Nikolić Danko
Max-Planck-Institute for Brain Research, Deutschordenstr. 46, D-60528, Frankfurt am Main, Germany.
J Comput Neurosci. 2007 Feb;22(1):5-19. doi: 10.1007/s10827-006-9441-7. Epub 2006 Sep 19.
Neuronal synchronization is often associated with small time delays, and these delays can change as a function of stimulus properties. Investigation of time delays can be cumbersome if the activity of a large number of neurons is recorded simultaneously and neuronal synchronization is measured in a pairwise manner (such as the cross-correlation histograms) because the number of pairwise measurements increases quadratically. Here, a non-parametric statistical test is proposed with which one can investigate (i) the consistency of the delays across a large number of pairwise measurements and (ii) the consistency of the changes in the time delays as a function of experimental conditions. The test can be classified as non-parametric because it takes into account only the directions of the delays and thus, does not make assumptions about the distributions and the variances of the measurement errors.
神经元同步通常与小时间延迟相关，并且这些延迟会随刺激特性而变化。如果同时记录大量神经元的活动并以成对方式（如互相关直方图）测量神经元同步，那么对时间延迟的研究可能会很繁琐，因为成对测量的数量会呈二次方增加。在此，提出了一种非参数统计检验方法，利用该方法可以研究：（i）大量成对测量中延迟的一致性；（ii）作为实验条件函数的时间延迟变化的一致性。该检验可归类为非参数检验，因为它仅考虑延迟的方向，因此不对测量误差的分布和方差做假设。
