Summer Course 2024: 14 Electroencephalography (EEG) with & without simultaneous fMRI by Pete Molfese
71
Published 2024-07-24
1. **Exogenous and Endogenous Peaks**:
- Exogenous peaks are driven by the characteristics of the stimuli (e.g., louder sounds produce larger N100 peaks).
- Endogenous peaks, like the P300, are related to internal processing and can be observed even with rapid stimulus presentation.
2. **Measurement and Analysis**:
- Early programs for measuring EEG peaks focused on amplitude, latency, and consistency across electrodes.
- Peaks like the P100 and P300 are well-studied, with specific characteristics and known sources.
3. **Comparability Across Lifespan**:
- Event-related potentials (ERPs) can be used to gather comparable data from infants and adults.
- ERPs can predict various behaviors and clinical conditions, such as reading ability and risk of dyslexia in infants.
4. **EEG Recording Procedures**:
- Described the process of recording EEG from infants, including head measurement, electrode placement, and data collection.
- Emphasized the importance of measuring impedance and conducting electrical activity properly.
5. **Signal-to-Noise Ratio**:
- Discussed the relationship between the number of trials and the signal-to-noise ratio in ERP studies.
- Advocated for high sampling rates and numerous electrodes to capture accurate data.
6. **Time-Frequency Analysis**:
- Introduced the concept of time-frequency analysis, including wavelets and Fast Fourier Transforms (FFTs).
- Highlighted the importance of analyzing oscillatory dynamics and spectral content in EEG data.
7. **Principal Component Analysis (PCA)**:
- PCA is used to identify temporal and spatial components in EEG data.
- PCA helps in testing differences between conditions and participants and is considered more reliable than peak analysis alone.
8. **Peaks and Variations**:
- Peaks in EEG data can be influenced by trial variations and the number of trials averaged.
- Emphasized the need to regularize or normalize peak sizes for accurate analysis.
9. **Comparing Infant and Adult EEG**:
- Showed differences in EEG responses to the same stimuli between infants and adults.
- Adult EEG responses are faster, more dynamic, and more integrated compared to infants.
Several Source Analysis topics are also covered
1. **Motivation for Source Analysis**:
- The interest in source analysis increased due to the appealing visual representations of brain activity seen in fMRI, leading EEG researchers to develop more refined source analysis methods.
2. **Importance of Data and Electrode Count**:
- Adequate data quantity and high electrode count (e.g., 256 channels) are crucial for effective source analysis. More sensors provide better spatial resolution and help in constructing accurate models.
3. **Geometry and Inverse Models**:
- Understanding the geometry of sensors and creating a connectivity model is essential. Choosing the right inverse model is important for accurate source localization.
4. **Minimum Norm Estimates (MNE)**:
- MNE is the de facto standard for source analysis. It assumes all brain areas can be active simultaneously but to different extents. MNE provides computational efficiency and reasonable solutions, though it may produce spatially smeared results.
5. **Beamforming**:
- Beamforming applies spatial filters to data to localize brain activity, allowing the analysis of specific brain regions with unique frequencies. SAM (Synthetic Aperture Magnetometry) is a notable beamforming technique developed at the NIH.
6. **Forward and Inverse Problems**:
- The forward problem involves predicting scalp topography given known brain sources and propagation models. The inverse problem, which is more challenging, involves deducing brain sources from scalp topography. Solutions to these problems require mathematical constraints and accurate models.