Transfer entropy—a model-free measure of effective connectivity for the neurosciences

By Raul Vicente, Michael Wibral, Michael Lindner, Gordon Pipa. Journal of Computational Neuroscience, 2010

1,021 citations recorded by the source corpus.

Understanding causal relationships, or effective connectivity, between parts of the brain is of utmost importance because a large part of the brain’s activity is thought to be internally generated and, hence, quantifying stimulus response relationships alone does not fully describe brain dynamics. Past efforts to determine effective connectivity mostly relied on model based approaches such as Granger causality or dynamic causal modeling. Transfer entropy (TE) is an alternative measure of effective connectivity based on information theory. TE does not require a model of the interaction and is inherently non-linear. We investigated the applicability of TE as a metric in a test for effective connectivity to electrophysiological data based on simulations and magnetoencephalography (MEG) recordings in a simple motor task. In particular, we demonstrate that TE improved the detectability of effective connectivity for non-linear interactions, and for sensor level MEG signals where linear methods are hampered by signal-cross-talk due to volume conduction.

Subfields: Artificial Intelligence, Cognitive Neuroscience

View Paper: https://www.semanticscholar.org/paper/c85fa1d222a5847a97e385046497d5e2a6d98dba

How researchers see this paper

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

In Artificial Intelligence

In Cognitive Neuroscience

Citations in the corpus

JIDT: An Information-Theoretic Toolkit for Studying the Dynamics of Complex Systems (2014, methodology): “This is done (following (Chávez et al., 2003; Verdes, 2005; Lindner et al., 2011; Lizier et al., 2011a; Vicente et al., 2011; Barnett and Bossomaier, 2012; Wibral et al., 2014b)) via permutation testing to construct appropriate surrogate populations of timeseries and measurements under the null hypo”
Causation entropy identifies indirect influences, dominance of neighbors and anticipatory couplings (2014, background): “Recently, it becomes increasingly popular to use transfer entropy for causality inference in networks of neurons [10, 11] and in coupled dynamical systems with parameter mismatches [12], anticipatory couplings [13], and time delays [14].”
Application of Graph Theory for Identifying Connectivity Patterns in Human Brain Networks: A Systematic Review (2019, methodology): “In addition, TE does not assume any particular model as underlying the interactions, therefore, its sensitivity to all order correlations becomes a privilege for exploratory analyzes over GC or other model-based methods (Vicente et al., 2011; Montalto et al., 2014).”
HERMES: Towards an Integrated Toolbox to Characterize Functional and Effective Brain Connectivity (2013, background): “This is particularly relevant when the detection of some unknown non-linear interaction is required (Vicente et al. 2011).”
Phase transfer entropy: A novel phase-based measure for directed connectivity in networks coupled by oscillatory interactions (2014, methodology): “TE may be a good candidate metric for phase-based connectivity analyses and has been used to detect directed interactions in both sensor- and source-space broadband MEG/EEG signals (Wibral et al., 2011; AC C EP TE D M AN U SC R IP T Vicente et al., 2011).”
Communication dynamics in complex brain networks (2017, methodology): “Alternative statistical methods used to infer the directionality of information flow, such as Granger causality 71 , transfer entropy 72,73 and others 74 , show that dominant patterns of information flow are associated with specific frequencies in neuro physiological recordings 75,76 .”
A critical assessment of connectivity measures for EEG data: A simulation study (2013, background): “There exists also nonlinear variants of Granger causality (Marinazzo et al., 2008; Vicente et al., 2011), which are not considered here.”
Granger Causality: A Review and Recent Advances (2021, methodology): “Model-free methods, like transfer entropy (Vicente et al., 2011) or directed information (Amblard and Michel, 2011), can detect nonlinear dependencies between past and future with minimal assumptions on the predictive relationships.”
Neural Granger Causality (2021, methodology): “For example, in neuroscience it is important to determine how brain activation spreads through brain regions [1], [2], [3], [4]; in ﬁnance it is important to determine groups of stocks with low covariance to design low risk portfolios [5]; and, in biology, it is of great interest to infer gene…”
Exploring Causal Relationships in Visual Object Tracking (2015, background): “TE has since found its place in many areas, including again neurosciences [36], chemistry [3] and others [23].”
Correlations of neural predictability and information transfer in cortex and their relation to predictive coding (2025, unknown): “We used a multivariate measure of transfer entropy to infer brain connectivity (Fig 2A) from intracranial EEG recordings of human epileptic patients (Fig 1A) [29].”
Intracranial EEG signals disentangle multi-areal neural dynamics of vicarious pain perception (2024, methodology): “The choice of using TE to index effective connectivity was based on previous work 94 , which showed that TE, as compared to Granger Causality and other model-based approaches, is a model-free measure of effective connectivity based on information theory and better at detecting effective…”
Neural interactions in the human frontal cortex dissociate reward and punishment learning (2024, methodology): “To generate the distributions of II and TE reachable by chance, we respectively shuffled the PE variable across trials for the II and the gamma power across trials of the source for the TE (Vicente et al., 2011).”