An improved graphical method for pattern recognition from spike trains of spontaneously active neurons

Spontaneous activity and rhythmical oscillations are common features of large neuronal networks in mammals. Detection of repetitive spike patterns or pacemaker activity during electrophysiological recording of spontaneous action potentials from single neurons can be difficult if a "noisy" background is present. This paper describes an improved method for an online spike train analysis based on joint interval histograms (JIH, Rodiek et al. 1962). By means of higher ordered JIH the discrimination of spike patterns with repetitive bursting activity or oscillations is possible even when randomly distributed action potentials appear. Examples of simulated spike trains and those recorded from cultured hippocampal neurons are presented.