Neural adaptations for social learning: Structural and functional investigations of action observation networks in macaques, chimpanzees, and humans

Social learning is an important ability in primate life, and human specializations for social learning are part of what set us apart from the rest of the animal kingdom. In particular, humans' ability to copy not only the outcomes of observed actions but also their movement details has been linked to the emergence of cumulative culture. Social learning involves an action observation network that is distributed across frontal, parietal, and occipitotemporal cortex. This dissertation reports species differences in the structure and function of these networks that may underlie species differences in social learning. First, diffusion tensor imaging studies revealed progressively greater parietal and inferotemporal connectivity from macaques to chimpanzees to humans. These structural differences parallel, and may underlie, functional differences. FDG-PET neuroimaging studies in chimpanzees revealed that like humans and unlike macaques, chimpanzees have overlapping brain responses for performed action, observed transitive action, and observed intransitive action. Since chimpanzees and humans but not macaques are capable of copying movement details (imitating), this suggests that the ability to "mirror" not only action outcomes but also movement details is a correlate to the ability to copy those movement details. Furthermore, the chimpanzee neural response to observed action was situated mainly in prefrontal cortex, which may reflect top-down processing related to a conceptual, abstract representation of the observed action, while humans had greater parietal and occipitotemporal activation, which may reflect greater bottom-up processing on the details of movements, body parts, and objects. This may explain why humans tend to copy movement details while chimpanzees tend to copy action outcomes. Finally, chimpanzees with greater activation in ventral premotor cortex and lateral occipital cortex performed better in a separate behavioral test on copying action outcomes/movements and tool use, suggesting that selection pressure for social learning behavior could act on brain responses to observed action. These results are relevant to the evolution of action observation, social learning, and possibly culture.