Discrimination of bursts and tonic activity in multifunctional sensorimotor neural network using the extended hill-valley method

Abstract

Individual neurons can exhibit a wide range of activity, including spontaneous spiking, tonic spiking, bursting, or spike-frequency adaptation, and can also transition between these activity types. Manual identification of these activity patterns can be subjective and inconsistent. The extended hill-valley (EHV) analysis discriminates tonic spiking and bursts in a spike train by detecting fluctuations in a local, history-dependent analysis signal derived from the spike train. Consequently, the EHV method is not susceptible to changes in baseline firing rate and can identify different types of activity patterns. In addition, output from the EHV method can be used to identify more complex activity patterns such as phasotonic bursting, in which a burst is immediately followed by a period of tonic spiking. NEW & NOTEWORTHY Neurons exhibit diverse spiking patterns, but automated activity classification has focused mainly on detecting bursts. The novel extended hill-valley algorithm uses a smoothed, history-dependent signal to discriminate different types of activity, such as bursts and tonic spiking.