Motor organization of Tritonia swimming. I. Quantitative analysis of swim behavior and flexion neuron firing patterns.

Abstract

Escape swimming of the marine mollusk T. diomedea consists of a series of alternating dorsal and ventral flexions. The timing of the swim behavior was determined quantitatively. The efferent neuron pools mediating the swim behavior as defined. This analysis provides the basis for comparison of the swim behavior and the motor-output pattern with a similar analysis of a premotor interneuronal network believed to generate the swim pattern. The swim behavior of freely moving and semirestrained animals was assessed by frame-by-frame analysis of filmed swims. Each swim cycle was divided into 4 segments: a rapid ventral flexion, a ventral pause, a rapid dorsal flexion and a dorsal pause. As cycle period increased during a swim, the duration of the 2 flexion segments remained relatively constant while the duration of the 2 pauses increased. The swim behavior is best characterized as phase constant with each dorsal flexion occurring at a constant phase of 0.5 relative to the cycle period, defined by the onset of successive ventral flexions. The organization of the putative motor neurons was examined byintracellular recordings in isolated brain and whole-animal preparations. Efferent neurons that burst during swimming were largely restricted to the dorsal surface of the pedal ganglia. The relative size of 4 operationally defined motor pools were estimated. The dorsal flexion neurons (DFN, 56% of cells sampled) and the ventral flexion neurons (VFN, 16% of cells sampled) fired alternating bursts, which corresponded 1- for 1 with the dorsal and ventral flexion movements, respectively. The class III (7% of sampled cells) fired 2 bursts per swim cycle and were coactive with the bursts in the DFN and VFN. The class IV (3% of sampled cells) also fired 2 bursts per cycle but these bursts occurred during the silent period between the DFN and VFN bursts. The VFN population was relatively homogeneous and fired with short, constant-duration bursts. The DFN can be divided into 2 classes. The DFN-A (55% of DFN sampled) had long-duration bursts that increased in duration as the swim progressed. These bursts began at nearly a constant latency after the preceding VFN burst. The DFN-B (45% of DFN sampled) had bursts that were short and constant in duration and began at nearly a constant phase of 0.5 relative to the VFN bursts. The timing of the VFN and DFN-B bursts closely resembled the timing of the 2 rapid flexion movements. The motor role of identified flexion neurons was evaluated in whole-animal preparations. The DFN-B and the VFN appeared to be excitatory motor efferents since their bursts were temporally correlated with the respective flexion movements; driving individual DFN-B or VFN gave rise to discrete flexion movements. Inhibiting the activity of single VFN during a swim decreased the amplitude and velocity of the ventral flexion. By these criteria, neither the DFN-A, nor class III, nor class IV cells provided significant peripheral excitation to the muscles that drove the flexion movements of a swim.