EegFun

Appearance

Plot Channel Spectrum ​

This demo demonstrates frequency domain analysis of EEG channels using power spectral density (PSD) visualization.

What is Power Spectral Density? ​

Power Spectral Density (PSD) shows the distribution of signal power across different frequencies:

  • Decomposes time-domain signal into frequency components

  • Reveals dominant rhythms (alpha, theta, beta, etc.)

  • Detects artifacts like line noise (50/60 Hz)

  • Quantifies frequency band power for analysis

EEG Frequency Bands ​

Standard frequency bands in EEG analysis:

BandFrequency Range
Delta0.5-4 Hz
Theta4-8 Hz
Alpha8-13 Hz
Beta13-30 Hz
Gamma>30 Hz

Visualization Options ​

Single Channel Spectrum:

  • View PSD for one channel

  • Identify dominant frequency peaks

  • Detect line noise or artifacts

Multiple Channel Comparison:

  • Compare spectral profiles across electrodes

  • Identify spatial patterns (e.g., posterior alpha)

  • Assess differential effects across regions

All Channels:

  • Overview of entire dataset

  • Quickly spot problematic channels

  • Identify global artifacts

Common Use Cases ​

1. Quality Control:

  • Line noise detection: Look for sharp peaks at 50/60 Hz

  • Artifact identification: High power in unexpected frequency ranges

  • Channel malfunction: Abnormal spectral profile compared to neighbors

Interpretation Guidelines ​

Posterior alpha peak:

  • Should be dominant in occipital channels

  • Typically 8-13 Hz

  • Suppressed with eyes open

Line noise:

  • Sharp peak at exactly 50 or 60 Hz

  • Often includes harmonics (100, 150, 200 Hz)

  • Indicates need for notch filtering

Technical Details ​

FFT Parameters:

The function uses Welch's method for PSD estimation:

  • Window size: Automatic based on data length

  • Overlap: 50% between windows

  • Window type: Hamming window

  • Frequency resolution: Determined by window size

Best Practices:

Data requirements:

  • Minimum duration: A few seconds (longer = better frequency resolution)

  • Sample rate: Should capture frequencies of interest (≥2× max frequency)

  • Artifact-free data: Remove extreme values before spectral analysis

Filtering considerations:

  • High-pass: Remove DC offset and slow drifts (≥0.1 Hz)

  • Low-pass: Anti-aliasing already applied during acquisition

  • Notch filtering: Consider before spectrum if you want to see line noise

Selecting channels:

  • Single channel: For targeted analysis (e.g., Oz for alpha)

  • Regional selection: Compare frontal vs posterior, left vs right

  • All channels: Initial quality control and overview

Workflow Summary ​

This demo shows a simple workflow:

  1. Load data (raw continuous data)

  2. Create EEG structure with channel layout

  3. Visualize spectra:

  • All channels at once (overview)

  • Single channel (e.g., Fp1 for EOG artifacts)

  • Channel groups (e.g., frontal channels for beta analysis)

Code Examples ​

Show Code
julia
# Demo: Channel Spectrum Plotting
# Shows power spectrum visualization for selected channels.

using EegFun
# Note: EegFun.example_path() resolves bundled example data paths.
# When using your own data, simply pass the file path directly, e.g.:
# dat = EegFun.read_raw_data("/path/to/your/data.bdf")

# read raw data
dat = EegFun.read_raw_data(EegFun.example_path("data/bdf/example1.bdf"));
dat = EegFun.create_eegfun_data(dat);

# Plots
EegFun.plot_channel_spectrum(dat)
EegFun.plot_channel_spectrum(dat, channel_selection = EegFun.channels([:Fp1]), title = "Fp1 Power Spectrum")
EegFun.plot_channel_spectrum(dat, channel_selection = EegFun.channels([:Fp1, :Fp2, :F3, :F4]), title = "Frontal Channels Power Spectrum")

See Also ​