EegFun

Appearance

How to Filter EEG Data ​

This guide shows you how to apply different types of filters to your EEG data.

High-Pass Filtering ​

Remove slow drifts and DC offsets:

julia
using EegFun

# Apply 1 Hz high-pass filter (removes frequencies below 1 Hz)
EegFun.highpass_filter!(dat, 1.0)

# Standard 0.1 Hz filter (removes very slow drifts only)
EegFun.highpass_filter!(dat, 0.1)

When to use: Always apply high-pass filtering (0.1-1 Hz) to remove slow drifts.

Low-Pass Filtering ​

Remove high-frequency noise:

julia
# Apply 40 Hz low-pass filter (removes frequencies above 40 Hz)
EegFun.lowpass_filter!(dat, 40.0)

# For very clean data or sleep studies
EegFun.lowpass_filter!(dat, 30.0)

When to use: Typically 30-50 Hz to remove muscle artifacts and line noise.

Band-Pass Filtering ​

Isolate a specific frequency range:

julia
# Apply high-pass + low-pass to isolate a band
EegFun.highpass_filter!(dat, 8.0)
EegFun.lowpass_filter!(dat, 12.0)

There is no dedicated `bandpass_filter!` — combine a high-pass and low-pass filter instead.

Choosing Filter Parameters ​

Cutoff frequency — For ERP work, 0.1 Hz high-pass is standard. Use 1 Hz only for ICA preprocessing. Low-pass at 30-40 Hz removes most muscle noise.

Filter order — EegFun defaults to order 1 for high-pass and order 3 for low-pass (Butterworth IIR). Higher orders give sharper rolloff but increase ringing and phase distortion. The defaults are safe for most EEG work.

IIR vs FIR — IIR (default) is fast and suitable for most use cases. FIR filters have linear phase but require many more taps; use filter_method="fir" if you need guaranteed zero phase distortion beyond what filtfilt provides.

Zero-phase filtering — filtfilt (default) applies the filter forwards and backwards, eliminating phase distortion but effectively doubling the filter order.

julia
# Customise filter parameters
EegFun.highpass_filter!(dat, 0.1; order=2, filter_method="iir")
EegFun.lowpass_filter!(dat, 30.0; order=4, filter_method="fir")

Verifying Filter Response ​

julia
# Create a filter and inspect it
filter_info = EegFun.create_highpass_filter(0.1, 256.0)
EegFun.print_filter_characteristics(filter_info)
EegFun.plot_filter_response(filter_info)

See Also ​