Questions: EEG Time-Frequency Analysis and Neural Oscillations

Question 1

A researcher averages 200 EEG trials time-locked to a stimulus onset and finds no consistent frontal theta power increase in the average, even though individual trials clearly show theta bursts. What is the most likely explanation?

• A. Theta is too low a frequency to be reliably detected from scalp EEG recordings
• B. The theta bursts are not phase-locked to the stimulus — they occur at variable latencies across trials, so averaging cancels them out
• C. 200 trials is statistically insufficient to detect theta power changes
• D. Frontal theta in this context is an artifact of eye movements rather than genuine neural activity

Answer: (B)

This is the fundamental distinction between ERPs and time-frequency analysis. Averaging preserves only phase-locked components — those occurring at the same latency on every trial. Theta oscillations related to working memory or encoding fluctuate in power across trials but not at a consistent post-stimulus latency, so averaging cancels them. Time-frequency analysis computes power in each trial's time windows independently before averaging the spectrograms, capturing this non-phase-locked activity. This is exactly why time-frequency methods are needed beyond standard ERP analysis.

Question 2

During a spatial attention task, a researcher observes alpha power increasing over right parietal cortex while subjects attend to the left visual field. What is the most accurate interpretation?

• A. Right parietal cortex is working harder to process the attended left-side information
• B. Right parietal cortex is being actively suppressed to prevent irrelevant right-side information from interfering with the attended left side
• C. The alpha increase is an artifact; genuine neural engagement always reduces alpha power
• D. Alpha increases reflect heightened arousal and general attentional readiness

Answer: (B)

Alpha is a marker of cortical inhibition, not engagement. Alpha increases over a region indicate active suppression of that region. During leftward spatial attention, right parietal cortex (which processes the right visual field) shows alpha increases because the system is inhibiting processing of irrelevant right-side input. The attended left visual field is handled by left parietal cortex, which shows alpha suppression. Option A inverts the relationship — engaged regions show alpha decreases, not increases.

Question 3

High gamma power in an EEG recording reliably indicates increased local cortical spiking activity.

• T. True
• F. False

Answer: False

High-frequency gamma can arise from EMG (electromyographic) muscle artifact — electrical signals from scalp and facial muscles that volume-conduct to EEG electrodes. This is one of the most common sources of false positives in cognitive EEG research. While genuine gamma can reflect local cortical processing and feature binding, the statement that high gamma reliably indicates neural spiking is false without careful artifact rejection. Gamma results must be interpreted with this confound in mind.

Question 4

Alpha suppression over a brain region is evidence that the region is actively engaged in processing relevant information.

• T. True
• F. False

Answer: True

Because alpha reflects cortical inhibition, a decrease in alpha (alpha suppression or desynchronization) over a region indicates that the inhibitory 'idle' state has been released and the region is actively processing. This makes alpha a dual-sided marker: increases reveal what is being suppressed; decreases reveal what is being engaged. Selective attention studies exploit this logic — attention to one location produces alpha suppression over regions processing that location and alpha increases over regions handling irrelevant locations.

Question 5

What is phase-amplitude coupling (PAC), and how does it offer a mechanistic account of working memory capacity limits?

Model answer: Phase-amplitude coupling is when the phase of a slow oscillation (e.g., theta, 4-8 Hz) modulates the amplitude of a fast oscillation (e.g., gamma, 30-100 Hz) — gamma bursts occur preferentially at specific phases of the theta cycle. This is thought to implement neural multiplexing: each gamma burst within a theta cycle can represent a different item held in working memory. The number of gamma cycles that fit within one theta cycle is finite, which constrains how many distinct items can be simultaneously maintained, providing a mechanistic explanation for working memory's limited capacity.

PAC operationalizes a hierarchical view of neural oscillations: slow oscillations provide temporal frames, fast oscillations encode content within those frames. The theta-gamma coupling model links a behavioral observation (capacity limits of ~4 items) to a physiological mechanism (the ratio of theta to gamma periods), generating testable predictions about how capacity limits should vary with oscillatory frequencies and cognitive load. It also connects EEG findings to hippocampal sequence coding, where theta-gamma coupling organizes spatial representations.