Lempel-ziv Complexity for Analysis of Neural Spikes

Analysis of neural spikes is well established [4]. Many methods can be used to uncover distinct patterns of spikes. Consider two spike trains: 10101010 and 11110000, recorded in 1 s (where 1 represents the occurrence of a spike in a bin of length 1000/8 ms). The firing rates of both spike trains are identical (4 Hz), however the patterns are different; that is, the first train shows an oscillatory behavior while the second train resembles a step function. If two separate stimuli elicited these patterns, then the firing rate in this example is inadequate for characterizing the stimulus and does not provide any information about the distinct patterns that might be observed in a spike train. Analyses of distinct spike patterns have revealed underlying nonlinear and chaotic dynamics of firing, and would benefit from complementary metrics for complexity analysis of neural spikes.