EEG: Introduction, Advantages and Disadvantages

This course focuses on EEG (electroencephalography), but the principles discussed also apply to MEG (magnetoencephalography) and LFP (local field potential) signals.
EEG involves measuring voltage fluctuations over time using electrodes placed on the scalp.

Typical EEG data consists of voltage fluctuations over time from multiple channels (e.g., 64) placed on the head.
Electrodes are connected to the scalp using gel or a sponge to measure voltage activity from neuron populations.
EEG signals exhibit small, rapid deflections (some from noise, some from high-frequency neural activity) and slower fluctuations.
Consistent activity across multiple electrodes suggests broader neural activity, while unique activity on one electrode may indicate noise contamination.
EEG data is often analyzed in epochs, time-locked to specific events.
Time-locking involves aligning data to the onset of an event (e.g., stimulus presentation) marked by TTL pulses or markers.
These markers synchronize brain activity with behavioral events, such as picture appearance or button presses.

Neurons communicate through electrochemical signaling, involving the flow of ions into and out of the neuron.
This ion flow creates an electrical field due to the spatial asymmetry of ion movement.
While tiny electrodes can measure the electrical field of a single neuron, EEG uses larger electrodes on the scalp.
EEG electrodes measure the collective synchronous activity of many neurons (tens of thousands to millions).
The electrical fields from these neurons volume conduct through brain tissue, skull, and skin to be measured non-invasively.
EEG signals (wiggles) reflect the instantaneous activity of cell populations.

While the existence of EEG is well-understood (based on neurobiology, biochemistry, biophysics, and engineering), the "content" of EEG is less clear.
The challenge is to understand what specific EEG patterns mean in terms of cognition, thinking, memories, and perception.
Cognition arises from complex interactions between diverse brain cells, not just large-scale voltage fluctuations.
Analyzing EEG signals to extract meaningful information about brain computations is a central goal.

Direct Measure of Brain Activity: EEG directly measures electrical activity, the primary medium of information computation in the brain.
High Temporal Resolution and Precision: EEG can measure brain activity at a high rate (e.g., 1 kHz or higher), capturing rapid cognitive processes occurring on the order of tens to hundreds of milliseconds.
Rich Data: EEG signals contain a wealth of complex information, allowing for various physiologically inspired analyses.
Cross-Scale Linking: EEG facilitates linking findings across different spatial and temporal scales, methods, and species, due to consistent patterns of electrical activity at various levels of the brain.
Neural Oscillations: EEG captures neural oscillations (rhythmic, sinusoidal-like activity), which are conserved across species.
Species Comparison: Dominant temporal patterns of rhythmic activity are remarkably consistent across species, from insects to humans, indicating fundamental similarities in neural systems.
Example: Gamma band oscillations are observed in humans, macaques, cats, rats, mice, and even insects like honeybees and locusts.

Limited to Synchronous Large-Scale Potentials: EEG cannot measure the activity of single neurons due to the weakness of their electrical fields at the scalp.
Undetectable Local Computations: Many computations in the brain occur at spatial scales too small to be measured by EEG.
Cancellation of Electrical Fields: Electrical fields from geometrically opposing dipoles can cancel each other out, leading to no measurable change in voltage, even with significant activity.
Uncertainty in Anatomical Localization: Determining the precise location in the brain from which an EEG signal originates is challenging, with inherent uncertainties in spatial resolution.
Complexity of Data: EEG data are complex, multidimensional, and contain noise, making physiologically relevant data analyses difficult, time-consuming, and potentially frustrating.
Temporal Precision Limitations: High temporal precision can be a disadvantage when studying slower cognitive processes with uncertain onset times.
Example: Studying hypothesis generation while reading, where the exact moment of hypothesis formation is unknown.

EEG is valuable for understanding brain function despite its limitations.
The next video will discuss artifacts in EEG data and their potential impact on analyses.