This chapter illustrates how computational models based on abstract neural net­ works can provide insights into questions such as where in the brain learning occurs and the mechanisms that bring it about. It serves as a tutorial on aligning quantitative learning theory with physiology, thereby establishing a potential conduit for communication between molar and molecular levels of analyses. One might say that such efforts are about bringing abstract models to life in real nervous systems. Specifically, it brings together two lines of research: (a) studies of brain circuits underlying classical conditioning of a simple skeletal response, the rabbit nictitating membrane response (NMR) and (b) developing computa­ tional models applicable to real-time conditioning phenomena such as response

topography and CS-US interval effects. Our work stands on a foundation provided by the “extended laboratory” (Gabriel, 1988) of those who investigate classical eye blink/NMR conditioning (Gormezano, Prokasy, & Thompson, 1987).