When we decide to go down the corridor and get a coffee we have several options. We can make our way at a slow leisurely pace (lento), walk quickly and purposefully (andante) or even engage a full other gait and start running down the corridor (allegro). The understanding of locomotion and its different speeds has traditionally focused on understanding networks of neurons called central pattern generators (CPGs) in the spinal cord, which are involved in generating and implementing the cyclic rhythmicity involved with locomotion: left, right, left, right … It has proven much harder to understand what is upstream of these spinal CPGs: how does the brain initiate locomotion, how does it control its speed and how does it relay this information down the spinal cord to the central patter generators which generate locomotion.
In a paper appearing this week in Neuron, Severi et al. investigate this question in larval zebrafish. Larvae want to stay in the same place with respect to their visual environment, so if you show them visual motion they will swim in the direction of the motion. The faster the motion, the quicker the larvae will swim. Severi et al. use this robust innate behavior, known as the optomotor response, to elicit swimming of different speeds in larvae.
A detailed quantitative behavioral characterization shows that larvae change many behavioral parameters when they try to swim faster. These parameters can be changed independently: they can move their tails for longer periods of time, they can move it more vigorously, they can take less time in between swim bouts or they can even recruit a total different type of bout (analogous to switching to a different gait such as humans with running or horses galloping). How does the brain tell the spinal cord which pattern to implement, and how can it change only particular aspects of the behavior?
In the larval zebrafish there is a small group (~300) of neurons that project from the brain down the spinal cord. In particular, there is a group of ~20 neurons with cell bodies in the midbrain called the nMLF (nucleus of the medial longitudinal fasciculus) that send axons down the spinal cord and that was previously identified in the Engert lab as being active during swimming (Orger et al. Nature Neuroscience 2008).
Severi et al. show that electrically stimulating this nucleus can induce swimming in larvae and that the strength of locomotion correlates with the duration and vigor of the swimming. They also show that selectively killing a group of 4 cells in the nMLF, which are identifiable from fish to fish, has severe impact on the ability of larvae to modulate their swimming speed.
Using two-photon functional calcium imaging in behaving larvae, they show that when swimming faster the nMLF does not recruit more neurons. Instead, neurons that were already active are simply more active. In particular, activity in two identifiable cells (MeLr and MeLc) is correlated with bout duration and tail beat frequency respectively. How can increased activity in these neurons modulate these parameters differentially? The explanation is yet to be found, although as shown by Wang and McLean (co-submission) it probably involves the different projection patterns of these neurons in the spinal cord. In the spinal cord there is a dorso-ventral gradient in the biophysical properties of neurons: ventral interneurons and motor neurons are easily recruited whereas dorsal ones require stronger input. Increased activity in the nMLF is likely to shift activity from the ventral to the dorsal spinal cord, resulting in more vigorous swims.
Overall, this work provides a valuable insight into the supra-spinal control of locomotion by attributing an important role to 4 identifiable cells in independently controlling two very specific behavioral parameters: bout duration and tail-beat frequency. By studying the inputs to these cells we hope, in future studies, to be able to shed light on how vertebrates “decide” to start moving and how they “choose” how to move.
Read more in Neuron (AOP) or download PDF