Activation of latent programs for maternal behavior

suggest that social transmission of pup retrieval may be less robust than previously reported, and180 that researchers should interpret conclusions drawn from small-sample studies with caution.181 Despite our findings, we should acknowledge some limitations. Our study focused on behavioral measures182 alone, without the direct assessment of neural activity, genetic differences, or hormonal states. While our183

strongly suggest that virgin females can acquire maternal behavior independent of visual observation,184 we cannot directly determine the underlying neural mechanisms. The pattern of substantially different “reach185 pup latencies” but intact “return pup latencies” supports the hypothesis that maternal behaviors exist as186 latent neural programs that direct pup exposure can trigger, independent of social observation. Future187 studies incorporating neural recordings or manipulations will provide valuable insights into the mechanisms188 underlying this behavioral plasticity.189 Methods190 Animals191 The Cantonal Veterinary Office Zurich, Switzerland, approved all animal procedures and experiments. All192 experiments used female C57Bl6/Crl1 mice (Jackson Lab) aged between 7 to 22 weeks at the start of the193 behavioral experiment. Animals were group-housed (around 2-4 litter-mates) in individually-ventilated cages194 (IVC) in a 12 h light/dark cycle room, and we provided them with food and water ad libitum. The subjects195 fell into two groups: dams (either primiparous or multiparous at the start of the experiment) and sexually-196 and pup-naive females (virgins), which were housed separately. Virgins were generated in-house, and we197 transferred them to a different pup- and breeding-free room upon weaning to avoid exposure to other pups198 and breedings. Virgins remained male-naive throughout the experiments. Dams and virgins were prescreened199 (“day zero”, d0) for retrieval and pup mauling before initiating the behavioral task. We only accepted those200 virgins that retrieved 1 pup or less on d0.201 8 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint Behavioral task202 The established pup retrieval task [6] consists of two consecutive stages, ’demonstration’ and ’testing’, which203 we repeated over 4 days after prescreening (P1-4 of mother’s pups). During the ’demonstration’ phase, the204 dam and the virgin were placed on opposite sides of a conventional 1800 cm 2 home cage (588 mm × 194 mm205 × 395 mm, NexGen IVC for rats 1800, Allentown), and an opaque or transparent barrier separated them206 at the midline. We gave the animals 20 minutes to acclimatize before each demonstration session began.207 The dam’s entire litter was grouped in a corner of the arena on the dam’s side, and we covered them with208 nesting material. After an additional 1–2 minutes of acclimatization, we removed one pup from the nest and209 pseudorandomly placed it in either of the 3 other corners of the dam’s cage partition, and we gave the dam210 a maximum of 2 minutes and 5 seconds to retrieve the displaced pup back to the nest. After 10 trials, the211 animals returned to their home cage, and a 30-minute break followed. We changed the used cage into a new,212 clean cage, and we placed the virgin on one of the partitions of the arena. After 20 minutes of acclimatization,213 the ’testing’ phase began, and we put the entire litter back into the corner of the testing arena, covering214 them with nesting material. After 1–2 minutes of acclimatization, we tested the virgin for pup retrieval215 in 10 trials, as we described above. At the end of the experiment, all the animals returned to their home216 cages. We applied two different conditions of this task: ’transparent barrier’, where the barrier dividing the217 cage was transparent (and therefore the virgins could observe the dam) and an ’opaque’ condition, which218 consequently impeded the virgin from observing the dam and litter.219 Behavioral analysis220 A top view B/W CMOS camera (60516, Stoelting Europe) integrated with an ANY-maze system (Stoelting221 Europe) recorded all mouse behavior. The following parameters were quantified for analysis: Reach pup222 latency: the time to approach the pup ending in a pick-up, from pup displacement from the nest to the223 pick-up. Return pup latency: the time to retrieve the pup to the nest, from pick-up to return to the nest.224 Total latency: the whole trial duration, which is the sum of reach pup latency and return pup latency.225 Retrieval success: we categorized a trial as ’successful’ if the female (mother/virgin) brought the pup226 back to the nest within the time allotted (maximum 2 minutes and 5 seconds), and ’failed’ otherwise (the227 experimenter then returned the pup back to the nest at the end of the allotted time).228 Statistical analysis229 We analyzed retrieval performance at the trial level (success = 1, failure = 0) and retrieval latencies using230 custom Python scripts combining parametric and mixed-effects modeling approaches. To assess the effect231 of barrier conditions on virgin performance across days, and given the repeated-measures structure of the232 data, we applied linear mixed-effects models (LMMs) fit using restricted maximum likelihood estimation233 (REML) to examine the effects of barrier type, session, and their interaction on both success rates and234 9 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint log-transformed latencies. In these models, we included pair IDs (for success rates) or the individual subject235 (for latency) as a random intercept to account for repeated measures, while fixed effects included group236 (mother vs. virgin), barrier, session, and relevant interactions. We additionally modeled trial-level outcomes237 using a binomial generalized linear model (GLM), with the number of successful and unsuccessful trials as238 the response variable. To follow approaches from previous studies [6], we additionally conducted a two-way239 ANOVA on the average success rates of animals in both barrier types.240 To assess the robustness of our findings, we conducted a simulation-based power analysis using the fixed-241 and random-effects estimates from the fitted LMMs. For this, we generated simulated datasets under the242 alternative hypothesis with varying numbers of animals per condition, and the same REML-based mixed-243 effects model analyzed each dataset. The proportion of simulations yielding significant interaction terms244 (α = 0.05) estimated statistical power, allowing us to determine the sample sizes required to detect effects245 of the magnitude that prior studies observed. Finally, we examined the relationship between maternal and246 virgin performance over days using Spearman correlation, to test whether maternal success predicted virgin247 acquisition within pairs.248 References249 [1] Johannes Kohl. “Parenting — a paradigm for investigating the neural circuit basis of behavior”. In:250 Current Opinion in Neurobiology 60 (Feb. 2020), pp. 84–91. issn: 09594388. doi: 10.1016/j.conb.251 2019.11.011. url: https://linkinghub.elsevier.com/retrieve/pii/S0959438819301229 (visited252 on 02/22/2024).253 [2] Catherine Dulac, Lauren A. O’Connell, and Zheng Wu. “Neural control of maternal and paternal be-254 haviors”. In: Science 345.6198 (Aug. 15, 2014), pp. 765–770. issn: 0036-8075, 1095-9203. doi: 10.1126/255 science . 1253291. url: https : / / www . science . org / doi / 10 . 1126 / science . 1253291(visited on256 11/29/2023).257 [3] Danielle S. Stolzenberg and Heather S. Mayer. “Experience-dependent mechanisms in the regulation258 of parental care”. In: Frontiers in Neuroendocrinology 54 (July 2019), p. 100745. issn: 00913022. doi:259 10 . 1016 / j . yfrne . 2019 . 04 . 002.url: https : / / linkinghub . elsevier . com / retrieve / pii /260 S0091302218301249 (visited on 05/25/2024).261 [4] Gen-ichi Tasaka et al. Dopamine Dynamics Underlying Alloparental Behavioral Acquisition Facilitated262 by Top-down Orbitofrontal Inputs in Female Mice . Feb. 4, 2023. doi: 10.1101/2023.02.03.527077.263 url: http://biorxiv.org/lookup/doi/10.1101/2023.02.03.527077 (visited on 07/10/2024).264 [5] Bianca J. Marlin et al. “Oxytocin enables maternal behaviour by balancing cortical inhibition”. In:265 Nature 520.7548 (Apr. 23, 2015), pp. 499–504. issn: 0028-0836, 1476-4687. doi: 10.1038/nature14402.266 url: https://www.nature.com/articles/nature14402 (visited on 11/29/2023).267 10 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint [6] Ioana Carcea et al. “Oxytocin neurons enable social transmission of maternal behaviour”. In: Nature268 596.7873 (Aug. 26, 2021), pp. 553–557.issn: 0028-0836, 1476-4687. doi: 10.1038/s41586-021-03814-7.269 url: https://www.nature.com/articles/s41586-021-03814-7 (visited on 11/29/2023).270 [7] Jennifer K. Schiavo et al. “Innate and plastic mechanisms for maternal behaviour in auditory cortex”.271 In: Nature 587.7834 (Nov. 19, 2020), pp. 426–431. issn: 0028-0836, 1476-4687. doi: 10.1038/s41586-272 020-2807-6. url: https://www.nature.com/articles/s41586-020-2807-6 (visited on 11/29/2023).273 [8] Shota Okabe et al. “Pup odor and ultrasonic vocalizations synergistically stimulate maternal attention274 in mice.” In: Behavioral Neuroscience 127.3 (June 2013), pp. 432–438. issn: 1939-0084, 0735-7044. doi:275 10.1037/a0032395. url: http://doi.apa.org/getdoi.cfm?doi=10.1037/a0032395 (visited on276 11/29/2023).277 11 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint Figures278 Figure 1: Barrier condition on pup retrieval task does not influence virgin performance. A. Task schematic of pup retrieval behavior task. A dam demonstrates pup retrieval for 10 trials, while a virgin observes through a transparent or opaque barrier. After a 30-minute break, the virgin is then tested for pup retrieval. The whole paradigm is repeated for 4 days. B. Heatmaps show virgin retrieval performance in transparent condition (top plot, N = 14 animals) and opaque condition (bottom plot, N = 13 animals). Dark green color = 100% retrieval rate; white color with ’X’ mark = 0% retrieval rate. C. Lineplots depict successful retrievals over days (median with IQR) for each experimental group (Mother:Transparent in blue, Mother:Opaque in cyan, Virgin:Transparent in red, Virgin:Opaque in orange). D. Cumulative distributions (Kaplan-Meier) for retrieval onset in Virgins:Transparent (red) and Virgins:Opaque (orange). The log-rank test did not show any statistical significance rates between the two barrier conditions (log-rank p = 0.310). 12 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint Figure 2: Latency to reach the pup, but not latency to return the pup, differs between mothers and virgins. Boxplots show median latency and IQR for Left. total trial duration, Middle. “reach pup latency” and Right. “return pup latency”. 13 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 6, 2025. ; https://doi.org/10.1101/2025.11.05.686743doi: bioRxiv preprint