| name | exploring-brain-networks-eeg-meg |
| description | Skill for exploring brain networks using noninvasive electrophysiological measurements (EEG/MEG) based on arXiv:2607.17602v1. Covers forward/inverse problems, source reconstruction, connectivity measures, and analysis pipelines. |
| license | Complete terms in LICENSE.txt |
| metadata | {"arxiv_id":"2607.17602v1","published":"2026-07-20","authors":["Unknown"]} |
| tags | ["eeg","meg","brain network","connectivity","source localization"] |
Exploring Brain Networks Using Noninvasive Electrophysiological Measurements
Based on arXiv:2607.17602v1 - "Exploring Brain Networks Using Noninvasive Electrophysiological Measurements: Methods and Applications"
Overview
This skill provides a practical guide for analyzing brain networks using EEG and MEG data. It covers the methodological foundations from forward/inverse modeling to connectivity analysis and practical workflows using open-source tools like Brainstorm.
Key Concepts
1. Forward and Inverse Problems
- Forward problem: Predicting sensor signals from known neural sources
- Inverse problem: Estimating neural sources from sensor measurements (ill-posed)
- Requires accurate head modeling and source reconstruction techniques
2. Source Reconstruction Techniques
- Minimum norm estimates (MNE)
- Beamforming (LCMV)
- Multiple sparse priors (MSP)
- Dipole fitting
3. Mitigating Volume Conduction and Signal Leakage
- Orthogonalization approaches
- Signal space projection (SSP)
- Surface Laplacian
- Imaginary part of coherency
- Phase lag index
4. Functional and Effective Connectivity Measures
- Functional (symmetric, undirected):
- Coherence
- Phase synchronization (PLV, PLI)
- Amplitude envelope correlation
- Effective (directed, causal):
- Granger causality
- Dynamic causal modeling (DCM)
- Transfer entropy
5. Analysis Pipelines
- Preprocessing (filtering, artifact removal)
- Source localization
- Connectivity estimation
- Statistical validation
- Visualization (brain networks, graphs)
Practical Workflow (Brainstorm-centric)
- Data Import: Load raw EEG/MEG files (EDF, BDF, FIF, etc.)
- Preprocessing:
- Bandpass filtering (typically 1-40 Hz)
- Artifact removal (ICA, SSP, regression)
- Bad channel detection/interpolation
- Head Modeling:
- Create volume conduction model (single sphere, realistically shaped)
- Align MRI with sensor positions
- Source Localization:
- Compute leadfield matrix
- Apply inverse method (MNE, beamforming)
- Extract source time series
- Connectivity Analysis:
- Choose appropriate measure based on hypothesis
- Compute connectivity matrices (frequency-specific if needed)
- Apply statistical thresholding (permutation testing, FDR)
- Network Analysis:
- Graph theoretical metrics (degree, betweenness, clustering, path length)
- Community detection
- Rich-club analysis
- Visualization:
- Source activations on cortical surface
- Connectivity matrices (circular, matrix plots)
- Brain networks (glass brain, force-directed layouts)
Emerging Approaches
- Time-varying connectivity: Sliding windows, state-space models, hidden Markov models
- Cross-frequency interactions: Phase-amplitude coupling, cross-frequency coherence
- Multivariate decoding: MVPA on source space, decoding networks
Tools and Resources
Validation and Best Practices
- Validate forward model with simulated dipoles
- Test inverse solutions with known source configurations
- Use surrogate data testing for connectivity measures
- Correct for multiple comparisons
- Report parameters and preprocessing steps for reproducibility
Activation Keywords
eeg meg brain network connectivity source localization brainstorm mne python neuroscience electrophysiology
References
- arXiv:2607.17602v1 - Exploring Brain Networks Using Noninvasive Electrophysiological Measurements: Methods and Applications
- Brainstorm tutorials and documentation
- MNE-Python documentation
- FieldTrip tutorials
