Abstract
AbstractWhen an organism explores a new environment or stimulus it varies its behavior to ensure proper sampling. As contingencies are learned, behavioral variance can give-way to routines and stereotypies. This phenomenon is common across animal species and non-biological learning systems but has not been well studied in the social domain, in which the stimulus an animal investigates, another individual, may react negatively to unexpected behaviors. Here we develop a set of generalizable measures of interaction variability to test the hypothesis that social relationship formation is associated with initially high variability, followed by stabilization of dyadic interaction patterns. We focus these tests on female degus (Octodon degus), rodents predisposed to form cooperative peer relationships. Degus were presented with a series of 20 minute, dyadic “reunion” sessions across days, interleaving exposures to familiar and unfamiliar same-sex conspecifics. We found that degu pairs could be distinguished from one-another by their session-to-session interactive behaviors, suggesting dyad-specific social relationships. Consistent with predictions, stranger interactions changed from the first to second exposure, after which they became less variable across days. Unexpectedly, however, long-term cagemates showed overall more interaction variability across sessions than strangers. Taking into account all interaction variability measures, differences between strangers and cagemates were only observed when considering many (> 10) interactions, or longer (>5 min, or session-to-session) time periods. Shorter scales, including interaction transitions, showed meaningful patterns but no reliable group differences. The results suggest that while variability of social behavior over minutes or across sessions may change as relationships develop, interactions between unfamiliar adult female degus are not systematically more variable than those of cagemates. Given known ecological pressures on female degus to form large, stable social networks, these patterns are consistent with the notion that interaction variability may be maladaptive for social coordination.
Publisher
Cold Spring Harbor Laboratory