Building spike representation in tetrodes

Abstract

This paper presents a new technique for analyzing the recorded information from tetrodes in freely behaving rats, based on independent component analysis (ICA). The ion-specific pumps and channels allow fast transfer of charges such as Na +, K +, Cl − and eventually Ca 2+ during each action potential (AP). These groups of charges under an electrical field have distinct spatial trajectories. Therefore, the generated signals within a tetrode are considered to be composed mainly by statistically independent signal sources that can be obtained by performing ICA. In order to compute the position of independent sources during AP generation, the triangulation method uses an iterative Newton-Raphson algorithm. The representation of the independent signal sources in three-dimensional tetrode space is then obtained. Since the charge movements are extensively spread on the neuron's surface, the representation in tetrode space reveals electrical spatial patterns of activation during each AP. The analysis of several spikes coming from the same neuron reveals small changes from spike to spike in the 3D shape. Since information within spikes is highly transferred by ionic fluxes these electrical patterns of activation reflect neuronal computation occurring during each AP.