Motor-unit recruitment in self-reinnervated muscle.

1. Recruitment order of motor units in self-reinnervated medial gastrocnemius (MG) muscles was studied in decerebrate cats 16 mo after surgical reunion of the cut MG nerve. Pairs of MG motor units were isolated by dual microelectrode penetration of ventral roots to measure their recruitment sequence during cutaneous reflexes in relation to their physiological properties. 2. Physiological properties of reconstituted motor units appeared normal, as expected. Also normal were the relationships among these properties: twitch and tetanic tension tended to increase with axonal conduction velocity and decrease with twitch contraction time. A small fraction of motor units (10/116) in reinnervated muscles produced either no measurable tension or unusually large amounts of tension compared with controls. This was the only distinct feature of the sample of reconstituted units. 3. In muscles reinnervated after nerve section, stretch was notably ineffective in eliciting reflex contraction of MG muscles or their constituent motor units (only 5/116 units). Incomplete recovery from nerve section was probably the cause of this impairment, because stretch reflexes were readily evoked in adjacent untreated muscles and in one reinnervated MG muscle that was studied 16 mo after nerve crush. In contrast with the ineffectiveness of muscle stretch, sural nerve stimulation succeeded in recruiting 49/116 units, a proportion fairly typical of normal MG muscles. 4. The contractions of the first unit recruited in cutaneous reflexes tended to be slower and less forceful than those of the other unit in a pair. By these measures, recruitment obeyed the size principle. This recruitment order with respect to unit contractile properties was not significantly different (P > 0.05) between untreated and reinnervated muscles but was significantly (P < 0.005) different from random order in both groups. The same recruitment pattern was observed for pairs of motor units sampled from the muscle reinnervated after nerve crush, whether units were recruited by muscle stretch or sural nerve stimulation. 5. The usual tendency for motor units with slower conduction velocity (CV) to be recruited in sural nerve reflexes before those with faster CV was not strong in reinnervated muscles. After nerve section the proportion of units exhibiting the usual recruitment pattern was not significantly different (P > 0.05) from a random pattern for CV. 6. The central finding is that the normal recruitment patterns recover from nerve injury in a muscle that is reinnervated by its original nerve. By contrast, stretch reflexes do not recover well from nerve section, and this deficiency may contribute to motor disability.