The beat goes on, and up and down. Focus on "bursts of information: coordinating interneurons encode multiple parameters of a periodic motor pattern".

The effective production of many motor behaviors requires the strict coordination of discrete motor pattern generating circuits (Hill et al. 2003). Locomotor behaviors usually require this coordination at several levels of the nervous system, within single limbs (interjoint coordination), across multiple limbs (intersegmental coordination), and between multiple central pattern generator (CPG) circuits (Büschges 2005). In most such systems, the strategies for achieving this coordination are poorly understood. Does a single CPG entrain all others in the system, what is the nature of the timing cues, how extensive are the interconnections between CPGs, are the interactions symmetric or directional, and what is the balance between centrally generated and peripherally originating (sensory) coordinating signals? The crayfish swimmeret system was one of the first examples of a centrally generated motor pattern (Hughes and Wiersma 1960; Ikeda and Wiersma 1964). In this issue of Journal of Neurophysiology (p. 850–861), Mulloney and Harness (2006) exploit this preparation in a new study of interactions between independent central pattern generators in the production of a coordinated functional motor pattern. The swimmerets are paired appendages located in each of four abdominal segments (A2–A5) and the rhythmic beating of the four pairs of swimmerets provides thrust for forward swimming. The swimmeret motor pattern consists of the strict antiphase activation of power and return stroke motor neurons with bilateral pairs beating in phase. The overall motor pattern has a fixed posterior to anterior phase relationship (phase difference 0.25) that is constant with changes in cycle period, producing a metachronal wave of beating along the body axis. This motor pattern is produced independently by CPG circuits present in each hemi-ganglion (Hughes and Wiersma 1960; Ikeda and Wiersma 1964; Murchison et al. 1993) and the isolated abdominal nerve cord can generate the motor pattern and maintain the strict phase relationship (Braun and Mulloney 1995; Mulloney 1997). The fundamental swimmeret CPG module consists of a small group of nonspiking local interneurons (Heitler and Pearson 1980; Paul and Mulloney 1985a,b) and intersegmental coordinating interneurons (Stein 1971) with local, independent CPGs in each hemi-ganglia. Coordination between the segmental oscillators is mediated by three bilateral pairs of interneurons that project to anterior (ASCE, ASCL) and posterior ganglia (DSC). These intersegmental neurons are necessary and sufficient to lock the intersegmental phase relationship and to establish the posterior-anterior sequence of activity (Namba and Mulloney 1999; Paul and Mulloney 1986; Tschuluun et al. 2001). Mulloney and Harness determined the nature of the signals conveyed by coordinating interneurons by relating their activity to the swimmeret motor pattern. Firing of these neurons is shown to be strongly correlated with the timing, duration and strength of motor neuron bursts. Each burst of spikes in an ASCE neuron signals the start and duration of a power-stroke (PS) motor neuron burst. The activity of the descending coordinating neuron, DSC, while poorly correlated with PS bursts, instead reports the same motor pattern parameters related to the activity of return-stroke (RS) motor neurons, but with slightly lower fidelity than ASCE neurons. These timing parameters, PS and RS burst start and duration, can provide obvious and perhaps essential, phase-specific information for coordination of a coupled CPG system. More interesting is the observation that the number of spikes in ACSE (DSC) bursts was highly correlated with the strength of the PS (RS) motor bursts, although the correlations were strongest when there were large variations in motor neuron burst strength. Given that the timing of ASCE and DSC bursts reflects the timing of PS and RS motor neuron bursts, what might be the purpose of also encoding a signal (spike number) that is proportional to motor neuron burst strength? In terrestrial locomotor systems, coordination of joints or limbs is very dependent on feedback from proprioceptors and load sensors (Akay et al. 2004; Bucher et al. 2003; Büschges 2005; Marder and Bucher 2001). Encoding motor neuron burst strength by the coordinating interneurons might therefore provide a similar, albeit internal, signal proportional to (intended) swimmeret force production to adjacent upstream and downstream CPG modules. The genesis of the posterior to anterior phase gradient does not appear to be reside in differences in excitability of individual ganglia (Mulloney 1997) and there is presently no satisfactory explanation for this functional characteristic. As reported here, simultaneous recording from ASCE and DSC neurons in A2–A5 revealed a posterior–anterior gradient in the burst duration, spikes/burst and spike frequency of these coordinating neurons. These differences in firing properties therefore provide evidence for the cellular basis of the posterioranterior sequence of the swimmeret motor pattern. This study provides new insights into the motor parameters used to coordinate a distributed motor pattern assembled from a set of independent pattern generation circuits. Simple ON-OFF timing cues may be inadequate in similar systems, given that a richer parameter set (motor neuron burst onset, duration, and strength) appears to be signaled in the swimmeret system. Address for reprint requests and other correspondence: Biological Sciences, Ohio University, Athens, Ohio (E:mail:[email protected]). J Neurophysiol 95: 589–590, 2006. 10.1152/jn.01121.2005.