HIPPOCAMPAL MEDIATION OF STIMULUS REPRESENTATION - A COMPUTATIONAL THEORY

The authors propose a computational theory of the hippocampal region's function in mediating stimulus representations. The theory assumes th at the hippocampal region develops new stimulus representations that e nhance the discriminability of differentially predictive cues while co mpressing the representation of redundant cues. Other brain regions, i ncluding cerebral and cerebellar cortices, are presumed to use these h ippocampal representations to recode their own stimulus representation s. In the absence of an intact hippocampal region, the theory implies that other brain regions will attempt to learn associations using prev iously established fixed representations. Instantiated as a connection ist network model, the theory provides a simple and unified interpreta tion of the functional role of the hippocampal region in a wide range of conditioning paradigms, including stimulus discrimination, reversal learning, stimulus generalization, latent inhibition, sensory precond itioning, and contextual sensitivity. The theory makes novel predictio ns regarding the effects of hippocampal lesions on easy-hard transfer and compound preexposure. Several prior qualitative characterizations of hippocampal function-including stimulus selection, chunking, cue co nfiguration, and contextual coding-are identified as task-specific spe cial cases derivable from this more general theory. The theory suggest s that a profitable direction for future empirical and theoretical res earch will be the study of learning tasks in which both intact and les ioned animals exhibit similar initial learning behaviors but differ on subsequent transfer and generalization tasks.