Single Unit Based Seizure Prediction.

描述（由申请人提供）：癫痫是最常见的神经系统疾病之一，影响全球5000万人，其中四分之一具有医学难治性病例。癫痫发作的不可预测性可能导致严重的身体伤害甚至死亡。因此，非常需要一种可靠的癫痫发作预测方法来保护癫痫患者免受生命线程事件的影响。此外，准确的癫痫发作预测将允许临床医生提供适时的药物输送或治疗性脑刺激。这种最佳时间干预的能力可以大大减少药物反应患者长期用药的副作用和毒性，并为难治性癫痫患者提供新的治疗方法。鉴于这一需要，美国癫痫学会已将癫痫发作预测确定为癫痫研究的首要任务。神经科学家，临床医生，数学家和生物工程师已经发展了致力于这一研究领域的合作。尽管癫痫发作预测研究取得了很大进展，但可靠的癫痫发作预测仍然是一项具有挑战性的任务。在过去的30年里，EEG一直被用于癫痫发作预测研究，但缺乏新的替代研究工具。EEG数据反映了弥漫性脑区中大量细胞的突触后电位的总和，因此缺乏时间和空间特异性。在这项拟议的研究中，我们将首次采用多通道，单单位记录技术，以记录多达80个单细胞在5个不同的脑区，已被证明是关键参与颞叶癫痫（TLE）在大鼠模型的TLE。记录单个单位的主要好处是它提供了高的时间和空间分辨率以及多维数据，这使我们能够在癫痫发作之前检测到大脑区域的细微变化。我们将应用这些信息与电池的单一单位和脑电图数据分析工具，试图检测神经活动的变化，预示着癫痫发作。在这项初步研究中，我们将离线分析数据，以建立一个可靠的癫痫发作预测算法。一旦建立了该算法，我们将应用真实的时间、前瞻性癫痫发作预测方法来检测真实的时间的癫痫发作，并且我们将联合收割机将该预测工具与深部脑刺激相结合，以防止癫痫发作的发展和进展。这项研究的长期目标是将动物实验结果转化为临床，用这种新方法治疗癫痫患者。 公共卫生相关性：准确的癫痫发作预测可以防止癫痫患者的身体伤害，并使临床医生能够及时给予药物或脑刺激治疗癫痫。本研究将发展一种新的多通道、单单位记录方法来预测大鼠癫痫模型的自发性发作。该项目的成功将有助于改善临床上的癫痫发作预测，并使数百万癫痫患者受益。PHS 398/2590（Rev.09/04）

DESCRIPTION (provided by applicant): Epilepsy is one of the most prevalent neurological disorders, affecting 50 million individuals worldwide, a quarter of which have medically intractable cases. The unpredictability of seizure onset can lead to severe bodily injury or even death. Thus a reliable seizure prediction method is greatly needed to protect epileptic patients from life threading incidents. Furthermore, accurate seizure prediction would allow clinicians to provide well-timed drug delivery or therapeutic brain stimulation. This ability to optimally time interventions could substantially reduce the side effects and toxicity of long term medications for medically responsive patients and provide a novel treatment for patients with intractable epilepsy. In light of this need, the American Epilepsy Society has identified seizure predication as a primary priority for epilepsy research. Neuroscientists, clinicians, mathematicians and bioengineers have developed collaborations devoted to this research field. In spite of great progress in seizure prediction research, reliable seizure prediction remains a challenging task. EEG has been employed for the last 3 decades of seizure prediction research, whilst novel alternative research tools are lacking. EEG data reflect the summation of postsynaptic potentials across a large population of cells in diffuse brain areas and thus lack temporal and spatial specificity. In this proposed study, we will for the first time employ a multiple channel, single unit recording technique to record up to 80 single cells in 5 different brain areas that have been shown to be critically involved in temporal lobe epilepsy (TLE) in rat model of TLE. The major benefits of recording single units are that it provides high temporal and spatial resolution and multiple dimensional data, which allow us to detect subtle changes in brain areas before seizure onset. We will apply this information with a battery of single unit and EEG data analytic tools to try to detect the neural activity changes that portend seizure onset. In this initial study, we will analyze the data off-line to establish a reliable seizure prediction algorithm. Once this algorithm has been established, we will apply real time, prospective seizure perdiction methods to detect seizures in real time and we will combine this prediction tool with deep brain stimulation to prevent seizure development and progression. The long term goal of this research is to translate animal experimental results to the clinic to treat epilepsy patients with this novel approach. PUBLIC HEALTH RELEVANCE: Accurate seizure prediction can prevent bodily injury for epileptic patient and enable clinicians to treat epilepsy with timely drug administration or brain stimulation. This research project will develop a novel multiple channel, single unit recording method to predict spontaneous seizures in rat model of epilepsy. The success of this project will be instrumental in improving seizure prediction in the clinic and benefit millions of patients with epilepsy. PHS 398/2590 (Rev. 09/04) Page Continuation Format Page