|
Description
|
Learning enables animals to track changing environments, but whether cognitive performance scales with brain size remains unclear, because most evidence comes from comparisons across species or populations. The common shrew (Sorex araneus) allows this to be tested within individuals: it undergoes Dehnel's phenomenon, a reversible, region-specific reduction in brain size in winter followed by partial regrowth in spring, with the neocortex and hippocampus shrinking most. We tested visual associative learning, path integration, and spatial learning in a maze across three seasons, comparing large-brained summer juveniles, small-brained winter subadults, and intermediate-brained spring adults, in wild-caught and captive individuals. Bayesian mixture models identified two latent performance modes in every task, separating execution quality from the probability of adopting an efficient strategy. Spatial learning was robust: all groups improved across trials and converged by the final trial, and neither season nor captivity reliably predicted strategy adoption. Visual associative learning was consistently inefficient but showed no winter decline; instead, shrews were most likely to adopt the high-performance strategy during spring regrowth. Path integration was the task most clearly impaired in winter, and efficiency declined across trials in all groups, suggesting that exploration is prioritised over navigational precision once a route becomes familiar. Captivity acted as a task-dependent filter, depressing associative learning while leaving spatial learning unaffected and increasing return-path straightness. Seasonal brain shrinkage therefore does not impair cognition uniformly: ecologically critical functions are preserved while less urgent ones are allowed to decline, and brain size alone cannot serve as a proxy for learning ability.
|
|
Keyword
|
associative learning, brain plasticity, captivity effects, Dehnel's phenomenon, path integration, seasonality, Sorex araneus, spatial cognition |
