A MATLAB toolbox for Granger causal connectivity analysis.

By Anil K. Seth. Journal of Neuroscience Methods, 2010

770 citations recorded by the source corpus.

Assessing directed functional connectivity from time series data is a key challenge in neuroscience. One approach to this problem leverages a combination of Granger causality analysis and network theory. This article describes a freely available MATLAB toolbox--'Granger causal connectivity analysis' (GCCA)--which provides a core set of methods for performing this analysis on a variety of neuroscience data types including neuroelectric, neuromagnetic, functional MRI, and other neural signals. The toolbox includes core functions for Granger causality analysis of multivariate steady-state and event-related data, functions to preprocess data, assess statistical significance and validate results, and to compute and display network-level indices of causal connectivity including 'causal density' and 'causal flow'. The toolbox is deliberately small, enabling its easy assimilation into the repertoire of researchers. It is however readily extensible given proficiency with the MATLAB language.

Subfields: Cognitive Neuroscience, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/7e5aa8ade633932fbd00aad59d0998fb7bc6186d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Analysing connectivity with Granger causality and dynamic causal modelling (2013, methodology): “The most common implementation of GC is via linear vector autoregressive (VAR) modelling of timeseries data, enabling both statistical significance testing and estimation of GC magnitudes [6 ,15 ,16].”
Wiener-Granger Causality: A well established methodology (2011, neutral): “For example, bandpass filtering may cause severe confounds in WGC analysis by introducing temporal correlations into MEG or EEG time series (Seth, 2010; Florin et al., 2009).”
Application of Graph Theory for Identifying Connectivity Patterns in Human Brain Networks: A Systematic Review (2019, methodology): “To tackle this issue, GC analysis is typically performed by fitting a linear vector autoregressive (VAR) to the time series (Seth, 2010; Friston et al., 2013; Seth et al., 2015).”
A critical assessment of connectivity measures for EEG data: A simulation study (2013, methodology): “If antisymmetrization is to be avoided, an analytic procedure using F-tests can be used for Granger causality to assess whether the full MVAR model adds significant explanatory power compared to the reduced model (e.g. Seth, 2010).”
Network modelling methods for FMRI (2011, neutral): “For the first, we used the “Causal Connectivity Analysis” toolbox (Seth, 2010) (www.”
HERMES: Towards an Integrated Toolbox to Characterize Functional and Effective Brain Connectivity (2013, methodology): “In contrast to existing toolboxes (Delorme and Makeig 2004; Zhou et al. 2009; Seth 2010; Tadel et al. 2011), HERMES encompasses both linear and nonlinear indexes of functional and effective connectivity, visualization tools and two different types of statistical tests to address the multiple…”
The MVGC multivariate Granger causality toolbox: a new approach to Granger-causal inference. (2014, methodology): “In these cases ﬁltering may be a valid and useful tool to improve stationarity; e.g. notch-ﬁltering of line-noise in electronically recorded time series data (Barnett and Seth, 2011), or ﬁnite di ﬀ erencing to eliminate drift (Seth, 2010).”
Constraints extraction from asymmetrical bimanual tasks and their use in coordinated behavior (2018, methodology): “1) Motion–motion coupling: For actions in which both arms employed position control we determined the coupled dimensions by analyzing the pair–wise interactions between the components of ν , ν using Granger Causality [43].”
Neural mechanisms of time-forward predictions for naturalistic auditory tone sequences (2026, unknown): “For each electrode pair, we computed pairwise GC using neural activity in the broadband (0.5 – 150 Hz) range, following the well-established time-domain GC analysis approach 29,30 .”
Directional prefrontal-thalamic information flow is selectively required during spatial working memory retrieval (2022, methodology): “…function (DTF) method, a multi-channel parametric method of analysis based on an autoregressive model, has been applied to a number of neurobiological systems for analysis of causal connectivity and flow pattern of activity in the frequency domain (Kaminski and Blinowska, 1991; Seth, 2010).”