A Comparison of Euclidean metrics in spike train space
Spike trains are observables when investigating neural activity - represent the response of a neuron to stimuli and are often modeled as realizations of stochastic point processes. The spike train space is non-euclidean, recently, however, two L 2
- like distances were introduced on that space:
the Elastic distance and Generalized Victor-Purpura (GVP) distance.
On this poster we briefly review these two distances and run several comparisons, including construction of the summary statistics, corresponding in ideas to mean and variance as well as classification capabilities. To allow comparisons between
metrics we propose an efficient algorithm for GVP summary statistics.