Probabilistic characterization of simultaneous nerve impulse sequences controlling dipteran flight

Abstract

A probabilistic method of analysis of spike trains is presented which provides a complete statistical description of spike sequences and allows the elucidation of some of the properties of the neural interconnections producing the output patterns. The flight motor system of the blowfly, Calliphora terraenovae, is analyzed by this method. Individual motor units show large, non-serially correlated, cycle-to-cycle variations in frequency superimposed upon long term frequency trends. These trends are apparently not generated by averaging the cycle-to-cycle variations in input excitation over a long time period. The different motor units share the same short term input excitation and the excitation causing long term trends. Units in different muscles show no preferred phase or latency relationships; they maintain similar frequencies but their phases drift through all possible values. Frequency control without phase control may be accomplished by shared excitation with a total input frequency many times the output frequency. Units in the same muscle maintain strong phase relationships. Constant phase relationships during variations in frequency may, among other models, be due to reciprocal inhibition or a common linearly rising input. Sensory feedback cannot account for the degree of phase or frequency regulation shown. Thus central patterning of the output sequence is necessary, as in the locust, and the two flight systems can be considered as integradable evolutionary variations.