We recently reported that somatosensory inputs from large areas of the body evoked long-latency, long-lasting depressions of spike activity in cerebellar Golgi cells (Holtzman et al. 2003) in anaesthetized rats. Here we ask which ascending pathways may contribute to these responses.

In deeply anaesthetized rats (Urethane, 1-1.2 g/kg, I.P.) single unit recordings were made from Golgi cells while various peripheral nerves were stimulated electrically. Hindlimb muscle nerves (Gastrocnemius and Hamstring), mixed nerves (Tibial at the ankle) and the principally cutaneous Sural nerve were exposed and mounted on pairs of silver wires in tubular electrodes, mounted in a mineral oil pool which was maintained at 37° C. The effectiveness of the stimulation (in terms of threshold of the most excitable fibres, T) was monitored by recording evoked volleys from the surface of the lumbosacral spinal cord (L2-3), which was exposed by laminectomy. At the end of the experiment all animals were killed by overdose of anaesthetic.

Nerves that contain a large proportion of cutaneous fibres (Sural and Tibial) evoked effects in most of the neurons tested: long-lasting depression of Golgi cell firing was elicited by stimulation at intensities ▓ge│ 2T in 6/6 cells from sural and in 12/13 cells from tibial. In contrast, stimulation of the muscle nerves (Gastrocnemius and Hamstring) never evoked responses at intensities ▓le│ 2T. At higher stimulus intensities (▓ge│ 5T) responses were evoked in some neurons (9/15 and 5/8 cells, from Gastrocnemius and Hamstring, respectively). High (electrical) threshold cutaneous fibres are also likely to contribute to this Golgi cell response, since the firing rate depression was greater in response to stronger stimuli.

These results provide evidence that low threshold cutaneous afferents can generate long-lasting Golgi cell firing depressions, but that low threshold (group I) muscle afferents are unlikely to. The classical ascending spinocerebellar pathways (dorsal and ventral spinocerebellar pathways) are therefore unlikely to have mediated the Golgi cell responses. Higher (electrical) threshold fibres from cutaneous and muscle nerves also elicit these responses. Golgi cells are inhibitory interneurons that, by inhibiting granule cells, modulate transmission from mossy fibre input to the cerebellar cortex. A system that selectively depresses Golgi, but not Purkinje cell firing in response to cutaneous inputs from a wide region of the body is likely to have a substantial impact on this transmission. These data are not readily compatible with the conventional view that Golgi cells mediate feedback control of transmission through the mossy fibre system.

We thank the Physiological Society for a vacation studentship award to C-L Phuah.