GENERATION OF RHYTHMIC ACTIVITY IN THE SPINAL CORD

A. Tscherter and J. Streit
Department of Physiology, University of Bern, Switzerland

During locomotion a neuronal network in the spinal cord generates rhythmic activity. This activity is directed to motoneurons, which activate the limbs in a rhythmic way. Until now, how this rhythmic activity is generated remains unclear. In order to investigate this neuronal network, one has to record from many neurons simultaneously. Therefore extracellular multisite recording was performed on fetal rat slice cultures grown on multielectrode arrays.
In these organotypic cultures a spontaneous rhythmic bursting of neurons could be induced by the disinhibition of the culture. This could be obtained by the application of bicuculline and strychnine, the antagonists of the GABA A and glycine receptors. The bursts were synchronized between both sides of the culture. The majority of bursts started in the ventral part of the slice and then propagated to more medial and dorsal parts in 40-80 ms. The application of APV, an NMDA receptor antagonist, and CNQX, a non-NMDA receptor antagonist, completely abolished the bursts. Therefore this rhythmic activity is mainly mediated by glutamatergic synaptic transmission. Between the bursts asynchronous activity at low rates was recorded. The rate of such activity mostly increased during the last second preceding the bursts.
The spike rate in the whole network decreased towards the end of the bursts, indicating a decrease of the network excitability. In addition, a correlation between burst duration and the previous interval could usually be observed, suggesting that the burst duration was dependent upon the recovery of the network from the previous burst. To analyze this idea we stimulated the network at different frequencies and indeed, extracellular stimulation by one electrode in the ventral part of the slice suppressed the spontaneous bursting, provided the stimulation frequency was higher than the spontaneous frequency. Under stimulation the burst duration was inversely proportional to the stimulation frequency and the time to peak of the network activity at the beginning of the bursts decreased with lower frequency. This means, it took more time to activate the whole network at higher frequencies than at lower frequencies.
These findings suggest the following hypothesis: The excitability of the neuronal network decreases during the burst, finally leading to cessation of the burst. The neuronal network then slowly recovers from this decrease in excitability and the increase of spontaneous activity induces a new burst.