Distinct dynamics of social motivation drive differential social behavior in laboratory rat and mouse strains

The study investigates whether commonly used laboratory mice and rats differ in their social investigation behavior, focusing on C57BL/6J mice versus Sprague Dawley (SD) rats. Social interactions in mammals are complex and dynamic, requiring methods that capture temporal behavior. Although mice have become the dominant genetic model, advances in rat genetics renew the need to understand inter-strain and inter-species behavioral differences. Prior recognition of differences between rats and mice has lacked direct quantitative comparisons. The authors hypothesize that the dynamics of social motivation differ between strains, leading to distinct patterns in social preference and novelty tests, and aim to quantify these dynamics and relate them to neural activation and motivational competition.

The paper situates the work within prior literature showing: (1) rodents as core models for social behavior and neuropathologies; (2) technological shifts enabling rat genetics; (3) acknowledged but rarely quantified species differences in social behavior; (4) previous development of automated systems to track social preference dynamics in small rodents; (5) known roles of brain regions like medial amygdala (MeA), nucleus accumbens (NAc), lateral septum (LS), and ventral tegmental area (VTA) in social motivation and reward; (6) reports that social isolation and sex of stimuli modulate c-Fos induction in mice; and (7) evidence from conditioned place preference that SD rats may find social interaction more rewarding than C57BL/6J mice. The authors note gaps: few direct, quantitative strain/species comparisons across matched assays, and almost no measurement of stimulus behavior during social tests.

Subjects: Adult male (and some female) C57BL/6J and BALB/c and ICR (CD-1) mice, and Sprague Dawley (SD) and Wistar Hannover rats (10–15 weeks). Stimuli were juvenile conspecifics (21–30 days), except sex-preference tests which used adult male/female. Group-housed, standard conditions; tests during dark phase under dim red light; IACUC-approved. Behavioral assays and setups: An automated, previously described system tracked investigation during social preference (SP) and social novelty preference (SNP) tests, adapted for species size. Protocol: 20-min arena habituation; 15-min habituation with empty chambers; 5-min SP test (social stimulus vs object); 15-min inter-test interval; 5-min SNP test (novel vs familiar social stimulus). Sex preference (SxP): 15-min habituation then 5-min exposure to adult male vs female. Social vs food competition: 15-min habituation then 5-min exposure to social stimulus in one chamber and inaccessible food pellets in the opposite chamber; same animals tested across deprivation levels (mice: satiety, 4 h, 24 h; rats: satiety, 24 h, 48 h). Free interaction assay: 5-min dyadic male–male interactions with juveniles, tracked by custom MATLAB algorithm. Behavioral measures: Investigation time binned (20 s); investigation bouts corrected for <0.5 s gaps and categorized as short (<6 s), intermediate (6–19 s), long (>19 s); transitions defined as onset of investigation of the alternate stimulus (20-s bin rates); relative differential investigation (RDI) = |social−other|/(social+other); center/periphery ratio from 20-min habituation for anxiety proxy. Stimulus movement monitoring: Triangular chamber floor equipped with six 27-mm piezoelectric disks in parallel; signals acquired (20 kHz, downsampled 2 kHz; 10–100 Hz band-pass). Major movements defined as >20% of max absolute signal. Analysis aligned subject investigation relative to stimulus movement, comparing investigation following major movement vs without major movement in the preceding 3.5 s. c-Fos immunohistochemistry: After 15-min habituation, exposure for 2 min to: (i) two empty chambers (control), (ii) a social stimulus in a chamber (social chamber), or (iii) freely moving social stimulus (free interaction). Ninety minutes later, brains processed for c-Fos (50-µm coronal sections), DAB labeling, manual counts in MeA, NAc, LSD, LSV (and later VTA); in mice, a separate cohort compared 2-min versus 5-min SP for NAc core/shell and MeA. Computational modeling: Discrete-time Markov model with four states: investigate stimulus 1 (S1), stimulus 2 (S2), stillness (S3), exploration (S4). Motivational parameters: time-varying reward for each stimulus, stress, and derived anxiety (function of stress and reward difference). Transitions constrained to pass through stillness before switching motivational states. Parameters fit via evolutionary multi-objective optimization to mouse SP data; SNP and rat simulations adjusted only by initial reward values/dynamics. Statistics: Normality by Kolmogorov–Smirnov/Shapiro–Wilk; t tests, Wilcoxon/Mann–Whitney as needed; (repeated-measures/mixed-model) ANOVAs/Welch’s ANOVAs with post hoc tests (Student’s t or Games–Howell). Alpha 0.05. Source data and code provided via supplementary and GitHub.

- SD rats display stronger and earlier social preference than C57BL/6J mice in the SP test. SD rats begin with intensive interaction with the social stimulus and few early transitions; mice show high early transitions and short bouts, with long social bouts increasing later. RDI significantly higher in SD rats than C57BL/6J mice (t=23.619, df=114, p<0.0001). - The difference concentrates in the first minute: SD rats have longer bout durations and lower transition rates than mice only in minute 1; mice have higher early transition rates (strain×time interaction for transitions: F≈6.704, p=0.010). - Anxiety proxy does not account for differences: mice had higher center/periphery ratios than rats (0.22±0.22 vs 0.08±0.05; t=3.251, df=71, p<0.001), indicating lower baseline anxiety. - Strain specificity: Outbred ICR mice and Wistar Hannover rats exhibited SP dynamics more similar to SD rats than to C57BL/6J mice (higher social preference, low early transitions, long bouts peaking early), but quantitatively intermediate (RDI Welch’s ANOVA F3,57.745=26.112, p<0.001; first-minute transition rate Welch’s ANOVA F3,58=22.071, p<0.001; first-minute bout duration Welch’s ANOVA F3,49.87=24.106, p=0.011). BALB/c mice resembled ICR, differing from C57BL/6J. - SNP dynamics: C57BL/6J mice showed SNP with similar dynamics to SP (early exploration then increased interaction). SD rats required extended SP (15 min) to show clear SNP; then SNP dynamics resembled mice (high early transitions, long bouts toward novel stimulus without significant time increase). - Stimulus movement effects: Piezo measurements showed similar stimulus movement rates early in tests, but opposite subject responses. In mice, social investigation decreased immediately following major stimulus movements (stimuli×movement interaction F≈4.26, p=0.04); object investigation unchanged. In rats, social investigation increased and object investigation decreased following major movements (stimuli×movement F≈7.008, p=0.014). Thus, subject behavior primarily drives strain differences. - Free interactions: SD rats showed greater total interaction time than C57BL/6J mice (t=3.909, df=24, p=0.001), driven by more long (>6 s) interactions (time×strain F1,24=6.444, p=0.018). - c-Fos induction: After 2-min SP exposure, SD rats showed significant c-Fos increases in MeA (F2,10=13.46, p=0.0015) and NAc (F2,10=5.62, p=0.0231), but not LSD/LSV; mice showed no significant induction in any region. In mice, extending SP to 5 min increased c-Fos in MeA (F2,12=7.161, p=0.009) and NAc shell (F2,12=3.767, p=0.057 trend), but not NAc core, LSD, LSV. VTA analyses in rats mirrored MeA/NAc findings. - Social vs food competition: Mice showed no social-over-food preference at satiety (t=−1.454, df=15, p=0.167) and preferred food after 4 h (t=3.351, df=15, p=0.004) and 24 h (t=4.118, df=15, p=0.0001) deprivation. Rats preferred social at satiety (t=−8.448, df=7, p<0.001) and did not prefer food over social at 24 h (t=1.799, df=7, p=0.115), but preferred food after 48 h (t=4.331, df=7, p=0.003). Under no overall preference conditions, mice started biased to food, rats to social, diverging significantly in minute 1 (time×strain F4,88=5.805, p<0.0001; minute-1 t=−4.842, df=22, p<0.001). Transition rates shifted with deprivation in opposite directions for mice (decrease; p=0.001) and rats (increase; p=0.034). - Computational model: A simple four-state Markov model, parameterized on mouse SP data, reproduced mouse SP/SNP dynamics; with only social reward dynamics adjusted, it reproduced rat SP/SNP dynamics, including early high social motivation in rats and delayed social motivation in mice. Overall, SD rats exhibit immediate, high social motivation that wanes with time, whereas C57BL/6J mice begin with low social motivation that grows, producing distinct early-phase behaviors and associated neural activation patterns.

The work directly addresses whether commonly used rat and mouse strains differ in social behavior by dissecting temporal dynamics of social investigation and relating them to motivation and neural activation. The pronounced early-phase differences in SP behavior—high social drive and prolonged social bouts in SD rats versus exploratory, high-transition behavior in C57BL/6J mice—support the hypothesis that strains differ in the induction and time course of social motivation. The results were robust across constrained and free interaction contexts, against an anxiety confound, and extended to other strains, indicating strain-specific rather than purely species-specific effects. Opposite behavioral responses to stimulus movements implicate subject-side motivational processes. Brain activation data (c-Fos) in MeA and NAc/VTA align with behavioral dynamics, suggesting that rapid engagement of social motivation circuitry underlies SD rats’ immediate social preference, while mice require longer exposure to recruit these circuits. The social-versus-food competition paradigm further demonstrates stronger baseline social motivation in rats and susceptibility of early-phase dynamics to motivational balance shifts, consistent with the computational model. These findings emphasize that strain selection critically influences observed social phenotypes and that accounting for motivational dynamics is essential for interpreting social behavior assays and modeling neuropathologies.

This study demonstrates marked, strain-specific differences in social investigation dynamics between SD rats and C57BL/6J mice. SD rats show immediate, strong social motivation with prolonged early interactions and c-Fos induction in MeA/NAc/VTA after brief exposure, whereas C57BL/6J mice initially explore both stimuli with short bouts and high transitions, showing delayed engagement of social motivation circuits. Behavioral responses to stimulus movements are opposite across strains, and competition with food reveals much stronger social drive in rats. A minimalist computational model reproduces these dynamics by altering social reward trajectories. These insights argue for careful selection of rodent strains and experimental designs that align with research questions on social behavior and pathology. Future work should map causal circuit mechanisms (e.g., MeA→NAc/VTA pathways) across strains, quantify neuromodulator dynamics during early versus late interaction phases, extend analyses to females and additional strains/species, and investigate how genetic or disease models interact with intrinsic strain-specific motivational dynamics.

- Generalizability is limited to the strains tested; while additional strains (ICR, BALB/c, Wistar Hannover) were probed, species-level conclusions cannot be definitively drawn. - c-Fos analyses capture early time windows and rely on group housing shortly before testing, potentially elevating baseline expression in mice; c-Fos is an indirect marker and lacks cell-type specificity. - The experimental SP/SNP setups restrict direct physical interaction; although free-interaction assays were included, most results derive from chamber-limited investigation. - The social vs food paradigm assumes comparable hunger/energy homeostasis across species/strains; physiological measures during deprivation were not collected here. - The computational model is intentionally simplified (four states, heuristic anxiety construct) and does not include direct transitions between motivational states or detailed neural mechanisms. - Most primary analyses emphasize males; although some female data are referenced, sex differences were not comprehensively explored.