MEG Reading Group

An information theoretic framework for neuroimaging data analysis: stimulus modulations, representational interactions and causal communication of specific information content

Information theory provides a principled and unified statistical framework for neuroimaging data analysis. A major factor hindering wider adoption of this framework is the difficulty of estimating information theoretic quantities in practice. We present a novel estimation technique that combines the statistical theory of copulas with the closed form solution for the entropy of Gaussian variables (Ince et al., 2016a). This results in a general, computationally efficient, flexible, and robust multivariate statistical framework that provides effect sizes on a common meaningful scale, allows for unified treatment of discrete, continuous, uni- and multi-dimensional variables, and enables direct comparisons of representations from behavioral and brain responses across any recording modality. Open-source Matlab and Python code implementing the new methods is available at: https://github.com/robince/gcmi