Causality Analysis of Multivariate Neural Data
DOI link for Causality Analysis of Multivariate Neural Data
Causality Analysis of Multivariate Neural Data book
This chapter reviews advances of causal influence measures for analyzing multivariate neural data under the framework of multivariate autoregressive (MAR) modeling. It discusses practical issues concerning how to estimate such measures from time series data with special emphasis on MAR model, from which the Granger causality and its partial measures can be derived. The chapter offers an overview of specific issues concerning multivariate data analysis, present available tools and give a few examples of their applications. It addresses the key technical issues of the causal measures concerning stationarity/nonstationarity, bivariate/multivariate, linearity/nonlinearity, and statistical significance. The chapter presents several selected applications in the analysis of spike trains, electroencephalogram, and Functional magnetic resonance imaging BOLD data, each representing a major recording modality used in neuroscience research, followed by a demonstration of clinical application in seizure localization in epilepsy patients. It also presents a variety of applications ranging from basic neuroscience research to clinical relevance.