ABSTRACT
We introduce a time-invariant neural spike train analysis method to detect oscillations in neurons, and a cross-interval-interspike-interval measure for detecting the synchronous spike firings between neurons. A renormalized measure of the ratio of interspike intervals, called firing trend index, is used to detect relative changes in firing rate so that true oscillations can be distinguished from other statistical local random fluctuations, while the cross-interval-interspike-interval analysis is used to detect the specific nature of the time-locked firing between neurons so that correlated spike firings can be distinguished from uncorrelated firings. Such a distinction of correlated firing is important to reveal whether the near-synchronous firings are tightly timelocked or uncorrelated due to chance coincidence. Simulation results showed that these analyses can uncover the spike generation process contributing to the phenomenon of oscillations and correlated synchronous firings in neurons.
