Marczynski T J, Burns L L, Marczynski G T
Brain Res. 1980 Mar 3;185(1):139-60. doi: 10.1016/0006-8993(80)90678-2.
The study addressed the problem of information transmission in mammalian brain as reflected in the emergence or disappearance of temporal patterns in extracellularly monitored single action potentials from the dorsal hippocampus of unrestrained cats during slow wave sleep (SWS), rapid eye movement sleep (REM), and motionless quiet wakefulness (QW). The spike trains were analyzed with a nonparametric technique. Chi-square statistics were used to measure deviation of firing patterns from the theoretical model which is based on the assumption that the intervals are random and/or independent from each other. The plots of the chi-square values for a given set of patterns represented the neuronal 'signatures' characteristic of a behavioral state. During SWS most neurons followed the theoretical model, i.e. their 'signatures' were flat and statistically non-significant. However, during REM sleep and QW their firing modes showed specific deviations from the theoretical model: some patterns occurred more often while others less often than expected, thus generating large and statistically significant 'signatures'. During REM sleep some neurons shared similar tendencies in their departures from the theoretical model. However, during QW the same neurons developed their individual 'signatures' which were significantly different from each other. Hence, the QW episodes were characterized by a greater differentiation of neuronal firing patterns. The mean firing rate and the shape of the time interval histogram were not necessarily correlated with the emergence of specific temporal patterns in spike trains. The results suggest that information transmission from one neuron to another depends on the emergence of repetitive and specific temporal patterns. The strong tendency of most neurons to lapse during SWS into a firing mode that closely follows the theoretical model constitutes the basis for a working hypothesis which states that the essence of SWS recovery in cognitive systems is the disappearance of temporal patterns, and that the 'noisy' interactions between neurons plays an important role in the recuperative processes.
该研究探讨了哺乳动物大脑中的信息传递问题，这一问题通过在慢波睡眠（SWS）、快速眼动睡眠（REM）和静止安静觉醒（QW）期间，对未受约束的猫的背侧海马体进行细胞外监测的单个动作电位的时间模式的出现或消失来反映。使用非参数技术对尖峰序列进行分析。卡方统计用于测量放电模式与理论模型的偏差，该理论模型基于间隔是随机的和/或相互独立的假设。给定一组模式的卡方值图代表了行为状态特有的神经元“特征”。在慢波睡眠期间，大多数神经元遵循理论模型，即它们的“特征”是平坦的且在统计学上不显著。然而，在快速眼动睡眠和静止安静觉醒期间，它们的放电模式显示出与理论模型的特定偏差：一些模式出现的频率比预期更高，而另一些模式出现的频率比预期更低，从而产生了大的且在统计学上显著的“特征”。在快速眼动睡眠期间，一些神经元在偏离理论模型方面具有相似的趋势。然而，在静止安静觉醒期间，相同的神经元形成了彼此显著不同的个体“特征”。因此，静止安静觉醒阶段的特征是神经元放电模式有更大的差异。平均放电率和时间间隔直方图的形状不一定与尖峰序列中特定时间模式的出现相关。结果表明，从一个神经元到另一个神经元的信息传递取决于重复和特定时间模式的出现。大多数神经元在慢波睡眠期间强烈倾向于进入一种紧密遵循理论模型的放电模式，这构成了一个工作假设的基础，该假设指出认知系统中慢波睡眠恢复的本质是时间模式的消失，并且神经元之间的“嘈杂”相互作用在恢复过程中起重要作用。
