Aplicação de análise multivariada em detecção de disparos de neurônios

Abstract

Spike sorting is a problem in contemporary neuroscience and consists of detecting and grouping neural activities, also called peaks. The problem exists because there is a difficulty when it comes to differentiating these activities, when several neurons are recorded simultaneously, causing different neural activities to be detected. This work aims to present a process for solving the spike sorting problem. For this we will, using software Rstudio, apply some statistical techniques on a real dataset about neural activities of a species of grasshopper Schistocerca americana, with the absolute median deviation and multivariate analysis techniques as analysis of main components and K-means.