Spinal cord segments containing key elements of the central pattern generators for three forms of scratch reflex in the turtle.

The immobilized, low-spinal turtle produces 3 forms of the fictive scratch reflex in response to tactile stimulation of specific sites on its body surface (Robertson et al., 1985). We used complete transections of the spinal cord at different rostrocaudal levels to reveal the minimum length of spinal cord sufficient to produce each scratch form. Additional transections revealed the progressive loss of elements of the motor pattern and the eventual loss of rhythmogenesis. We have identified, therefore, spinal cord segments containing key elements of each scratch form's central pattern generator (CPG). The turtle spinal cord consists of 8 cervical segments (C1-C8), 10 dorsal segments (D1-D10), 2 sacral segments (S1, S2) and about 16 caudal segments (Ca1-Ca16; Kusuma et al., 1979). The cell bodies of motor neurons innervating the hindlimb muscles are located in the hindlimb enlargement, segments D8-S2 (Ruigrok and Crowe, 1984). The receptive field for the rostral scratch is innervated by segments D3-D6; the pocket scratch receptive field is innervated by segments D6-D8; the caudal scratch receptive field is innervated by segments S2, Ca1, and more caudal segments (Mortin and Stein, 1985). A rostral scratch motor pattern could be produced with as few as 5 or 6 segments, i.e., segments D5-D9 or D3-D8. The anterior 3 segments of the hindlimb enlargement, D8-D10, could produce a pocket scratch motor pattern. A single segment, either D7 or D8, is capable of rhythmogenesis in response to stimulation of sites in its part of the pocket receptive field. A caudal scratch motor pattern could be produced by D8-End (the hindlimb enlargement and more caudal segments). The posterior 40-80% of the hindlimb enlargement is not necessary for the production of a rostral or pocket motor pattern. The anterior segment of the enlargement is necessary for the production of a normal caudal scratch motor pattern. Key elements of the CPG for each of the 3 scratch forms reside in segments D7-D10. The pattern-generating capacity of the anterior half of the hindlimb enlargement is greater than the posterior half; such an asymmetric distribution of pattern-generating elements in the enlargement of the spinal cord has been described for cat scratching (Deliagina et al., 1983). These results are consistent with the hypothesis that the CPGs producing different motor patterns for the hindlimb share neuronal elements (Grillner, 1981; Robertson et al., 1985; Currie and Stein, 1988, 1989).