Reconstruction of the electrical functional structure of the human brain based on electroencephalography data for various head models and comparison with magnetic encephalography data

The method of functional tomography based on multichannel measurement data transforms a set of time series into a spatial distribution of electrical or magnetic sources and is successfully used to study biological and physical objects. This paper examines the method application to electroencephalography data for various head models. The inverse problem was solved using the equivalent current dipole approximation. A single-layer and a three-layer spherical conductor were considered as approximate models of the human head. In this study, we used electroencephalography and magnetic resonance imaging data obtained from the publicly available OpenNEURO data repository. The experiments were conducted at the University of Oklahoma Health Sciences Center, USA. 600-second electroencephalograms were recorded during a quiet, awake state with eyes closed. Thirty-four healthy volunteers participated. Data were recorded on a 128-channel electroencephalograph, allowing for a detailed examination of brain activity. To localize and visualize brain sources, T1-weighted structural magnetic resonance imaging scans with a resolution of 1 mm were acquired from all subjects. We performed a Fourier transform on the complete time series, which allowed us to obtain detailed multichannel spectra. All spectra show a broad alpha rhythm peak in the frequency band from 8 to 13 Hz. For all sources in this band, the inverse problem was solved in single-layer and three-layer head models, and three-dimensional maps of brain activity were constructed – electrical functional structures of alpha rhythm sources. It was found that the source distributions found from the electroencephalogram in the single-layer and three-layer models were in reasonable agreement, with the sources in the three-layer model being distributed more widely and in good agreement with the magnetic encephalography data. The method can be used to study the spatial distribution of brain activity based on electroencephalography data in realistic head models.