Modification of epileptiform bursting using chaos control

Recently, attempts have been made to apply chaos control techniques to manipulate the electrical discharges in the brain (interictal bursts) that are characteristic of epilepsy. These techniques would offer the advantage of using small and relatively infrequent stimuli to revert a seizure. However, questions have since arisen as to whether these results were truly chaos control or simply demand pacing. We have previously demonstrated evidence - including unstable periodic orbit (UPO) detection - that such epileptiform bursting is chaotic. We have investigated the potential for chaos control algorithms to manipulate extracellular bursts in rat hippocampal slices exposed to high levels of potassium. Interburst intervals (IBIs) were measured in real time using a threshold-detection circuit, embedded into two-dimensional state space, and analyzed for the presence of UPOs. The detection of a period-1 UPO strongly suggested the presence of chaos in the data and is a prerequisite for most forms of chaos control. Evaluation of control was aided by distinguishing whether IBIs were stimulated or natural. We investigated the effect of control region size on control efficacy. Complications to obtaining control exist, including (1) intrinsic system noise, (2) large instabilities (Lyapunov exponent) of the UPO, (3) difficulties estimating stable manifolds, (4) nonstationarity, and (5) neuronal plasticity. We have examined methods for surmounting these obstacles, including blocking synaptic plasticity and dynamically tracking the fixed point location.