Development Of A System For Real-time Seizure Detection And Closed-loop Deep Brain Stimulation

Recently, deep brain stimulation has received arising attention in treating epilepsy. However, during current researches most of the stimulating modes are open-loop, which cannot emit stimulating pulses automatically according to the status of patients. Moreover, there are many disadvantages on feasibility and universality of nowadays’ existing closed-loop electrical stimulators. As a result, it is of great significance for researches of epilepsy treatment with electrical stimulation to design a closed-loop system, which can detect epileptic seizures automatically and emit electrical stimuli accordingly.In the thesis, we designed such a system using the Lab VIEW software, NI USB-6251data acquisition card, amplifier and stimulator. The system could detect epileptic seizures and emit electrical stimuli to the deep brain, and was designed mainly to deal with local field potentials of the rat epileptic models. The main work and methods involved in this thesis were:(1) Development of the closed-loop control program using LabVIEW, so that the system can accomplish the function of real-time detection of epileptic seizures, and output of electrical stimuli according to predefined stimulating parameters;(2) Design of a new algorithm to detect epileptic seizures. The algorithm must be able to work fast while maintaining a relatively high accuracy;(3) Making of rat epileptic models, to which we delivered the stimulating pulses of different parameters and assessed results of suppressing epileptic seizures.The main results of the thesis were summarized as follows:(1) We developed a closed-loop control program with LabVIEW. The program was real-time, flexible, and could control the NI USB-6251data acquisition card to collect brain signals and send out stimulation triggering signals. The trigging signal would in turn lead to the output of stimulating pulses, and the time delay was only around0.12s. The system was real-time, flexible and universal.(2) By combining the signal’s amplitude, slope and coastline index, We developed a new algorithm to detect epileptic seizures. The new algorithm was fast, and reached a good detection rate. Experimental test results of rat epileptic models in the hippocampal region showed that, with this algorithm, the system was able to detect epileptic seizures with an accuracy rate91.3%and false rate8%.(3) On-line experimental test results of rat epileptic models showed that, the high frequency electrical stimulation can suppress seizures when the rat was experiencing the seizure period. But it might also strengthen seizure activities when the stimulation happened in a rat’s inter-seizure period. Furthermore, stimulating pulses lasting for10s had a better suppression effect on seizures than their5s counterparts, but no obvious suppression was observed when the pulses were extended for more than1min.In conclusion, the closed-loop electrical stimulation system we developed in this thesis can detect epileptic seizures in local field potentials automatically, and send out stimulating pulses of certain frequency, amplitude and duration as customers had set. Experimental test results of rat epileptic models showed that the high frequency electrical stimulation can suppress seizures. The system was real-time, flexible and universal, and could be a useful tool in researches of treating epilepsy with deep brain stimulation.