Abstracts

Clinical evidence that cortical stimulation might stop seizures has led to an ongoing multicentre trial of a responsive neurostimulation system for intractable epilepsy. Little is known however about the effects of stimulation on epileptiform activity. A large scale neural network simulation with realistic cortical architecture has been undertaken to investigate stimulation effects. This study investigates stimulation electrode geometry, frequency, and pulse profile. The validity of the model is supported by its representation of the time-frequency characteristics of epileptiform activity., The model consists of an approximately 1 mm X 1 mm region of simulated cortex (16,384 neurons), and includes seven neuron classes organized by cortical layer, synaptic properties, and electrophysiological characteristics. The cell dynamics are governed by a modified version of the Hodgkin-Huxley equations. Axonal connections are patterned after histological data and published models of local cortical wiring. Stimulation induced action potentials take place at the axon initial segments (AIS), according to threshold requirements on the applied electric field distribution. The calculations are performed on a 16 node distributed 64-bit processor system., Clear differences in seizure evolution are presented for stimulated versus control networks. Figure 1 (A) demonstrates the model Layer II/III pyramidal cell activity under the electrode during stimulation, (B) shows the decrease in network activity after the stimulation. Data is provided for frequency dependent stimulation effects as well as spatial effects due to electrode position or height above the cortex. Data for stimulation effects from a dipole electrode system are also presented, which demonstrate clear orientation dependent effects if there are strong underlying connection asymmetries. The relative roles of axon initial segment versus distal axon branch induced action potentials are presented, as well as stimulation induced changes in intracellular calcium dynamics., This computational modeling effort makes predictive assessments of stimulation characteristics used to alter or stop seizure activity. In general, there is a mild increasing efficacy of stimulation as a function of frequency between 60 to 120 Hz. Efficacy from 120 to 200 Hz remains constant. The effects are highly local in the context of the model, being strongest under the stimulation electrode.[figure1], (Supported by: WSA was supported by the Epilepsy Foundation via a Research and Training Fellowship for Clinicians.)