Central pattern generators (CPGs) have traditionally been modeled as sets of coupled bistable oscillators [2]. We present a framework for constructing models which avoid the shortcomings of these traditional models, while remaining biologically plausible. We demonstrate our approach by generating a novel model of lamprey locomotion. However, we suggest that the methods presented here can be more generally applied to modeling any neural system which produces and controls dynamic patterns of forces. Key words: Central pattern generator; CPG; Neural computation; Lamprey locomotion; Coupled oscillators
Chris Eliasmith, Charles H. Anderson