Neurophysiology of the subthalamic nucleus

Possibly as many as half the neurones in the STN have axon collaterals that branch off from the main axon and re-innervate the nucleus. This suggests that rather than working autonomously as was previously thought, the neurones of the STN can operate together as a network. Computer models of the STN showed that the level of interconnectivity within the STN would be huge, even if each axon collateral only contacted a small number of the total neurones with dendritic fields that overlapped with it. A network model showed that such a system was capable of switch-like behaviour. At low levels of activity the neurones would act autonomously. However, excitatory inputs could increase the degree of non-synchronous correlation between the activity of neurones in the STN leading them all to enter a high activity state. A single cell model was then developed in order to look at how this high activity state could be terminated. An interesting problem arose in the construction of this model; no known kinetics for the voltage-gated sodium and potassium channels could replicate the high frequency (500Hz) firing rates that are obtained by STN neurones

Intracellular recordings were made in vitro to investigate the mechanisms underlying high-frequency firing in the STN. Using a two-pulse protocol the speed of recovery from inactivation was measured giving an estimate of the inactivation characteristics of the ion channels in these neurones. These experiments showed that the neurones have very slow inactivation kinetics suggesting that STN neurones may have a much shortened refractory period, enabling high frequency firing. Such a mode of operation requires a large, fast potassium current. A potential candidate for this current is the Kv3.1 potassium channel, which is strongly expressed by STN neurones.

Extracellular recordings were used to look for evidence of functional interconnections between cells within the STN. These experiments showed that blocking any interconnections with a glutamatergic antagonist had no effect on the resting firing pattern or rate of STN neurones. However, when the neurones were depolarised using increased levels of potassium in the perfusing solution, the normally regular firing pattern of the neurones was disrupted and became irregular. The glutamatergic antagonist attenuated this disruption showing that it was at least partially mediated through glutamatergic synapses, the best candidate for which are those at the interconnections between the STN neurones.

Having investigated high frequency firing in the STN, and how such increased levels of activity could influence co-ordinated firing within the STN, the effects of one of the STNs targets was assessed. Lesions of the globus pallidus have been shown to create a chronic increase in the levels of STN activity in vivo. At three and six weeks after such lesions a marked reduction was found in the number of neurones in the substantia nigra that stained positive for tyrosine hydroxylase (marking them as dopaminergic cells). These data provide evidence supporting the excitotoxic hypothesis for the progressive loss of dopaminergic cells that is seen in Parkinson's Disease.