Interaction between spike waveform classification and temporal sequence detection

Abstract

In vivo extracellular recordings have allowed researchers to study the response properties of neurons to behaviorally relevant stimuli. In this paper we use multiple tetrode recordings from the hippocampus of the freely behaving rat to show that the action potential amplitude of a given cell can vary in a systematic and activity dependent manner over behaviorally relevant time scales. Since the discrimination algorithms used by experimenters to isolate cells from extracellular recordings are based on differences in waveforms, we show how these systematic changes in waveform shape can lead to non-random errors in single cell isolation. We further demonstrate that these non-random errors can lead to apparent temporal ordering effects between neurons in the absence of any specific temporal relationship. A firm understanding of these naturally occurring physiological changes is therefore critical for the evaluation of higher order phenomena such as the temporally correlated firing of ensembles of neurons.