The vertebrate spinal cord generates oscillations that ultimately produce the alternating left–right body movements that characterize running, walking and swimming. It has long been hypothesized that the circuitry that produces these spinal cord rhythms resides in the spinal cord itself. The idea is that the circuit lays dormant until higher brain regions send tonic unpatterned excitation down to the spinal cord. This generalized excitatory signal (start moving!) awakens separate oscillatory networks within the spinal circuitry. This hypothesis was first proposed by Graham Brown, and has been widely accepted for ∼100 years. But is it actually true? In a recent paper published in the Journal of Physiology, Stephen Soffe, Alan Roberts and Wen-Chang Li tested this long-standing idea by studying locomotor circuits in Xenopus tadpoles.

One population of neurons (reticulospinal neurons) in the hindbrain (a higher brain region just forward of the spinal cord) have been thought to provide the primary excitatory drive to spinal networks in other systems. The authors reasoned that if Brown's model were true, then these neurons would excite spinal neurons, but not show any kind of rhythmic patterns during locomotor rhythms. To test this, the authors simultaneously recorded from reticulospinal neurons while also recording from motor neurons in the tadpole spinal cord. They found that there are indeed direct excitatory connections descending from reticulospinal neurons in the hindbrain onto motor neurons. However, when they looked at reticulospinal and motor neuron activity during bouts of swimming motor patterns they found something that didn't fit Brown's model. Instead of just tonic (i.e. continuous, unpatterned) excitation to the motor neurons, the reticulospinal neurons were firing in discrete bursts just before motor neurons on every cycle of the swimming rhythm. This suggests that reticulospinal neurons do more than just generally excite spinal circuits – the neurons appear to actively drive every cycle of the rhythm.

But perhaps the tonic excitation comes from even higher order brain regions? Soffe and colleagues decided to test this with a brute force approach. They systematically recorded from hundreds of individual cells in more anterior brain regions and measured when these cells were active relative to reticulospinal neurons during swimming rhythms. The authors saw no evidence for any tonically active descending inputs to the spinal cord in any of these recordings. They found that the reticulospinal neurons were indeed the first neurons in the brain active during every cycle. This suggests that oscillations in these neurons are really what's driving the tadpole spinal cord.

If reticulospinal neurons are driving every cycle of the swimming rhythm, then how do these cells themselves oscillate? The authors show that an interaction between the natural electrical properties of reticulospinal neurons and rhythmic inhibitory inputs onto the neurons could plausibly generate the appropriate rhythmic output. This suggests that the key circuitry for generating spinal cord rhythms does not actually reside in the spinal cord itself.

The work of Soffe and colleagues is important because it challenges a generally accepted hypothesis that has become entrenched. Through a series of careful experiments, they show that in their particular vertebrate (tadpole), the predictions of the entrenched model do not hold true. The authors' then provide an alternative model, one that (even if it's not true in every other vertebrate) will doubtless influence the direction that spinal cord research takes in future years.

S. R.
W. C.
Defining the excitatory neurons that drive the locomotor rhythms in a simple vertebrate: insights into the origin of reticulospinal control.
J. Physiol.