Towards a white matter ephaptic coupling model of 1/f spectra

Log-linear 1/f spectral densities are widely observed in the brain - across many levels of abstraction. However, the intrinsic relationship between the structural attributes of the brain and 1/f biorhythms remains largely unclear. We recently proposed a spike propagation model in white matter fiber bundles - based on cable theoretic principles capable of generating inter-hemispheric transfer times observed empirically. Our biophysical model incorporated the presence of unmyelinated and myelinated neurons in white matter fiber bundles across a range of axon calibers within a recursive computation framework. Here, we extend this biophysical model to include the effects of ephaptic coupling, or the ability of neurons to interact with their neighbors by altering the electric potential of the extracellular medium. We tested this model using diffusion MRI (dMRI) and MEG data derived from the Human Connectome Project (HCP) 1200 data set, both with and without ephaptic coupling. When applied in a generative capacity, our computational model with ephaptic coupling was capable of predicting realistic 1/f spectral signatures observed empirically in the HCP MEG data. We suggest spike volleys traveling along (deep) white matter tracts may superimpose with superficial gray matter cortical activity, and thus contribute to EEG/MEG spectral observations. Including these effects in source localization models may lay the foundation for improved noninvasive brain computer interfaces.