Modelled optimisation of communication for ecologically relevant dyadic group decision-making

Group decision-making allows cognitive resources and information to be pooled, compensating for individual decision deficits which improves decision accuracy. However, unhelpful group dynamics and biases can sometimes emerge resulting in poor performance. Importantly, it is unclear how communication between members can be optimised for group decision performance. We addressed this by modelling data acquired from participants in an ecologically relevant perceptual decision-making task. These computational agents were coupled using three different inter-agent communication conditions, including two novel, and optimal pairing was investigated. Continuous, imbalanced communication between optimised pairs was found to have the best mean performance.