Arun, D. | bioRxiv Science

bioRxiv2025

Subadult macaws (Ara glaucogularis) copy better than adults but are less likely to imitate in problem-solving tasks

Copying via imitation (replicating others bodily actions) or emulation (replicating others action goals) underpins the cultural transmission of social norms and technical skills in humans. In this study, we investigated whether blue-throated macaws (Ara glaucogularis) are capable of copying conspecifics using two problem-solving devices: door and plug. The door device could be solved by sliding a panel either UP or DOWN, while the plug device required removing a stopper via either a PUSH or PULL action. A control group (N=5), which received no demonstrations, consistently preferred a single solution for each device-- DOWN for door and PULL for plug. In the subsequent test, adult (N=5) and subadult (N=5) macaws from the test group observed demonstrators (N=2) performing the non-preferred action for each device. Significantly more test birds copied the demonstrated action (UP) for door, but not PUSH for plug, because the subadults were less likely to imitate the PUSH action than the adults. However, the subadults took significantly less time to approach and manipulate the devices than the adults. Overall, our study shows that macaws can copy conspecifics in problem-solving tasks, which facilitates information transfer and transmission of foraging techniques, a prerequisite for foraging culture to arise.

animal behavior and cognition↗

bioRxiv2025

Extreme Value Theory for Modeling Category Decision Boundaries in Visual Recognition

Several possibilities exist for modeling decision boundaries in category learning, with varying degrees of human fidelity. This paper finds evidence for preferentially focusing representational resources on the extremes of the distribution of visual inputs in a generative model as an alternative to the central tendency models that are commonly used for prototypes and exemplars. The notion of treating extrema near a decision boundary as features in visual recognition is not new, but a comprehensive statistical framework of recognition based on extrema has yet to emerge for category learning. Here we suggest that the statistical Extreme Value Theory [Coles et al., 2001] provides such a framework. In Experiment 1, line segment stimuli that vary in a single dimension of length [Hsu and Griffiths, 2010] are used to assess how human subjects and statistical models assign category membership to a gap region between two categories shown as reference stimuli. A Weibull fit better predicts an observed human shift when moving from uniform to enriched or long tails as reference stimuli. In Experiment 2, more complex 2D rendered face sequences drawn from morphspaces [Folstein et al., 2012] are used as stimuli. Again, the Weibull fit better predicts an observed human shift when reference stimuli are sampled differently. An extrema-based model lends new insight into how discriminative information may be encoded in the brain with implications for the understanding of how decision making works in category learning.

neuroscience↗