Read H L, Siegel R M
Center for Molecular and Behavioral Neuroscience, Rutgers University, Newark, NJ 07102, USA.
Neuroscience. 1996 Nov;75(1):301-14. doi: 10.1016/0306-4522(96)00227-8.
Aperiodic entrainment to rhythmic sensory input was obtained with either a single neuron or an excitatory network model, without addition of a stochastic or "noisy" element. The entrainment properties of primary sensory neurons were well captured by the dynamics of the Hodgkin-Huxley ordinary differential equations with a quiescent resting state or threshold for spike output. The frequency-amplitude parameter space was compressed and aperiodic regimes were small in comparison to those of periodically activated pacemaker-like neurons. Transitions between phase-locked and aperiodic entrainment patterns were predictable and determined by the equation dynamics, supporting the contention that some aperiodicities observed in situ arise from the inherent membrane properties of neurons. When the rhythmically activated neuron was embedded in an excitatory network of Hodgkin-Huxley neurons with heterogeneous synaptic delays, aperiodic entrainment patterns were more frequently encountered and these were associated with asynchronous output from the network. Embedding the rhythmically activated neuron in a network with synaptic delays greatly reduced the range of entrained spike frequencies and increased the variability in the neuronal firing. The temporal coding of sensory stimuli may be dependent on these findings. Sensory stimuli are signaled in the periphery by a mixture of periodic and irregular interspike intervals. Most models of such temporal codes assume intrinsic rhythmicity arising from the ionic currents, with variations attributed to membrane or synaptic noise. In contrast, we demonstrate irregular neural codes that arise completely in the absence of noise. In the proposed model, the sources of these irregular sensory patterns are the extensive cross-connections and resultant interactions between neurons. The balance between the regular and irregular entrainment of a neuron in situ could uniquely identify a stimulus. Other biological mechanisms of modifying the entrainment properties and promoting aperiodic entrainment are discussed.
在不添加随机或“噪声”元素的情况下,使用单个神经元或兴奋性网络模型实现了对节律性感觉输入的非周期性夹带。霍奇金 - 赫胥黎常微分方程的动力学很好地捕捉了初级感觉神经元的夹带特性,该方程具有静止的静息状态或动作电位输出阈值。与周期性激活的起搏器样神经元相比,频率 - 振幅参数空间被压缩,非周期性状态较小。锁相和非周期性夹带模式之间的转变是可预测的,并由方程动力学决定,这支持了原位观察到的一些非周期性源于神经元固有膜特性的观点。当节律性激活的神经元嵌入具有异质突触延迟的霍奇金 - 赫胥黎神经元兴奋性网络中时,更频繁地遇到非周期性夹带模式,并且这些模式与网络的异步输出相关。将节律性激活的神经元嵌入具有突触延迟的网络中大大降低了夹带动作电位频率的范围,并增加了神经元放电的变异性。感觉刺激的时间编码可能依赖于这些发现。感觉刺激在外周由周期性和不规则的峰间间隔混合信号表示。大多数此类时间编码模型假设离子电流产生内在节律性,变化归因于膜或突触噪声。相比之下,我们展示了完全在无噪声情况下出现的不规则神经编码。在所提出的模型中,这些不规则感觉模式的来源是广泛的交叉连接以及神经元之间由此产生的相互作用。神经元原位的规则和不规则夹带之间的平衡可以唯一地识别刺激。还讨论了改变夹带特性和促进非周期性夹带的其他生物学机制。