How fast can you go?

Rhythmic activities such as walking need tight coordination. In mice, pace is tweaked by a specific set of spinal-cord neurons that, surprisingly, make the animals walk faster by inhibiting the underlying circuit.

Watch your step — walking may seem simple, but is actually quite a complex task. As with other rhythmic motor behaviours (breathing or swallowing, say), locomotion relies on a finely tuned neuronal network that is headquartered in the spinal cord^1,2. The ensemble of spinal neurons that generates a coordinated rhythmic activity is known as a central pattern generator. The rhythm and periodicity of this network determines movement features such as the alternation between left and right, or the speed of walking, jumping or swimming³. Understanding how this circuit operates and the specific roles of the different neurons that participate in it is difficult, but Gosgnach and colleagues⁴ have taken up the challenge. On page 215 of this issue, they report that the activity of a group of spinal-cord neurons controls the speed of locomotor behaviour in the mouse.

Neuronal circuits are formed by a network of interconnected excitatory and inhibitory neurons. In a very simplistic model, the former group switches the circuit on and the latter turns it off. Gosgnach et al. studied the role of a subclass of inhibitory spinal neurons known as V1 neurons, which are thought to be part of the central pattern generator. The researchers used detailed information about the gene-regulatory factors that underlie the development and specialization of these neurons, to generate mutant mice in which V1 neurons were either eliminated or silenced acutely during the experiment, and examined the changes in locomotor activity.

Counterintuitively, they found that removing the inhibition caused by V1 neurons slows the speed of the locomotor rhythm. Mutant mice lacking V1 neurons are unable to walk fast, but they can maintain normal motor behaviour at a slower pace. The authors demonstrate that this is because motor neurons connecting the spinal cord to the muscle are not sufficiently inhibited, because connections from the V1 neurons are missing in the mutant animals. The activity of motor neurons needs to be interleaved with precise periods of silence to generate a faster pace. These results underscore the value of inhibition in the nervous system: the delicate balance and fine-tuning of neuronal activity set by inhibitory connections is not only important in quieting down the system, but can also change core features of nervous-system function.

Even though Gosgnach and colleagues' mice slow down, other parameters of their locomotor activity remain intact, such as the alternation of left and right limbs necessary for coordinated stepping. As the authors showed previously⁵, a different class of spinal neurons (V0) is responsible for left–right coordination. In mutant mice lacking V0 neurons, the left and right motor neurons fire at the same time, rather than alternating. These are significant insights into how the work is distributed among the vast collection of spinal neurons that make up the central pattern generator.

How are the circuits of the central pattern generator established? When exploring the formation of circuits, scientists have focused on two principal alternative theories. One of these proposes that a genetically driven programme predetermines the identity of the neurons that participate in a given circuit and dictates how and between which of them connections are made⁶. The second proposes that early in development, neurons exhibit spontaneous electrical activity that affects the process of neuronal specialization and the establishment of appropriate connections^7,8. Changes in activity during development can indeed lead to resetting of the intrinsic excitability of neurons as well as reconfiguration of connections⁹. In particular, formation of motor central pattern generators requires early spontaneous electrical activity, because when this activity is disrupted, motor neuron fibres fail to follow their normal trajectories¹⁰.

However, Gosgnach et al.⁴ find that the activity of V1 neurons does not seem to be involved in the formation of the central pattern generator. Their results show that loss of V1 neurons early in development is not compensated for by any reconfiguration of the circuit — the mutant mice that chronically lost V1 neurons at early stages exhibit the same slow locomotor rhythm as that observed when neurons are acutely silenced after the circuit is formed.

Nevertheless, studies of rewiring during development are complex; they require a rather dynamic approach because exclusive examination of the final end point may not reveal intermediate remodelling processes that are crucial to the resulting network. In this regard, it would be interesting to test what happens to the circuit and locomotor activity if neurons are silenced at different times during development. Formation of circuits is likely to depend on both genetically driven and activity-shaped processes. We need to keep our minds open if we are to understand the interplay of these driving forces.

Developmental biologists and physiologists have tended to approach the problem of circuit formation and function from very different angles, without much dialogue between them. But genetic tools are now available to reconcile the results from these two disciplines and establish a fruitful interaction. Information about signatures of gene-regulatory factors that determine the fates of neuronal populations can now serve as tools with which to elucidate the functions of a given set of neurons. The work presented by Gosgnach et al. sets up a firm bridge across the river that has so far divided developmental biology and physiology.