Abstract
The signals produced by neural network models can be broadly divided into two classes: (1) cognitive (objective) representations, and (2) affective (non-cognitive) representations. It is not particularly difficult to understand what cognitive representations represent. Representations are said to be cognitive if they represent objects, either as entities or as events, or feature fragments that can be assembled to make a representation of an entity or event. Neurologically, cognitive representations are closely tied to the sensory cortices and to the association cortices in the temporal, parietal, and occipital lobes of the neocortex. Psychologically, cognitive representations are closely linked to the notions of intuitions and concepts.