AbstractFor certain brain functions, the theoretical networks presented here almost certainly show how neurons are actually connected. Stripped of details such as redundancies and other error-correcting mechanisms, the basic organization of synaptic connections within some of the brain’s building blocks is likely to be less complex than it appears. For some brain functions, the network architectures can even be quite simple.Flip-flops are the basic building blocks of sequential logic systems. Certain flip-flops can be configured to function as oscillators. The flip-flops and oscillators proposed here are composed of two to six neurons, and their operation depends only on minimal neuron capabilities of excitation and inhibition. These networks suggest a resolution to the longstanding controversy of whether short-term memory depends on neurons firing persistently or in brief, coordinated bursts. Oscillators can also generate major phenomena of electroencephalography.For example, cascaded oscillators can produce the periodic activity commonly known as brainwaves by enabling the state changes of many neural structures simultaneously. (The function of such oscillator-induced synchronization in information processing systems is timing error avoidance.) Then the boundary separating the alpha and beta frequency bands is
where μd and σd are the mean and standard deviation (in milliseconds) of delay times of neurons that make up the initial oscillators in the cascades. With 4 and 1.5 ms being the best estimates for μd and σd, respectively, this predicted boundary value is 14.9 Hz, which is within the range of commonly cited estimates obtained empirically from electroencephalograms (EEGs). The delay parameters μd = 4 and σd = 1.5 also make predictions of the peaks and other boundaries of the five major EEG frequency bands that agree well with empirically estimated values.The hypothesis that cascaded oscillators produce EEG frequencies implies two EEG characteristics with no apparent function: The EEG gamma band has the same distribution of frequencies as three-neuron ring oscillators, and the ratios of peaks and boundaries of the major EEG bands are powers of two. These anomalous properties make it implausible that EEG phenomena are produced by a mechanism that is fundamentally different from cascaded oscillators.The cascaded oscillators hypothesis is supported by the available data for neuron delay times and EEG frequencies; the micro-level explanations of macro-level phenomena; the number, diversity, and precision of predictions of EEG phenomena; the simplicity of the oscillators and minimal required neuron capabilities; the selective advantage of timing error avoidance that cascaded oscillators can provide; and the implausibility of a fundamentally different mechanism producing the phenomena.The available data are too imprecise for a rigorous statistical test of the cascaded oscillators hypothesis. A simple, rigorous test of the hypothesis is suggested. The neuron delay parameters μd and σd, as well as the mean and variance of the periods of one or more EEG bands, can be estimated from random samples. With standard tests for equal means and variances, the EEG sample statistics can be compared to the EEG parameters predicted by the delay time statistics.