Testing computational models of learning from social, real, and fictive feedback in human and nonhuman primates
Final Report Abstract
The collaboration project has advanced our understanding of the computational principles and neuronal underpinnings of learning from own action outcomes and from observing social conspecifics act and receive outcomes in human and non-human primates. However, cross-species comparison has been limited by a number of factors that, in part, are difficult to control. Implementing tasks that can be equally well performed by humans and macaques has turned out more difficult than originally expected. A major factor in this are differences in motivation: whereas mildly water-deprived monkeys with relatively little entertainment in their home cages are highly motivated to participate in experiments and even find passive Pavlovian conditioning very interesting, human participants are much harder to motivate and often bored by conditioning experiments. The motivational values of the rewards are hardly comparable between the species. On the other hand, it is very difficult to implement fictive (counterfactual) outcomes in a way understood by monkeys, such that we decided to omit this condition. Nevertheless, the results suggest that the general principles of feedback- and observation-based learning and decision making are similar in human and non-human primates. In both species, homolog areas of the medial frontal cortex appear to represent prediction error signals as well as observed errors. Moreover, we could show differences in modelling parameters as well as EEG correlates of monitoring of own compared to observed outcomes. In follow-up projects we will continue the collaboration with our Japanese partners to elucidate the function of the posterior medial frontal cortex in social and non-social performance monitoring and decision-making.
Publications
- (2018) Performance monitoring in the medial frontal cortex and related neural networks: from monitoring self-actions to understanding others’ actions. Neurosci Res 137:1-10
Ninomiya T, Noritake A, Ullsperger M, Isoda M
(See online at https://doi.org/10.1016/j.neures.2018.04.004) - (2018) Social reward monitoring and valuation in the macaque brain. Nat Neurosci 21:1452-1462
Noritake A, Ninomiya T, Isoda M
(See online at https://doi.org/10.1038/s41593-018-0229-7) - (2019). The feedback-related negativity indexes prediction error in active but not observational learning. Psychophysiology, e13389
Burnside R, Fischer AG, Ullsperger M
(See online at https://doi.org/10.1111/psyp.13389)
