Longitudinal gradients in the spinal cord of Xenopus embryos and their possible role in coordination of swimming.

We consider the hypothesis that the normal rostrocaudal spread of motor activity during swimming in Xenopus embryos depends on a longitudinal, rostrocaudal gradient in synaptic drive received by the pattern generating neurons in the spinal cord. Anatomical evidence shows that there is a rostrocaudal gradient in excitatory and inhibitory premotor interneuron numbers. Intracellular recordings were therefore made from neurons at different longitudinal positions to seek direct evidence for a gradient in synaptic drive. The main finding was that during fictive swimming the amplitude of the tonic excitatory synaptic input and the mid-cycle inhibition declined in a rostrocaudal direction and caudal to the twelfth postotic segment inhibition was absent altogether. Experimental manipulation of the gradients by the caudal application of pharmacological agents lead to changes or even reversal of the normal direction of spread of motor activity. To test whether a gradient in synaptic drive could organize a rostrocaudal spread of activity we have used a 48 neuron, 12 segment model of the spinal pattern generating network. This has shown that the gradient hypothesis can provide a sufficient explanation for much of the embryo's motor coordination. A gradient in synaptic drive provides a flexible mechanism to control sequences of motor activity.