Abstracts

Intractable epilepsy that remains resistant to drug treatment is a significant public health problem with few alternative therapeutic options besides surgery. Our laboratory has been exploring protocols by which electrical stimulation applied directly to a seizure focus could modulate an epileptic seizure and, ideally, terminate it. We have been investigating the use of control feedback as one such technique for supplying the necessary algorithms for the stimulation in order to control the output of the seizure focus. Specifically, we have explored the use of PID (proportional, integral, derivative) control paradigms. One desired characteristic of PID controllers is that they permit reduction in tracking error of a system without increasing overall instability. Bipolar electrodes with a cannula stem are stereotactically implanted in the CA3 region of hippocampus in one hemisphere of male Sprague-Dawley rats (300-400gm) and anchored in place with small screws and dental epoxy. Animals remain fully anesthetized (ketamine+xylazine) throughout the experiment. Intracranial EEG signals are measured from one of the bipolar electrodes by amplification and input via A/D converter to a PC. After a baseline recording (30min), penicillin-G (500IU; 20[mu]l) is injected through the cannula resulting in recording of strong persistent episodic electrical seizure activity that can last from 10-40s. Because the electrodes and cannula are inserted as a single tight unit, recording and stimulating electrodes are extremely proximal to the seizure focus. The recorded electrical activity by one electrode is used as the template (an error signal) for insertion into the control algorithm. Different values of controller characteristics including complex gain and delay are being examined. In addition, self-tuning techniques to adapt the controller to a dynamic change in the system characteristics is being implemented. Current- or voltage-based stimulation by the other electrode is then applied. Different combinations of controller parameter values of this PID feedback controller are currently being tested for their efficacy at modulating seizures. At present, short control epochs ([underline][lt][/underline] 10sec) can transiently quiet synchronized bursting although a more thorough analysis of controller parameter values may significantly improve performance. We present early results of a technique in which stimulation applied to a seizure focus is modulated by an adaptive feedback model in order to manipulate electrical bursting in the CA3 region of rat hippocampus. Because of the complexity of this control paradigm, many different control algorithms are being investigated for efficacy. If successful, this technique may provide an alternative therapy for seizure control. (Supported by funding from NIH (NINDS 046375) to DJM.)