The basal ganglia are a collection of interconnected subcortical nuclei which have been implicated inmotor, cognitive and limbic functions. The subthalamic nucleus is the sole excitatory structure within the basal ganglia. Given its central position influencingmany basal ganglia nuclei, it is likely to play an important role in the processing that is performed by the basal ganglia. In this thesis a theoretical analysis of the subthalamic nucleus is presented. In order to explore the multiple facets of processing that may be occurring, models that are designed to capture aspects of the subthalamic nucleus at different levels are developed. These include anatomical, network processing and single neuron multi–compartmental models. Through the integration of the results obtained from these models a new and coherent view of the processing of the subthalamic nucleus is presented. It is predicted that the subthalamic nucleus be considered as a massively connected excitatory network. Two distinct modes of asymptotic behaviour exist in such a network: a low resting state and a high self–sustained state. The single neuron multi– compartmental model demonstrates that the calcium T–type channel is the primary determinant of characteristic neuron behaviour. Such behaviour includes a slowaction potential, initial spike clustering, and a post-response quiescence. The network and single neuron results taken togetherprovide an intrinsicmechanismfor termination of uniform high activity generated by the excitatory network. It is therefore predicted that large regions of the subthalamic nucleus respond uniformly to stimuli, in the form of a pulse of activity with a sharp rise and fall. In addition, the single neuron model indicates that pulses will occur in pairs. It is proposedthat the subthalamic nucleus acts as a “braking mechanism”. It can induce, via intermediate structures, awide-spread pulse of inhibition on basal ganglia target nuclei. Furthermore, the sequence of two pulses can generate a window of disinhibition over the basal ganglia targets. The width of this time window may be under direct striatal control. Variable interpulse duration implies a role for the subthalamic nucleus in temporal processing.