Chapter 6 - Oscillating Neural Synchrony

Abstract

This chapter applied stable information theory and Reduced Error Logistic Regression (RELR) to oscillating neural synchrony phenomena in neural computation. The concept of oscillating neural synchrony was introduced as measured by the electroencephalography (EEG). The relationship between the EEG and local potential fields to the probability of binary spiking in neural populations was then introduced, and the suggestion was that local field potentials are good proxies for the probability of binary spiking in an underlying neural population. The relationship between EEG frequency and rhythmic structure and cognitive effects related to consciousness was also reviewed. The Hubel and Wiesel hierarchical integration model of visual perception was contrasted with binding by synchrony proposal of Singer. The Gestalt psychology roots of Singer's proposal were summarized. An RELR model of square perception was then shown to be consistent with binding by synchrony in the case of illusionary Kanizsa square perception. The concept of spike-timing plasticity was also connected to oscillating neural synchrony models of brain learning. This was also suggested to be consistent with RELR. Finally, evidence on the timing of cognitive operations as reflected in rhythmic oscillating neural synchrony patterns in the EEG was reviewed. Similarities to the rhythmic structure of music were suggested.