values for lever extension. See Tables 2 and 3 for all correlations. Thus, we ran additional analyses examining the associated neural activity of recorded trials completed during the first, middle, and last 15 min of each (45 min) session (Fig. 5A–L). We analyzed the AUC (Fig. 5M–P) and max peak (Fig. 5Q–T) across the 15 min bins with responder sex as the between-subject factor, peer sex, and time-bin (nominal; first, middle, last) as the within-subject factors. Note: Figures 5M–Q and 6A–C are derived from Figure 5A–L.
Striatal dopamine responses, area under curve, and max peak across 15 min bins. Striatal dopamine responses for lever extension and lever press during the first (A–D), middle (E–H), and last (I–L) 15 min of each session in the NAc core and DMS for male and female rats. The dotted line represents the moment of lever extension or lever press. Calculated AUC (M–P) and maximum peak amplitude (Q–T) as a function of time-bins in NAc core and DMS at lever extension and leading up to lever press. MM, male responder–male peer; MF, male responder–female peer; FM, female partner–male peer; FF, female partner–female peer. #Significant difference between male versus female responder, p < 0.05. *Post hoc significant difference between male responder with a male versus female peer, p < 0.05. n = 78–205 trials per condition. See Figure 6 for corresponding dopamine responses across the 15 min bins at lever press in the NAc core and DMS.
Area under curve and max peak across 15 min bins at lever press. Calculated AUC (A,B) and maximum peak amplitude (C,D) as a function of time-bins in NAc core and DMS at lever press. MM, male responder–male peer; MF, male responder–female peer; FM, female partner–male peer; FF, female partner–female peer. n = 78–205 trials per condition.
Lever extension
AUC
Like the prior analysis, independent of the peer sex, in both the NAc core and DMS, the males had higher AUC values than the females (Fig. 5M,O). Additionally, in the NAc core, the males had higher AUC values for a lever extending that was associated with a female peer. However, in both the NAc core and DMS, the AUC values were higher during the first 15 min bin than in the 30 and 45 min bins. The NAc core analysis showed significant effects of responder sex [F(1,33.2) = 11.9, p = 0.002], time-bin [F(2,125.3) = 25.0, p < 0.0001], and responder sex × peer sex interaction [F(1,98.2) = 7.0, p = 0.009]. The DMS analysis showed significant effects of responder sex [F(1,31.2) = 9.7, p = 0.004], peer sex [F(1,72.3) = 4.5, p = 0.037], time-bin [F(2,131) = 11.8, p < 0.0001], and peer sex × time-bin interaction [F(2,135.9) = 3.4, p = 0.036].
Max peak. Independent of the peer sex, in both the NAc core and DMS, the males had higher max peak values than the females (Fig. 5Q,S). In the NAc core, males had higher max peak values for a lever extending that was associated with a female peer. These effects were largely selective to the first 15 min of the sessions. The NAc core analysis showed significant effects of responder sex [F(1,27.1) = 13.4, p = 0.001], time-bin [F(2,169.4) = 29.1, p < 0.001], and responder sex × peer sex × time-bin [F(2,40.1) = 3.3, p = 0.048]. The DMS analysis showed significant effects of responder sex [F(1,23.2) = 8.1, p = 0.009] and time-bin [F(2,120.5) = 14.7, p < 0.001].
Pre-press
AUC
There was a small decrease in AUC measures leading up to the lever press in both the NAc core and DMS across the session (Fig. 5N,P). The NAc core analysis showed a main effect of time-bin [F(2,101) = 9.3, p = 0.0002]. The DMS analysis also showed a main effect of time-bin [F(2,108) = 8.7, p = 0.0003].
Max peak. There was a small decrease in max peak measures leading up to the lever press in both the NAc core and DMS across the session (Fig. 5R,T). The NAc core analysis showed a main effect of time-bin [F(2,91) = 5.2, p = 0.0072]. The DMS analysis also showed a main effect of time-bin [F(2,107) = 10.3, p < 0.0001].
Lever press
AUC
In the NAc core but not DMS, females but not males showed higher AUC values when lever presses were rewarded with a female peer (Fig. 6A). Independent of sex, in the DMS, AUC values were higher at lever press for a female peer (Fig. 6C). The NAc core analysis showed significant effects of responder sex [F(1,63.5) = 24.7, p < 0.001], peer sex [F(1,79.9) = 4.1, p = 0.047], and responder sex × peer sex [F(1,79.9) = 7.1, p = 0.009]. The DMS analysis also showed significant effects of peer sex [F(1,70.1) = 8.1, p = 0.006], and responder sex × peer sex × time-bin interaction [F(2,93.5) = 4.7, p = 0.011]; The interaction effect is due to the fact that the AUC value for males paired with a female peer during the last 15 min interval was significantly higher than that for females paired with a male peer during the first 15 min bin.
Max peak. The NAc core and DMS analysis showed main effects of time-bin [F(2,91) = 5.2, p = 0.0072] and [F(2,107) = 10.3, p < 0.0001]. In both the NAc core and DMS, the females had higher max peak values when lever presses were rewarded with a female peer (Fig. 6B,D). These effects were independent on the session time-bin. The NAc core analysis showed significant effects of responder sex [F(1,70.1) = 6.5, p = 0.013], peer sex [F(1,79.2) = 5.8, p = 0.018], and responder sex × peer sex [F(1,79.2) = 5.2, p = 0.026]. The DMS analysis also showed significant effects of responder sex [F(1,48.0) = 9.3, p = 0.004], and responder sex × peer [F(1,60.3) = 8.5, p = 0.005].
Together, the time course analyses show a similar pattern of sex-dependent dopamine response in NAc core and DMS at lever extension, leading up to lever press, and at lever press associated with same- or opposite-sex peers. These analyses also showed that the dopamine response to the lever extending and leading up to the lever press decreased within-session, with maximal responding during the first 15 min of the session. This time-dependent effect was not observed for lever press.
Demand assessment of same- versus opposite-sex peer
Both male and female rats lever pressed for access to a social peer at different FR schedules and responding for a social peer decreased with increased FR response requirements (Fig. 7). We fitted the exponentiated demand equation to the number rewards and analyzed the data using nonliner mixed effect model with responder sex as the between-subject factor, and peer sex and FR schedule (continuous; 1, 2, 4, 8, 16) as the within-subject factors.
Economic assessment of FR schedule changes for same- versus opposite-sex peer. Log–log plots of the number rewards earned at the different FR schedules in male (A) and female (B) rats for a same- or opposite-sex peer. Lines are best fits from the exponentiated demand equation. *Post hoc significant differences between male responder for a male versus female peer Q0, p < 0.05. n = 5 males; 6 females.
The analysis of the demand data showed no significant main or interaction effects (p-values >0.05, Table 1) for Q0 (free consumption measure). However, when we analyzed the data separately for males and females, there was a significant effect of peer sex for Q0 (higher Q0 for a female peer) for males [F(1,40) = 4.9, p = 0.03] but not females (p > 0.05). The analysis also showed a significant responder sex × peer sex interaction [F(1,86) = 11.2, p = 0.0012] for α (change in consumption as a function of unit price), reflecting lower demand elasticity for social interaction in males that lever press for a female peer.
Since we were unable to synchronize the behavioral and neural data reliably and accurately during the economic demand sessions, we correlated individual Q0 and α values with individual AUC and max peak measures (averaged across the recording sessions before the behavioral economic assessment) to determine if striatal dopamine activity predicts economic demand measures. The analysis mostly showed no significant correlations across all comparisons except for two instances (Tables 4, 5). In general, these results suggest that prior striatal dopamine activity in the NAc core and DMS did not predict economic demand.
Together, the economic demand data indicate that male rats completed more trials for social interaction when paired with a female peer than with a male peer when the price of social interaction was “free.” The analysis also showed that males had lower α value when paired with a female peer, indicating that their operant responding for a female peer decreased at a slower rate than for a male peer as the unit price increased (decreased demand elasticity), suggesting higher motivation to lever press for a female peer.
Discussion
We studied if peer sex influences operant social interaction and the role of estrous cycle and striatal dopamine in same- versus opposite-sex social interaction in rats. We report three main findings. First, in males but not females, operant responding was higher for the opposite-sex peer. Behavioral economic analysis also suggests decreased demand elasticity in males but not females who lever-press for social interaction with an opposite sex. Second, independent of the peer sex, the estrous cycle had no significant effect on operant responding in male or female rats. Third, our data suggest that dopamine release in the NAc core and DMS, as determined via GRABDA signal recordings with fiber photometry, reflect features of operant social interaction that are sex- and region-dependent. The main findings are that at lever extension (i.e., a cue that indicates the possibility to make an action to obtain social interaction), males showed more dopamine release in NAc core in response to a lever that signals the availability of a female peer. Dopamine activity leading up to a lever press (pre-press, the period leading up to decision to make an action) was similar for both sexes. In contrast, at lever press (i.e., the execution of the action to gain access to a peer), females showed greater dopamine release in NAc core when lever presses were rewarded with a female peer.
Effect of peer sex on operant responding for social interaction
One goal of our study was to determine if a same- or opposite-sex peer would influence operant social interaction. This question is of interest because prior studies on operant social interaction only used same-sex peers (e.g., Baldwin et al., 2022; Chow et al., 2022; Venniro et al., 2022; Smith et al., 2023). We assessed rate of responding during acquisition and maintenance of operant social interaction under an FR1 reinforcement schedule and also used an economic demand assessment of social interaction (Vanderhooft et al., 2019; Schulingkamp et al., 2023). The main finding from the FR1 schedule assessment is that operant responding was higher for the opposite-sex peer in males but not in females (Fig. 2A). There are two main findings in the economic demand assessment. The first is that in males, but not females, the Q0 (consumption of “free” reward) is higher for the opposite peer condition, confirming the FR1 schedule assessment. The second is a significant responder sex × peer sex interaction for α (rate of change as a function of unit price), because the males decrease their responding for a female peer at a slower rate, suggesting that a female peer for male rats serves as a less elastic reward (Fig. 7A).
Together, these data indicate that for males, a female peer is a more effective reinforcer while for females both male and female peers have similar reinforcing effects. A question for future research is whether this conclusion will persist in choice procedures where rats can choose between a same- versus opposite-sex peer under other experimental conditions (see Hackenberg et al., 2021).
Effect of estrous cycle on operant responding for social self-administration
Another goal of our study was to determine the effect of the female's estrous cycle on operant social interaction. Previous studies have shown that female rats interact more often with male rats during the proestrus/estrus phases of the cycle, which coincides with the time when female rats are receptive to sexual behavior (Eliasson and Meyerson, 1975; Perez et al., 2019). Thus, if operant responding for social interaction for an opposite-sex peer is driven by sexual motivation, there should be increased responding during those phases when males lever press for access to females and females lever press for access to males. Our results did not show any effect of the estrous cycle phase on operant social interaction. However, these results should be interpreted with caution because social interaction in our study was done through a perforated screen that prevented full social contact, including mounting. Additionally, our subjects were sexually naive, and there is evidence that sexual experience enhances the rewarding effects of opposite-sex social interaction (Tenk et al., 2009). Consistent with such complexity, the literature on the role of estrous cycle with other reinforcers such as food and addictive drugs is also mixed (Cifani et al., 2012; Richard et al., 2017; Becker and Chartoff, 2019; Fredriksson et al., 2021; Nicolas et al., 2022).
Characterization of dopamine transients during operant responding for social interaction
The final goal of our study was to characterize striatal dopaminergic signaling during operant social interaction. Our data suggest that dopamine activity during operant social interaction (lever extension vs lever press) is dependent on sex and striatal region. At lever extension, males showed higher dopamine responses in both the NAc core and DMS. More important, males showed higher responses in NAc core during the extension of a lever that was associated with a female peer, suggesting that NAc core dopamine activity reflects higher reinforcing effects of opposite-sex interaction that we observed in the behavioral results. There were no differences in dopamine responses leading up to the lever press (pre-press). In contrast, at lever press, females showed higher dopamine responses in both the NAc core and DMS. Furthermore, females showed higher dopamine responses in NAc core when lever presses were rewarded with a female peer. This effect does not have a clear correlate in social self-administration in females: in both the FR1 schedule and behavioral economic assessment there were no significant differences in operant responding for a male or female peer. Finally, we observed a time-dependent dissociation in dopamine activity during different events (lever extension vs lever press) of operant social interaction, where dopamine activity at lever extension and leading up to lever press decreased across the session, while at lever press, dopamine activity stayed relatively constant throughout.
Our data on the differential dopamine activity in DMS and NAc core in response to appetitive rewards agree with results from previous studies using male rats (Roitman et al., 2004; Gan et al., 2010; Saddoris et al., 2015). It also agrees with recording studies that have targeted the NAc core in rodents which showed that dopamine levels increase during social interaction (Robinson et al., 2011; Gunaydin et al., 2014; Dai et al., 2022). Our data extends beyond these findings by showing that this putative value representation in these regions is also seen in response to cues associated with the opportunity to engage in appetitive social interaction. Furthermore, like the correlates of prediction errors of nonsocial rewards, the higher dopamine responses to cues that predict female peer interactions in male rats agrees with previous work showing that NAc core dopamine release is higher in males interacting with females in comparison to a nonaggressive male (Louilot et al., 1986), and that dopamine release in NAc core determines sexual preference for females characteristic in male mice (Beny-Shefer et al., 2017). Our results also agree with the study by Solié et al. (2022), which proposed that dopamine neuron firing reflects a “social” prediction error signal. While our study did not examine errors in social expectations, it did confirm that dopamine release in both NAc core and DMS reflect social value prediction errors in response to cues associated with different forms (i.e., sex of the peer) of social interaction.
Our study extends results from previous studies on the important role of dopamine in different striatal regions in social reward in rats as assessed by the conditioned place preference procedure (Manduca et al., 2016; Vanderschuren et al., 2016). Finally, our data also mirror behavioral pharmacology findings showing that in male adolescent rats, increasing extracellular dopamine levels by a dopamine transporter reuptake inhibitor increases lever pressing for access to a same-sex peer, while blockade of dopamine receptors has an opposite effect (Achterberg et al., 2014, 2016).
One final consideration is that we previously showed that when rats are single-housed they will lever-press more for social interaction for a same-sex peer than when pair-housed (Chow et al., 2022). Prior studies have shown that housing condition alters dopaminergic function in rats. For example, long-term social isolation caused increased D1 dopamine receptor density (Gill et al., 2013) and basal dopamine levels in NAc (Miura et al., 2002), and decreased dopamine spine densities (Wang et al., 2012). Thus, a question for future research is whether the pattern of striatal dopamine response observed in our study generalizes to socially housed rats.
Concluding remarks
Our study on the role of same- versus opposite sex on operant social self-administration was inspired by our earlier studies showing that rats given mutually exclusive choice between operant social interaction and access to a same-sex peer significantly decrease and often completely abstain from methamphetamine, cocaine, and heroin self-administration (Venniro et al., 2018, 2019, 2021, 2022). Here, we showed that the peer sex had a selective effect on operant responding for social interaction in males, and, in sexually naive rats, the estrous cycle had no significant effect on operant social interaction in either males or females. Finally, our initial investigation of neuronal correlates suggest that striatal dopamine activity play different roles in operant social self-administration in males and females.
Footnotes
This work was supported by the Intramural Research Program of NIDA (ZIA-DA000434-17, Y.S.) and the NIH CCB Fellowship (J.J.C.) and NIH NIGMS PRAT Fellowship (FI2GM142476, J.J.C.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Jonathan J. Chow at jonathan.chow{at}nih.gov or Yavin Shaham at yavin.shaham{at}nih.gov.
