Decoding network activity from LFPS: A computational approach

Abstract

Cognition is one of the main capabilities of mammal brain and understanding it thoroughly requires decoding brain's information processing pathways which are composed of networks formed by complex connectivity between neurons. Mostly, scientists rely on local field potentials (LFPs) averaged over a number of trials to study the effect of stimuli on brain regions under investigation. However, this may not be the right approach when trying to understand the exact neuronal network underlying the neuronal signals. As the LFPs are lumped activity of populations of neurons, their shapes provide fingerprints of the underlying networks. The method presented in this paper extracts shape information of the LFPs, calculate the corresponding current source density (CSD) from the LFPs and decode the underlying network activity. Through simulated LFPs it has been found that differences in LFP shapes lead to different network activity.