EegFun

Appearance

Channel Repair ​

This demo demonstrates channel interpolation methods for repairing bad electrodes using spatial interpolation techniques.

What is Channel Repair? ​

Channel repair (interpolation) estimates the signal at bad electrodes using data from neighboring good channels. This preserves data quality while maintaining electrode count for spatial analyses.

When to Repair Channels ​

Channel repair is appropriate when:

  • Individual electrodes have poor contact/noisy signal

  • Isolated channels show excessive noise or artifacts

  • A small number of channels are affected while most data is clean

  • You need to maintain electrode count for spatial analyses (e.g., source localization)

Do not repair when:

  • Too many channels are bad

  • The entire dataset is noisy

  • Bad channels cluster together spatially

Interpolation Methods ​

Neighbor Interpolation:

  • Weighted average of spatially nearby electrodes

  • Fast and computationally efficient

  • Good for isolated bad channels

  • Requires neighbor calculation based on distance threshold

Spherical Spline:

  • Uses spherical spline functions to model scalp potential distribution

Best Practices ​

Timing:

  • Identify bad channels using quality metrics first

  • Repair before re-referencing (reference calculation needs all channels)

  • Repair before averaging or statistical analysis

Limits:

  • Avoid interpolating too many channels

  • Consider rejecting datasets with an excessive number of bad channels

Validation:

  • Visually verify repair quality (before/after comparison)

  • Check that interpolated channels match neighbors

Workflow Summary ​

This demo shows channel repair workflows:

Identify Channels to Repair ​

  • Load and preprocess data

  • Select specific channels for demonstration

  • Verify channels exist in the dataset

Calculate Neighbor Relationships ​

  • Compute spatial neighbors based on 3D electrode positions

  • Use distance threshold to define neighborhood

  • Required for neighbor interpolation method

Apply Interpolation ​

  • Test neighbor interpolation method

  • Test spherical spline method

  • Store original data for comparison

Validate Repair Quality ​

  • Compare interpolated vs. original data

  • Verify that interpolation changed the data as expected

  • Use visual inspection to confirm successful repair

Code Examples ​

Show Code
julia
# Demo: Channel Repair
# Shows neighbor interpolation and spherical spline for repairing bad 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"));

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

dat = EegFun.create_eegfun_data(dat, layout)

# minimal preprocessing
EegFun.rereference!(dat, :avg)
EegFun.highpass_filter!(dat, 1)

EegFun.plot_databrowser(dat)

# Select a channel to repair and make this channel noisy!
channel_to_repair = :Cz
dat.data[!, channel_to_repair] .+= randn(size(dat.data[:, channel_to_repair])) * 200 # v. noisy!

# We can now see this noisy channel in the databrowser
# NB. we can actually press "R" and select Cz and apply the repair in the browser
EegFun.plot_databrowser(dat)

# Try neighbor interpolation
EegFun.repair_channels!(dat, [channel_to_repair], method = :neighbor_interpolation)

# Cz is now repaired
EegFun.plot_databrowser(dat)

# Try neighbor interpolation
EegFun.repair_channels!(dat, [channel_to_repair], method = :spherical_spline)

# Cz is now repaired
EegFun.plot_databrowser(dat)

See Also ​