Introduction: Scalp-recorded EEG signals can be used to calculate the locations of electrical sources in the brain. This procedure can improve the investigation of the functional organization of the human brain, exploiting the high temporal resolution of EEG to follow the temporal dynamics of information processing. As for today, the uncertainties about the effects of in-homogeneities due to brain lesions preclude the adoption of EEG functional mapping on patients with lesioned brain. Objectives: The aim of this work is to quantify the accuracy of distributed source localization methods in recovering extended sources of activated neural activity when lesions are introduced. Methods: A realistic brain structural connectivity network will be used to simulate brain neuronal dynamics by a system of neural masses coupled to one another with strengths proportional to the structural properties at each edge. Different structural lesions obtained removing nodes and connections within spatially defined regions will be introduced to simulate altered dynamics of the brain activity. For each simulated lesion and equivalent cerebral activity, pathological four-layer Boundary Element Method head models will be developed and forward and inverse calculations will be carried out to quantify localization errors. The simulation analysis will allow to determine the sensitivity of the inverse solution to both geometrical and conductivity parameters of the lesion. Expected results: These findings will allow to increase the confidence in source localization procedures, maintaining and assuring the accuracy of solutions of EEG inverse problems also in presence of a lesion. The validity of the developed methodology will be tested by comparing reconstructed neural activity from already available simultaneous extra- and intra-cranial real EEG data of epilepsy patients with brain lesion.
Expected results
The expected results of the project will provide an important theoretical advancement in the framework of source localization procedures, by extending their application to subjects with lesioned brains. The described methodology, based on physiologically plausible hypothesis and on the use of advanced techniques of signals processing, will allow to overcome the limitations imposed by traditional methods of analysis.
The proposed project will provide:
- A quantitative evaluation of the effects of the presence of brain lesions (with different geometrical and electrical properties) to the reconstruction of a realistic source-distributed activity by means of BEM head models;
- An improvement in the accuracy of the estimation of neural activity, allowing to selectively adjust the model complexity and include the a priori information about the lesion in the source estimation algorithms only when strictly necessary;
- An evaluation of the applicability of the developed methodology to appropriate real datasets of simultaneous EEG and Stereoelectroencephalography (sEEG) of epilepsy patients.
Impact
The described methodology, based on the use of advanced techniques of signals processing, will go beyond the state of the art by providing the scientific community with new tools, the first to be explicitly developed for the evaluation of the influence of brain lesions in the source localization procedures using BEM head models to estimate realistic source-distributed activity. This will provide guidelines about the conditions in which a source analysis can be performed without the need to build time/resources consuming models of cortical lesions and those in which such step is strictly necessary.
The new modeling approach will provide an important theoretical advancement in the framework of source localization procedures, with a potential clinical impact, advancing the diagnostic value of the EEG in the characterization and localization of functional abnormalities of the brain. This methodology will allow to take better advantage of the available high temporal resolution of EEG, extending the study of brain connectivity analysis in subjects with lesioned brains, so far limited to scalp EEG signals.
This new method for the study of neural activity will be a preliminary step toward the increase of the understanding of mechanisms at the basis of cortical plasticity and reorganization. From a clinical point of view, this may lay the foundation for future applications designed to the quantification of the structural and functional modifications in patients resulting from a neuro-rehabilitation intervention. Although the nature of this methodology will capture lesion effects only in their structural and dynamic impact, the future use of MRI-derived head models and functional information specific to the patient will allow the design of subject-specific simulated time courses for network activity; this will permit to probe directly the behavioral and cognitive effects of the lesion.
In perspective, such a model may point the way toward optimized rehabilitation strategies for patients with brain lesions, allowing the design of individualized treatments and location-specific rehabilitative interventions.