Cortical Specializations Underlying Fast Computations.

Cortical Specializations Underlying Fast Computations.
复制标题

DOI:
10.1177/1073858415571539
复制
发表时间:
2016-04
期刊:
The Neuroscientist : a review journal bringing neurobiology, neurology and psychiatry
影响因子:
--
通讯作者:
Volgushev M
Volgushev M
中科院分区:
其他
文献类型:
--
作者:
Volgushev M

文献摘要

相似文献

行为相关的环境事件的时间进程设置神经元处理的时间限制。哺乳动物的大脑如何利用新皮层日益复杂的网络,同时在合理的时间内做出决定和执行行为反应?决定神经元网络计算速度的关键参数是神经元集成处理其输入变化并将处理结果传达给下游神经元所需的时间间隔。理论分析确定了快速处理的基本要求:使用神经元群体进行编码,背景活动和神经元中动作电位的快速启动动力学。实验证据表明,新皮层神经元群体满足这些要求。事实上,它们可以在1到3 ms内非常快速地响应输入扰动来改变发射率,并通过将其尖峰锁定到高达300到1000 Hz的频率来编码输入的高频分量。这意味着皮质系综计算的时间单位只有几个,1至3毫秒,这是相当快的单个神经元的膜时间常数。皮层神经元集合在毫秒时间尺度上进行通信的能力允许在并行处理流中进行复杂的多步骤处理和神经元活动的精确协调,同时将行为反应的速度保持在环境设定的时间限制内。
The time course of behaviorally relevant environmental events sets temporal constraints on neuronal processing. How does the mammalian brain make use of the increasingly complex networks of the neocortex, while making decisions and executing behavioral reactions within a reasonable time? The key parameter determining the speed of computations in neuronal networks is a time interval that neuronal ensembles need to process changes at their input and communicate results of this processing to downstream neurons. Theoretical analysis identified basic requirements for fast processing: use of neuronal populations for encoding, background activity, and fast onset dynamics of action potentials in neurons. Experimental evidence shows that populations of neocortical neurons fulfil these requirements. Indeed, they can change firing rate in response to input perturbations very quickly, within 1 to 3 ms, and encode high-frequency components of the input by phase-locking their spiking to frequencies up to 300 to 1000 Hz. This implies that time unit of computations by cortical ensembles is only few, 1 to 3 ms, which is considerably faster than the membrane time constant of individual neurons. The ability of cortical neuronal ensembles to communicate on a millisecond time scale allows for complex, multiple-step processing and precise coordination of neuronal activity in parallel processing streams, while keeping the speed of behavioral reactions within environmentally set temporal constraints.