EegFun

Appearance

Baseline ​

This demo demonstrates baseline correction methods for ERP data, including different baseline intervals and correction approaches for continuous, epoch, and ERP data.

What is Baseline Correction? ​

Baseline correction removes the pre-stimulus mean from each trial, ensuring that activity is measured relative to a neutral reference period. This is essential for ERP analysis because:

  • Removes slow drifts: Eliminates DC offsets that vary across trials

  • Standardizes pre-stimulus activity: Ensures comparable starting points across conditions

  • Enables amplitude interpretation: Makes measurements interpretable relative to baseline activity

Common Baseline Windows ​

Pre-stimulus period (most common):

  • Standard: -200 to 0 ms before stimulus onset

  • Captures typical pre-stimulus activity

  • Assumes stable activity before stimulus

Baseline Methods ​

Mean correction (default):

  • Subtract the mean voltage during baseline period

  • Standard approach in ERP research

  • Appropriate when baseline is stable

Best Practices ​

Baseline interval selection:

  • Choose based on experimental design and paradigm timing

  • Standard: -200 to 0 ms for most ERP paradigms

  • Document chosen interval in methods section

  • Keep consistent across conditions and participants

Timing considerations:

  • Ensure baseline period is free from stimulus overlap

  • Avoid including activity from previous trials

  • Verify baseline period is stable (visually inspect)

When to baseline:

  • Apply after filtering and before averaging

  • Can apply to epoched or average ERPs

IMPORTANT ​

Baseline correction assumes the pre-stimulus period represents neutral brain activity. This assumption may be violated in designs with:

  • Very short inter-trial intervals

  • Anticipatory activity (e.g., motor preparation)

  • Sustained activity from previous trials

Workflow Summary ​

This demo demonstrates baseline correction for different data types:

Baseline Continuous Data ​

  • Load raw BioSemi data

  • Apply baseline to entire continuous recording

  • Visualize DC offset removal in databrowser

  • Baseline to specific timepoint (t=0)

Baseline Epoch Data ​

  • Extract epochs around experimental events

  • Apply baseline at different timepoints (t=0, t=0.5)

  • Visualize effect on individual trials

  • Compare baseline choices

Baseline ERP Data ​

  • Average epochs into ERPs

  • Baseline to single timepoint (t=0, t=0.5)

  • Baseline to time interval (-200 to 0 ms)

  • Compare effects across baseline choices

Code Examples ​

Show Code
julia
# Demo: Baseline Correction
# Shows baseline correction on continuous, epoched, and ERP data using
# different baseline intervals and visualizes the effects.

# 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")

using EegFun
using GLMakie

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

# read and prepare layout file
layout = EegFun.read_layout(EegFun.example_path("layouts/biosemi/biosemi72.csv"));
EegFun.polar_to_cartesian_xy!(layout)

# create EegFun data structure (EegFun.ContinuousData)
dat = EegFun.create_eegfun_data(dat, layout);

# Plot 
EegFun.plot_databrowser(dat) # DC shift visible

# Baseline stuff and replot 
EegFun.baseline!(dat)
EegFun.plot_databrowser(dat) # now zero mean over all samples (reduced DC shift)

# NB. such DC shift is removed when applying a high-pass filter (e.g., 0.1 Hz)

# baseline to timepoint 
EegFun.baseline!(dat, (0, 0)); # timepoint = 0
EegFun.plot_databrowser(dat)

# EpochData
# Some minimal preprocessing (average reference and highpass filter)
EegFun.rereference!(dat, :avg)
EegFun.highpass_filter!(dat, 0.1)

# Create some epoched data
epoch_cfg = [
    EegFun.EpochCondition(name = "ExampleEpoch1", trigger_sequences = [[1]]),
    EegFun.EpochCondition(name = "ExampleEpoch2", trigger_sequences = [[2]]),
]
epochs = EegFun.extract_epochs(dat, epoch_cfg, (-0.2, 1.0))  # -200 to 1000 ms

EegFun.plot_epochs(epochs, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(epochs, (0.0, 0.0)) # baseline to t=0
EegFun.plot_epochs(epochs, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(epochs, (0.5, 0.5)) # baseline to t=0.5 (just for demo purposes!)
EegFun.plot_epochs(epochs, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(epochs, (-0.2, 0.0)) # baseline from -200 to 0 ms t=0 (common pre-event baseline)
EegFun.plot_epochs(epochs, channel_selection = EegFun.channels([:Fp1]))


# Create some ERP data
erps = EegFun.average_epochs(epochs)

EegFun.plot_erp(erps, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(erps, (0.0, 0.0)) # baseline to t=0
EegFun.plot_erp(erps, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(erps, (0.5, 0.5)) # baseline to t=0.5 (just for demo purposes!)
EegFun.plot_erp(erps, channel_selection = EegFun.channels([:Fp1]))

EegFun.baseline!(erps, (-0.2, 0)) # baseline to t=-0.2 to 0.0
EegFun.plot_erp(erps, channel_selection = EegFun.channels([:Fp1]))

# We can see the influence of baseline interval interactively using the plot_erp_measurement_gui
EegFun.plot_erp_measurement_gui(erps)

See Also ​