Exciting times in the tadpole spinal cord

Abstract

Given the ‘relative’ simplicity of the spinal cord compared to other areas of the CNS spinal locomotor networks should have been among the easiest vertebrate networks to characterise. Brown provided the first evidence for a spinal central pattern-generating network almost a century ago (Brown, 1911). He also provided insight into their basic organisation that has formed the foundation for all subsequent analyses of spinal networks. However, the characterisation of these networks has not been straightforward: given the difficulties of understanding much simpler networks this should probably not have come as a surprise. Understanding any network at the very least requires information on the network organisation (the neurons that the network contains and their direct synaptic interactions), as well as the functional properties of network cells and synapses. While there are still significant gaps in the understanding of mammalian locomotor networks, there is some insight into locomotor networks in lower vertebrates, the lamprey and tadpole. The organisation of the spinal cord locomotor network in the lamprey is claimed to be fully characterised (e.g. Grillner et al. 2005 for a recent example): these claims are widely accepted, but they ignore uncertainty over the identification of network neurons and the significant gaps in our knowledge of network connectivity (see Parker, 2006). While it has traditionally been less celebrated, insight into the tadpole locomotor network has increased dramatically over the last several years. Li, Soffe and Roberts in Bristol have developed a tadpole preparation that has allowed paired recordings to be made routinely from identified classes of spinal cord neurons. As a result insight into the organization and functional properties of the tadpole locomotor network has increased markedly. While not widely regarded as such, it has arguably become the best-understood spinal cord locomotor network in terms of network organization and functional properties. The study by Li et al. (2009) in this issue of The Journal of Physiology has continued their characterisation of the locomotor network by examining electrical coupling between excitatory network interneurons (dINs) in the hindbrain and spinal cord, and the significance of this coupling during swimming. They describe a very high (but not absolute) degree of exclusivity in electrical coupling between the dINs through the use of pairwise recordings from identified cell types. It should not need emphasizing (but still does) that while it is difficult and time-consuming, detailed electrophysiological analyses of this sort are essential to any genuine characterization of networks. The data also highlight the impossibility of extrapolating within or between cell populations. The authors suggest that the electrical coupling between dINs contributes to the high degree of reliability of spiking in these neurons during swimming, and move on to try and link the effects of this coupling to the network output. This link is generally difficult to make, and where claims have been made they inevitably rest on assumptions and extrapolations rather than direct causal links. Here, the authors have tried to make this link by blocking the electrical connections between dINs using the gap junction blocker 18-β-glycyrrhetinic acid (18-β-GA). They found that 18-β-GA decreased the reliability of dIN spiking and the synchronization across the dIN population during swimming. In addition, it drastically shortened the duration of swimming episodes. They conclude that electrical coupling between dINs underlies the reliability and synchronization of spiking within the dIN population and the ability of these neurons to sustain swimming episodes. It thus provides a potential mechanism for sustaining rhythmic activity in response to a brief initial stimulus. These claims inevitably carry certain assumptions: there is the possibility of an effect of 18-β-GA on electrical connections between MNs, and the possibility of non-specific effects of gap junction blockers, a feature that complicates the use of these drugs in network analyses. The authors went some way to addressing these issues. Firstly, from their previous analyses electrical synapses are found between motor neurons (Perrins & Roberts, 1995) but there is no evidence for inputs to the dINs or other premotor interneurons, and the sustained activity in hindbrain dINs is not affected by removing the spinal cord (and thus motor neurons; Li et al. 2006), both features suggesting that the effects of 18-β-GA on network activity were not due to effects on electrical connections made by motor neurons. Secondly, they chose 18-β-GA from four gap junction blockers based on analyses of non-specific cellular and synaptic effects of these drugs. 18-β-GA was chosen because of its lack of non-specific effects on the properties they studied (although other effects cannot be ruled out). These features strengthen the link between the coupling between dINs and the changes in the network output, and thus make a reasonable link between cellular properties and network behaviour. This study thus continues the impressive series of analyses in this system.