It has been suggested that heterosegmentally activated slow positive potentials (HSP), recorded from the spinal cord of rat and humans, are feedback inhibitory potentials. The present study was carried out to define ascending and descending spinal tracts and the sites of central nuclei involved in the production of these HSP, and the effects of ketamine on these central nuclei. The spinal cords in ketamine-anaesthetized rats were transected to determine the ascending and descending tracts involved in the production of hindpaw (HP) and forepaw (FP) HSP, respectively. Lesions of the brain at various levels were performed stereotactically during ketamine anaesthesia. Dorsal one-third resection of the cord at the T8-9 level did not affect HSP significantly, while contralateral lesion of the dorsal two-thirds of the cord decreased FP-HSP but not HP-HSP during ketamine. Bilateral transection of the ventral one-third of the cord abolished both HSP. Ablation of the cerebral cortex, cerebellum, thalamus, midbrain and pons did not affect HSP significantly. However, transection of the middle medulla decreased, while transection of the most caudal part of the medulla completely abolished both HSP. Ketamine decreased HSP even in the medulla-spinal cord preparation and the segmental slow positive wave in spinalized animals. In ketamine-anaesthetized rats, ascending and descending spinal tracts involved in the production of HP-HSP and FP-HSP are located bilaterally in the ventrolateral quadrant and in the contralateral lateral funiculus and ventrolateral quadrant, respectively. Principal central nuclei feeding back HSP might be situated diffusely in the medulla down to the caudal part. Ketamine is suggested to suppress these inhibitory feedback potentials predominantly at, and partly even below, the level of the medulla.