IRIBHM, Université Libre de Bruxelles, Brussels, Belgium.
Theoretical Neurobiology & Neuroengineering, University of Antwerp, Campus CDE, Universiteitsplein 1, 2610, Wilrijk, Antwerp, Belgium.
Eur J Neurosci. 2018 Jan;47(1):17-32. doi: 10.1111/ejn.13761. Epub 2017 Dec 2.
Ensembles of cortical neurons can track fast-varying inputs and relay them in their spike trains, far beyond the cut-off imposed by membrane passive electrical properties and mean firing rates. Initially explored in silico and later demonstrated experimentally, investigating how neurons respond to sinusoidally modulated stimuli provides a deeper insight into spike initiation mechanisms and information processing than conventional F-I curve methodologies. Besides net membrane currents, physiological synaptic inputs can also induce a stimulus-dependent modulation of the total membrane conductance, which is not reproduced by standard current-clamp protocols. Here, we investigated whether rat cortical neurons can track fast temporal modulations over a noisy conductance background. We also determined input-output transfer properties over a range of conditions, including: distinct presynaptic activation rates, postsynaptic firing rates and variability and type of temporal modulations. We found a very broad signal transfer bandwidth across all conditions, similar large cut-off frequencies and power-law attenuations of fast-varying inputs. At slow and intermediate input modulations, the response gain decreased for increasing output mean firing rates. The gain also decreased significantly for increasing intensities of background synaptic activity, thus generalising earlier studies on F-I curves. We also found a direct correlation between the action potentials' onset rapidness and the neuronal bandwidth. Our novel results extend previous investigations of dynamical response properties to non-stationary and conductance-driven conditions, and provide computational neuroscientists with a novel set of observations that models must capture when aiming to replicate cortical cellular excitability.
皮质神经元集合可以跟踪快速变化的输入，并在其尖峰序列中传递这些输入，这远远超出了膜被动电学特性和平均发放率所施加的限制。最初在计算机中进行探索，后来在实验中得到证实，研究神经元如何对正弦调制刺激做出反应，比传统的 F-I 曲线方法提供了对尖峰起始机制和信息处理的更深入了解。除了净膜电流外，生理突触输入还可以诱导总膜电导的刺激依赖性调制，而标准电流钳协议无法再现这种调制。在这里，我们研究了大鼠皮质神经元是否可以在噪声电导背景下跟踪快速的时间调制。我们还在一系列条件下确定了输入-输出传递特性，包括：不同的突触前激活率、突触后发放率和变异性以及时间调制的类型。我们发现所有条件下的信号传递带宽都非常宽，类似的大截止频率和快速变化输入的幂律衰减。在缓慢和中等输入调制下，随着输出平均发放率的增加，响应增益降低。随着背景突触活动强度的增加，增益也显著降低，从而推广了早期关于 F-I 曲线的研究。我们还发现动作电位起始速度与神经元带宽之间存在直接相关性。我们的新结果将动态响应特性的先前研究扩展到非平稳和电导驱动的条件，并为计算神经科学家提供了一组新的观察结果，当模型旨在复制皮质细胞兴奋性时，模型必须捕捉这些观察结果。
