Algorithmic procedures for Bayesian MEG/EEG source reconstruction in SPM

• Published in: NeuroImage (Orlando, Fla.) Vol. 84; pp. 476 - 487
• Main Authors: López, J.D., Litvak, V., Espinosa, J.J., Friston, K., Barnes, G.R.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Inc 01-01-2014; Elsevier; Elsevier Limited; Academic Press
• Subjects: Algorithms; Artificial Intelligence; Bayes Theorem; Bayesian analysis; Bayesian model selection; Biological and medical sciences; Brain - physiology; Brain Mapping - methods; Data Interpretation, Statistical; Electroencephalography - methods; Free energy; Fundamental and applied biological sciences. Psychology; Humans; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Magnetoencephalography - methods; MEG/EEG inverse problem; Multiple Sparse Priors; Noise; Pattern Recognition, Automated - methods; Reproducibility of Results; Sensitivity and Specificity; Technical Note; Vertebrates: nervous system and sense organs; Bayesian model selection; MEG/EEG inverse problem; Multiple Sparse Priors; Free energy; Magnetoencephalography; Electrophysiology; Electroencephalography; Models
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1016/j.neuroimage.2013.09.002

The MEG/EEG inverse problem is ill-posed, giving different source reconstructions depending on the initial assumption sets. Parametric Empirical Bayes allows one to implement most popular MEG/EEG inversion schemes (Minimum Norm, LORETA, etc.) within the same generic Bayesian framework. It also provides a cost-function in terms of the variational Free energy—an approximation to the marginal likelihood or evidence of the solution. In this manuscript, we revisit the algorithm for MEG/EEG source reconstruction with a view to providing a didactic and practical guide. The aim is to promote and help standardise the development and consolidation of other schemes within the same framework. We describe the implementation in the Statistical Parametric Mapping (SPM) software package, carefully explaining each of its stages with the help of a simple simulated data example. We focus on the Multiple Sparse Priors (MSP) model, which we compare with the well-known Minimum Norm and LORETA models, using the negative variational Free energy for model comparison. The manuscript is accompanied by Matlab scripts to allow the reader to test and explore the underlying algorithm.
•We revisit the Bayesian source inversion algorithm for M/EEG source reconstruction.•We provide a didactic and practical guide with software examples.•The aim is to help standardize the development of other schemes.