Catecholamine Challenge Uncovers Distinct Mechanisms for Direct Versus Indirect, but Not Social Versus Non-Social, Learning

Abstract

1Evidence that social and individual learning are at least partially dissociable sustains the belief that 2 humans possess adaptive specialisations for social learning. However, in most extant paradigms, 3 social information comprises an indirect source that can be used to supplement one's own, direct, 4 experience. Thus, social and individual learning differ both in terms of social nature (social versus 5 non-social) and directness (indirect versus direct). To test whether the dissociation between social 6 and individual learning is best explained in terms of social nature or directness, we used a 7 catecholaminergic challenge known to modulate learning. Two groups completed a decision-making 8 task which required direct learning, from own experience, and indirect learning from an additional 9 source. The groups differed in terms of whether the indirect source was social or non-social. The 10 catecholamine transporter blocker, methylphenidate, affected direct learning by improving 11 adaptation to changes in the volatility of the environment but there was no effect of 12 methylphenidate on learning from the social or non-social indirect source. Thus, we report positive 13 evidence for a dissociable effect of methylphenidate on direct and indirect learning, but no evidence 14 for a distinction between social and non-social. These data fail to support the adaptive specialisation 15 view, instead providing evidence for distinct mechanisms for direct versus indirect learning. 16

Social learning in models and minds

“…Their research with this non-social control has revealed domain-specific processing but not for the social domain. They have shown that indirect learning differs from direct learning (Cook et al, 2018), and that meta-learning differs from first-order learning (Cook et al, 2019), but they have not found that different models fit the data when people believe they are getting information from other agents rather than rigged roulette wheels (Cook et al, 2018).…”

“…A more reliable way of solving the social specificity problem is to use contrastive experiments with an 'inanimate' or 'non-social' control condition (Cook et al, 2014(Cook et al, , 2018Heyes, 2014aHeyes, , 2014bLockwood et al, 2020). Cook and colleagues have pioneered this approach in studies of social learning.…”