Towards the neural modelling of mental spaces

This work relates the theory of Mental Spaces with neural models that sustain associations between patterns. The theory of context-dependent matrix associative memories is used to establish a neural counterpart for the connectors between mental spaces. Two applications of these neural models concerning mental space builders for linguistic topics and prepositions, respectively, are described. Finally, it is shown that this relation of mental spaces and matrix memories, establish a link with LSA that may help to develop a physiological approach to semantic networks.