2017/08/23 by Eero Satuvuori, Satuvuori, Eero, Thomas Kreuz +1 · 2 citations
Engineering · Neuroscience · #Advanced Memory and Neural Computing #Data Analysis #FOS: Biological sciences #FOS: Physical sciences #Neural dynamics and brain function #Neurons and Cognition (q-bio.NC) #Neuroscience and Neural Engineering #Statistics and Probability (physics.data-an)
paper · pdf · doi:10.48550/arxiv.1708.07508
openalex publication_date 2017/08/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Background: It is commonly assumed in neuronal coding that repeated\npresentations of a stimulus to a coding neuron elicit similar responses. One\ncommon way to assess similarity are spike train distances. These can be divided\ninto spike-resolved, such as the Victor-Purpura and the van Rossum distance,\nand time-resolved, e.g. the ISI-, the SPIKE- and the RI-SPIKE-distance.\n New Method: We use independent steady-rate Poisson processes as surrogates\nfor spike trains with fixed rate and no timing information to address two basic\nquestions: How does the sensitivity of the different spike train distances to\ntemporal coding depend on the rates of the two processes and how do the\ndistances deal with very low rates?\n Results: Spike-resolved distances always contain rate information even for\nparameters indicating time coding. This is an issue for reasonably high rates\nbut beneficial for very low rates. In contrast, the operational range for\ndetecting time coding of time-resolved distances is superior at normal rates,\nbut these measures produce artefacts at very low rates. The RI-SPIKE-distance\nis the only measure that is sensitive to timing information only.\n Comparison with Existing Methods: While our results on rate-dependent\nexpectation values for the spike-resolved distances agree with\n citetChicharro11, we here go one step further and specifically investigate\napplicability for very low rates.\n Conclusions: The most appropriate measure depends on the rates of the data\nbeing analysed. Accordingly, we summarize our results in one table that allows\nan easy selection of the preferred measure for any kind of data.\n