A computational model of a global neuronal workspace with stochastic connections

This paper presents a model of long-range cortical communication by means of a global neuronal workspace similar to that proposed by Dehaene and Naccache (2001). The model resembles that of Shanahan (2008), which was based on reverberating circuits of one-to-one connections, but uses a stochastic wiring regime in place of the highly regular scheme used there. The paper offers a systematic analysis of the influence of certain parameters on the dynamics of networks built according to this regime. Armed with a fuller understanding of the origins of the observed behaviour in the model, the qualitative behaviour of the previous model is replicated, and further unexplored choices are examined for ongoing work.