Automated pup-level analysis reveals distinct effects of prenatal CBD and THC exposure on maternal retrieval

This study explored the effects of prenatal exposure to CBD and THC on maternal caregiving behavior in C57BL/6J mice. To address the limitations of manual behavioral scoring, researchers developed a machine-automated scoring system called Machine-Automated Scoring of the Pup Retrieval Test (MAS-PRT). This system combines Multi-Animal DeepLabCut for tracking the dam and pup, along with Detectron2 for reconstructing dynamic nest structures.

The pipeline was validated against 170 manually annotated retrieval trials, demonstrating high concordance with manual measurements and enabling reproducible extraction of encounter latency, retrieval latency, and locomotor trajectories.

The results showed that prenatal exposure to either CBD or THC did not impair general nest-building or overall maternal care in the dams' home cages. However, there were distinct effects on the pups: those exposed prenatally to either compound exhibited reduced body weight at postnatal day 5. Further analysis using Cox proportional hazards modeling revealed that CBD-exposed dams had a weight-dependent increase in the probability of encountering and retrieving lighter pups, regardless of the pup's sex.

Spatial tracking also indicated that dams spent significantly shorter total trajectories retrieving female progeny, whether they were exposed to CBD or THC.

When examining the learning dynamics of the dams, longitudinal trial-by-trial analysis demonstrated intact task acquisition in both control and CBD-exposed dams. In contrast, THC-exposed dams displayed a flattened learning curve, characterized by lower retrieval latencies on initial trials. Overall, these findings suggest that prenatal cannabinoid exposure does not lead to generalized disruption of maternal care but rather results in compound-specific alterations in maternal reactivity, retrieval kinematics, and learning dynamics.