We have investigated the patterns of activation of epaxial musculature during both swimming and overground stepping in an adult newt (Pleurodeles waltl) with the use of electromyographic (EMG) recordings from different sites of the myomeric muscle dorsalis trunci along the body axis. The locomotor patterns of some limb muscles have also been investigated. During swimming, the epaxial myomeres are rhythmically active, with a strict alternation between opposite myomeres located at the same longitudinal site. The pattern of intersegmental coordination consists of three successively initiated waves of EMG activity passing posteriorly along the anterior trunk, the midtrunk, and the posterior trunk, respectively. Swimming is also characterized by a tonic activation of forelimb (dorsalis scapulae and extensor ulnae) and hindlimb (puboischiotibialis and puboischiofemoralis internus) muscles and a rhythmic activation of muscles (latissimus dorsi and caudofemoralis) acting both on limb and body axis. The latter matched the activation pattern of epaxial myomeres at the similar vertebral level. During overground stepping, the midtrunk myomeres express single synchronous bursts whereas the myomeres of the anterior trunk and those of the posterior trunk display a double bursting pattern in the form of two waves of EMG activity propagating in opposite directions. During overground stepping, the limb muscles and muscles acting on both limb and body axis were found to be rhythmically active and usually displayed a double bursting pattern. The main conclusion of this investigation is that the patterns of intersegmental coordination during both swimming and overground stepping in the adult newt are related to the presence of limbs and that they can be considered as hybrid lampreylike patterns. Thus it is hypothesized that, in newt, a chain of coupled segmental oscillatory networks, similar to that which constitutes the central pattern generator (CPG) for swimming in the lamprey, can account for both trunk motor patterns if it is influenced by limb CPGs in a way depending on the locomotor mode. During swimming, the segmental networks located close to the girdles receive extra tonic excitation coming from the limb CPGs, whereas during stepping, the axial CPGs are entrained to some extent by the limb oscillators.

An in vitro brain stem-spinal cord preparation from an adult urodele (Pleurodeles waltl) was developed in which two fictive rhythmic motor patterns were evoked by bath application of N-methyl-D-aspartate (NMDA; 2.5-10 microM) with D-serine (10 microM). Both motor patterns displayed left-right alternation. The first pattern was characterized by cycle periods ranging between 2.4 and 9. 0 s (4.9 +/- 1.2 s, mean +/- SD) and a rostrocaudal propagation of the activity in consecutive ventral roots. The second pattern displayed longer cycle periods (8.1-28.3 s; 14.2 +/- 3.6 s) with a caudorostral propagation. The two patterns were inducible after a spinal transection at the first segment. Preliminary experiments on small pieces of spinal cord further suggested that the ability for rhythm generation is distributed along the spinal cord of this preparation. This study shows that the in vitro brain stem-spinal cord preparation from Pleurodeles waltl may be a useful model to study the mechanisms underlying the different axial motor patterns and the flexibility of the neural networks involved.

The ability of mammalian spinal cords to generate rhythmic motor behavior in nonlimb moving segments was examined in isolated spinal cords of neonatal rats. Stimulation of sacrocaudal afferents (SCA) induced alternating left-right bursts in lumbosacral efferents and in tail muscles. On each side of the tail, flexors, extensors, and abductors were coactive during each cycle of activity. This rhythm originated mainly in the sacrocaudal region because it persisted in sacrocaudal segments after surgical removal of the thoracolumbar cord. Sacrocaudal commissural pathways were sufficient to maintain the left-right alternation of lumbar efferent bursts, because their timing was unaltered after a complete thoracolumbar hemisection. The lumbar rhythm originated in part from sacrocaudal activity ascending in lateral and ventrolateral funiculi, because efferent bursts in rostral lumbar segments were nearly abolished on a particular side by lesions of the lateral quadrant of the cord at the L(4)-L(5) junction. Intracellular recordings from S(2)-S(3) motoneurons, obtained during the rhythm, revealed the presence of phasic oscillations of membrane potential superimposed on a tonic depolarization. Bursts of spikes occurred on the depolarizing phases of the oscillation. Between these bursts the membrane input conductance increased, and hyperpolarizing drive potentials were revealed. The inhibitory drive and the decreased input resistance coincided with contralateral efferent bursts, suggesting that crossed pathways controlled it. Our studies indicate that pattern generators are not restricted to limb-moving spinal segments and suggest that regional specializations of pattern-generating circuitry and their associated interneurons are responsible for the different motor patterns produced by the mammalian spinal cord.

The rhythmogenic capacity of the tail-innervating segments (L4-Co3) of the spinal cord was studied in isolated spinal cord and tail-spinal cord preparations of neonatal rats. Bath-applied serotonin/N-methyl-D-aspartate (NMDA) failed to produce a robust sacrococcygeal rhythmicity following midlumbar transection of the spinal cord. By contrast, a regular alternating left-right rhythm could be induced in the sacrococcygeal segments by application of noradrenaline (NA) or NA and NMDA before and after midlumbar transection of the cord. This rhythm was accelerated with the concentration of NMDA and was blocked by alpha1 or alpha2 adrenoceptor antagonists. The efferent bursts induced by NA/NMDA were accompanied by rhythmic tail movements produced by alternating activation of the left and right tail muscles and by coactivation of flexors, extensors, and abductors on a given side of the tail. This coactivation implies that reciprocal inhibitory pathways were not activated during the rhythm. Lesion experiments revealed that the rhythmogenic circuitry is distributed along all or most of the sacrococcygeal segments. The NA/NMDA-induced rhythm persisted in the isolated sacrococcygeal (S1-Co3), sacral (S1-S4), coccygeal (Co1-Co3), and smaller isolated regions of the sacrococcygeal cord. The rhythm also could be maintained in longitudinally split sacrococcygeal hemicords in which flexor, extensor, and abductor motoneurons are coactivated. This finding indicates that neither left/right nor flexor/extensor inhibitory interactions are required for rhythmogenesis in the sacrococcygeal cord. A slow rhythm lacking the alternating left-right pattern was induced by NA/NMDA in tail-innervating caudal lumbar segments of isolated L4-Co3 preparations. This rhythm was independent of the concurrent sacrococcygeal rhythm and the activity pattern of the tail musculature and it does not seem to contribute to rhythmic tail movements under these conditions. Comparative studies of the rhythm produced in the isolated caudal lumbar, sacrococcygeal cord, and caudal thoracic-rostral lumbar segments revealed that the S1-Co3 rhythm was faster than the L4-L6 pattern and slower than the T6-L3 rhythm. It is suggested that the caudal lumbar and sacrococcygeal segments of the cord are normally driven by the faster rostral lumbar central pattern generators. The relevance of the findings described above to pattern generation in the mammalian spinal cord is discussed.