Food discovery is associated with different reliance on social learning and lower cognitive flexibility across environments in a food caching bird
Data files
Mar 10, 2021 version files 20.47 MB
-
ArrayVisits_ForNetwork.txt
1.44 MB
-
BirdAges.txt
52.31 KB
-
LearningTest.txt
2.77 MB
-
NovelFeederDiscoveryData.txt
14.42 MB
-
README.txt
6.05 KB
-
ReversalTest.txt
1.77 MB
Apr 26, 2021 version files 20.47 MB
-
ArrayVisits_ForNetwork.txt
1.44 MB
-
BirdAges.txt
52.31 KB
-
LearningTest.txt
2.77 MB
-
NovelFeederDiscoveryData.txt
14.42 MB
-
PreferredArrays.txt
3.58 KB
-
README.txt
6.33 KB
-
ReversalTest.txt
1.77 MB
Abstract
Social learning is a primary mechanism for information acquisition in social species. Despite many benefits, social learning may be disadvantageous when independent learning is more efficient. For example, searching independently may be more advantageous when food sources are ephemeral and unpredictable. Individual differences in cognitive abilities such as spatial memory, which affect an individual’s environmental predictability, can also be expected to influence social information use. We investigated how resident food-caching chickadees discovered multiple novel food sources in both harsher, less predictable high elevation and milder, more predictable low elevation winter environments. Chickadees at high elevation were faster at discovering multiple novel food sources and discovered more food sources than birds at low elevation. Using network-based diffusion analyses, we found that while birds at both elevations used social information, the contribution of social learning to novel food discovery was significantly lower at high elevation. At both elevations, chickadees with better spatial cognitive flexibility were slower at discovering novel food sources, likely because spatial cognitive flexibility allows better tracking of changing information about food sources, making the environment more predictable. Overall, our study supported the prediction that harsh and unpredictable environments should favor less reliance on social learning.
See README.txt