Research Of Preictal State Characteristics Perception And Automatic Drug Release Pretreatment Model

Epilepsy is a chronic disease because of nervous system disorders, with the characteristics of recurrence and difficult to cure. The perniciousness of its attack brought great danger to people with epilepsy. Therefore, study the method of epileptic seizures prediction, and give the people with epilepsy automatic preventive treatment with a non-invasive manner is very meaningful. The epilepsy pre-treatment model reported in the literature home and abroad is only limited to the methods assessment of the epilepsy prediction, and there is nearly no research on micro-pump drug delivery model of epilepsy, which makes the epilepsy treatment process hindered. In this paper, the pre-ictal state prediction model of epilepsy in wavelet domain was put forward, using the nonlinear dynamics methods, and based on the pre-ictal state perceptual system and automatic drug delivery system. In this model, when epilepsy patients at the onset of pre-state, the pre-ictal state perceptual system will trigger the early warning device to issue alerts and micro-pump to release drug, in order to achieve the purpose of non-invasive treatment of patients.In the pre-ictal state prediction model, the nonlinear dynamic characteristics of person’s seizure EEG include of the normal, pre-ictal and ictal state were extracted, using the methods of Lyapunov exponent, approximate entropy, power spectal entropy and the Kolmogorov entropy. Then conducted the comprehensive analysis and evaluation of the results. The results indicated that four methods can effectively predict pre-ictal state and approximate entropy show best performance.The thesis designed the micro-pump with a double membrane structure in order to make the release flow controlled, and the static analysis and modal analysis were also conducted. Static analysis focuses on the film deformation of influence factor which exert an influence on the micro-pump flow, and include, the factors include of membrane thickness, radius and drive voltage and other parameters; the modal shape analysis of the pump membrane was finished, the modal analysis determined the natural frequency under some micro-pump structure parameters and the vibration form of micro-pump membrane under maximum operating efficiency. The pre-ictal state prediction and automatic release pretreatment modal was simulated by LabVIEW Virtual Instrument, with80%sensitivity and90.9%specificity. The application experiments verified the validity of the model through prediction of epileptic EEG and the model forecast data provide the foundation for clinical diagnosis and brain science research.The framework of the system is open, the signal feature extraction methods could be added to optimize the micro-pump parametric model and to perfect the pre-treatment model of epilepsy. The system has broad application prospects in the field of disease diagnosis and monitoring, brain science and brain cognition.