A Cognitive and Neural Model for Adaptive Emotion Reading by Mirroring Preparation States and Hebbian Learning

Two types of modelling approaches exist to reading an observed person's emotions: with or without making use of the observing person's own emotions. This paper focuses on an integrated approach that combines both types of approaches in an adaptive manner. The proposed models were inspired by recent advances in neurological context. Both a neural model and a more abstracted cognitive model are presented. In the first place emotion reading is modelled involving (preparatory) mirroring of body states of the observed person within the observing person. This involves a recursive body loop: a converging positive feedback loop based on reciprocal causation between preparations for body states and emotions felt. Here emotion reading involves the person's own body states and emotions in reading somebody else's emotions: first the same feeling is developed by mirroring, and after feeling the emotion, it is imputed to the other person. In the second place, as an extension an adaptive process is modelled based on Hebbian learning of a direct connection between a sensed stimulus concerning another agent's body state (e.g., face expression) and an emotion imputation state. After this Hebbian learning process the emotion is imputed to the other agent before it is actually felt, or even without it is felt. Both the mirroring and Hebbian learning processes first have been modelled at a neural level, and next, in a more abstracted form at a cognitive level. By means of an interpretation mapping the paper shows the relation between the obtained cognitive model and the neurological model. In addition to specifications of both models and the interpretation mapping, simulation results are shown, and automated verification of relevant emerging properties is discussed. © 2010 Elsevier B.V.