Self-organized criticality in bat auditory cortex

Abstract

Presents evidence that biological cortical networks operate in a self-organizing critical state, and may dynamically modify their connection schemes in order to most efficiently solve a particular task. It has been theorized that the amount of interconnection that results in self-organizing criticality (SOC) permits an optimal quantity of information to propagate through a network. SOC arises in systems in which the degree of parallelism is such that the resultant networks are poised between ordered and chaotic behavior. The authors have found some unexpected correlates of SOC activity from analysis of noise recorded from the auditory cortex of the big brown bat (Eptesicus fuscus). In order to simulate echolocation, the authors presented stationary bats with recordings of vocalizations and echoes and recorded the synchronized responses of ensembles of neurons to these stimuli from a variety of auditory cortex locations. Two stimulus paradigms were employed: one consisting of a sequence in which a single echo followed a vocalization, and another consisting of a sequence in which 2 echoes followed a vocalization. Cortical potentials thus recorded were averaged and the resultant waveform is shown. Logarithmic power spectra were taken from intervals of cortical noise lasting 200 ms before the evoked response for each stimulus paradigm. A prototypical log power spectrum is shown. The log power spectrum is approximately linear; this implies a power spectrum that is proportional to (1/f)/sup n/ where n is some exponent close to 1, and f is frequency. Power spectra of this shape are considered hallmarks of SOC behavior.