Vida Imre, Bartos Marlene, Jonas Peter
Anatomisches Institut, Universität Freiburg, Albertstr. 17, D-79104 Freiburg, Germany.
Neuron. 2006 Jan 5;49(1):107-17. doi: 10.1016/j.neuron.2005.11.036.
Networks of GABAergic neurons are key elements in the generation of gamma oscillations in the brain. Computational studies suggested that the emergence of coherent oscillations requires hyperpolarizing inhibition. Here, we show that GABA(A) receptor-mediated inhibition in mature interneurons of the hippocampal dentate gyrus is shunting rather than hyperpolarizing. Unexpectedly, when shunting inhibition is incorporated into a structured interneuron network model with fast and strong synapses, coherent oscillations emerge. In comparison to hyperpolarizing inhibition, networks with shunting inhibition show several advantages. First, oscillations are generated with smaller tonic excitatory drive. Second, network frequencies are tuned to the gamma band. Finally, robustness against heterogeneity in the excitatory drive is markedly improved. In single interneurons, shunting inhibition shortens the interspike interval for low levels of drive but prolongs it for high levels, leading to homogenization of neuronal firing rates. Thus, shunting inhibition may confer increased robustness to gamma oscillations in the brain.
γ-氨基丁酸能神经元网络是大脑中γ振荡产生的关键要素。计算研究表明,相干振荡的出现需要超极化抑制。在此,我们表明,海马齿状回成熟中间神经元中γ-氨基丁酸A(GABA(A))受体介导的抑制是分流性的,而非超极化性的。出乎意料的是,当将分流性抑制纳入具有快速且强突触的结构化中间神经元网络模型时,相干振荡会出现。与超极化抑制相比,具有分流性抑制的网络具有几个优点。首先,振荡在较小的紧张性兴奋性驱动下产生。其次,网络频率被调谐到γ频段。最后,对兴奋性驱动异质性的鲁棒性显著提高。在单个中间神经元中,分流性抑制在低水平驱动时缩短峰间间隔,但在高水平驱动时延长峰间间隔,导致神经元放电率均匀化。因此,分流性抑制可能会增强大脑中γ振荡的鲁棒性。