Dynamic control of modeled tonic-clonic seizure states with closed-loop stimulation
Dynamic control of modeled tonic-clonic seizure states with closed-loop stimulation
复制标题
DOI：
10.3389/fncir.2012.00126
复制
发表时间：
2013-02-06
影响因子：
3.5
通讯作者：
Netoff, Theoden I.
中科院分区：
文献类型：
--
作者：
Beverlin, Bryce, II;Netoff, Theoden I.
使用深部脑刺激 (DBS) 控制癫痫发作为难治性和耐药性癫痫患者提供了一种替代疗法。本文提出了新颖的 DBS 刺激方案来扰乱癫痫发作。提出了两种协议：开环刺激和闭环反馈系统，利用测量的放电率来调整刺激频率。使用 3000 个与抑制突触相连的兴奋性 Morris-Lecar (M-L) 模型神经元在计算模型中演示了刺激抑制。细胞使用二阶网络拓扑 (SONET) 连接，以模拟皮层网络中测量的网络拓扑。随着突触强度和神经元强直输入的减少，网络自发地从强直切换为阵挛。我们向该模型添加周期性刺激脉冲来模拟 DBS。周期性强迫可以同步或去同步振荡的神经元群体，具体取决于刺激频率和幅度。因此，可以延长或缩短癫痫发作的强直期或阵挛期。以神经元的发射速率施加的刺激通常使群体同步，而比发射速率稍慢的刺激则阻止同步。我们提出了一种自适应刺激算法，可测量神经元的放电率并调整刺激以维持相对刺激频率与放电频率，并在强直阵挛发作的计算模型中进行演示。这种自适应算法可以使用比开环控制小得多的刺激幅度来影响强直阶段的持续时间。
Seizure control using deep brain stimulation (DBS) provides an alternative therapy to patients with intractable and drug resistant epilepsy. This paper presents novel DBS stimulus protocols to disrupt seizures. Two protocols are presented: open-loop stimulation and a closed-loop feedback system utilizing measured firing rates to adjust stimulus frequency. Stimulation suppression is demonstrated in a computational model using 3000 excitatory Morris-Lecar (M-L) model neurons connected with depressing synapses. Cells are connected using second order network topology (SONET) to simulate network topologies measured in cortical networks. The network spontaneously switches from tonic to clonic as synaptic strengths and tonic input to the neurons decreases. To this model we add periodic stimulation pulses to simulate DBS. Periodic forcing can synchronize or desynchronize an oscillating population of neurons, depending on the stimulus frequency and amplitude. Therefore, it is possible to either extend or truncate the tonic or clonic phases of the seizure. Stimuli applied at the firing rate of the neuron generally synchronize the population while stimuli slightly slower than the firing rate prevent synchronization. We present an adaptive stimulation algorithm that measures the firing rate of a neuron and adjusts the stimulus to maintain a relative stimulus frequency to firing frequency and demonstrate it in a computational model of a tonic-clonic seizure. This adaptive algorithm can affect the duration of the tonic phase using much smaller stimulus amplitudes than the open-loop control.