Belt-and-suspenders as a biological design principle.

Recent studies have shown both the pFRG and the preBötC are sufficient to generate respiratory rhythm, and are hypothesized to do so via distinct mechanisms (Onimaru and Homma 2003; Mellen, Janczewski, Bocchiaro and Feldman 2003). The coexistence of mechanistically distinct, functionally matching networks (defined as degeneracy, Edelman and Gally 2001) is a ubiquitous feature of motor networks in both invertebrates (Selverston and Miller 1980) and vertebrates (DiDomenico, Nissanov and Eaton 1988). In almost all cases, a consensus exists about which subsystem is the "primary" rhythm generator, yet consistently, the effect of modulators on the isolated primary rhythm generator is qualitatively different than their effect on the more intact network (Ayali and Harris-Warrick 1999) and, in the intact animal, all rhythmogenic networks are active during motor pattern generation. Thus, at best, ascribing primacy to a particular network has weak support (since the other networks can produce qualitatively similar patterns; Prinz, Bucher and Marder 2004) and little explanatory power (since effects of modulatory inputs on the isolated "primary" rhythm generator do not persist in more intact networks). The ubiquity of degenerate networks for motor pattern generation suggests that a more useful question is why such an organization exists. We propose that degeneracy is ubiquitous because it reduces the phenotype's sensitivity to genetic mutation and environmental perturbation, and broadens the adaptiveness of motor patterns.