| Autor: |
Hidetoshi Konno, Yoshiyasu Tamura |
| Jazyk: |
angličtina |
| Rok vydání: |
2018 |
| Předmět: |
|
| Zdroj: |
AIP Advances, Vol 8, Iss 1, Pp 015118-015118-16 (2018) |
| Druh dokumentu: |
article |
| ISSN: |
2158-3226 |
| DOI: |
10.1063/1.5012547 |
| Popis: |
In neural spike counting experiments, it is known that there are two main features: (i) the counting number has a fractional power-law growth with time and (ii) the waiting time (i.e., the inter-spike-interval) distribution has a heavy tail. The method of superstatistical Poisson processes (SSPPs) is examined whether these main features are properly modeled. Although various mixed/compound Poisson processes are generated with selecting a suitable distribution of the birth-rate of spiking neurons, only the second feature (ii) can be modeled by the method of SSPPs. Namely, the first one (i) associated with the effect of long-memory cannot be modeled properly. Then, it is shown that the two main features can be modeled successfully by a class of fractional SSPP (FSSPP). |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
