A new approach to spatial localization of EEG-based electrical activity

Abstract

Background. Accuracy and reliability of methods for spatial localization of brain activity based on electroencephalography (EEG) data do not lose relevance. Existing localization methods are coupled to specific difficulties due to ambiguity of the resulting solution.Ojective: verification of a new method for localizing brain activity using EEG data «Virtually Implanted Electrode» and demonstratoin of its capabilities.Material and methods. The new method was verified using deep brain stimulation (DBS) data. There were used the data from scalp electrical activity induced solely by potentials from two electrodes implanted into deep areas of human brain. It was achieved by filtering EEG signals at the stimulation frequency and excluding signals from other brain areas. A clinical case of applying the new method for EEG analysis of patient paroxysmal activity is presented as well.Results. «Virtually Implanted Electrode» method allowed to quite accurately determine the localization of DBS current sources. The effectiveness of using the method for analyzing the mechanisms of paroxysmal activity was demonstrated: based on change in potentials calculated in 33 sites of the brain (brain areas), three groups of areas were identified specifically contributing to development of paroxysmal activity. For comparison, the same artificial DBS sources were localized using other verified methods: the BrainLoc (Russia) and sLORETA (Switzerland) software providing substantially worse data.Conclusion. Identifying spatial localization of electrical potentials recorded on the scalp surface by using the new approach with the «Virtually Implanted Electrode» method shows high efficiency and reliability, demonstrating a clear advantage over other known localization methods.