Everything Is a Relationship

Abstract

This chapter discusses the hypothesis that the strongly skewed nature of our perceptions and memory result from log-normal distributions of anatomical connectivity at both micro- and mesoscales, synaptic weight distributions, firing rates, and neuronal population activity. Nearly all anatomical and physiological features of the brain are part of a continuous but wide distribution, typically obeying a log-normal form. This organization implies that the interactions that give rise to this distribution involve multiplication or division of random factors, resulting in values that can span several orders of magnitude. Neuronal networks with such broad distributions are needed to maintain stability against competing needs, including wide dynamic range, redundancy, resilience, homeostasis, and plasticity. These features of the brain may explain the Weber-Fechner law: for any sensory modality, perceptual intensity is a logarithmic function of physical intensity. Neuronal systems organized according to log rules form brain networks that can produce good-enough and fast decisions in most situations using only a subset of the brain’s resources.