Dynamic causal modelling of distributed electromagnetic responses.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biophysics

Background:

• Dynamic Causal Modelling (DCM) is widely used for inferring neural mechanisms from neuroimaging data.
• Conventional DCM often relies on equivalent current dipole (ECD) models, which may not accurately represent distributed neural activity.
• Electroencephalography (EEG) and magnetoencephalography (MEG) provide high temporal resolution for studying brain dynamics.

Purpose of the Study:

• To introduce a novel variant of DCM for evoked responses using EEG/MEG.
• To extend DCM from ECD formulations to spatially distributed source estimates.
• To provide a more biologically plausible model for localized, distributed neuronal activity on the cortical manifold.

Main Methods:

• Developed a distributed DCM spatial model based on neural-field equations.
• Approximated cortical activity using coupled local standing-waves.
• Modeled sources as a mixture of overlapping patches on the cortical mesh.
• Propagated time-varying activity through lead-field or gain matrices to sensor data.

Main Results:

• The distributed DCM model offers improved accuracy for locally distributed source activity compared to ECD models.
• The spatial degrees of freedom can be specified and optimized via model selection.
• The model's linearity in spatial parameters simplifies model inversion.
• Comparative evaluation with conventional ECD models for auditory processing using EEG was performed.

Conclusions:

• The proposed distributed DCM provides a more appropriate framework for modeling spatially distributed neural activity measured with EEG/MEG.
• This approach enhances the precision of spatiotemporal source estimation.
• The method offers advantages in model specification, optimization, and inversion for neuroscientific research.