Filtering#

Apply frequency filters to your Wearable Sensing EEG data to remove noise and isolate signals of interest.

Applying Filters#

  1. Tools → Filter data

  2. Set filter parameters:

    • High-pass: Removes slow drifts (e.g., 0.5 Hz)

    • Low-pass: Removes high-frequency noise (e.g., 40 Hz)

    • Notch: Removes specific frequencies (60 Hz in North America, 50 Hz in Europe/Asia)

  3. Click Apply

Power Line Noise

For Wearable Sensing recordings in North America, use 60 Hz notch filter. For Europe/Asia, use 50 Hz. You may also want to remove harmonics (120 Hz, 180 Hz for 60 Hz systems).

Common Filter Settings#

Different analyses require different filter settings:

ERP Analysis#

Recommended: 0.1 - 50 Hz

For specific ERP components, see:

Zhang, G., Garrett, D. R., & Luck, S. J. (2024). Optimal filters for ERP research II: Recommended settings for seven common ERP components. Psychophysiology, 61, e14530. https://doi.org/10.1111/psyp.14530

Spectral Analysis#

Recommended: 0.5 - 50 Hz

  • Removes very slow drifts while preserving oscillatory activity

  • Adjust upper limit based on sampling rate and analysis goals

Time-Frequency Analysis#

Recommended: 1 - 40 Hz

  • Broader than spectral to avoid edge artifacts in wavelet analysis

Filter Types#

MNELAB uses FIR (Finite Impulse Response) filters by default:

Advantages:

  • Linear phase (no temporal distortion)

  • Stable and predictable

Considerations:

  • Introduces group delay

  • Can create edge artifacts (use longer recordings or crop after filtering)

Filter Order

MNELAB automatically calculates appropriate filter order based on your settings. Higher-order filters have sharper cutoffs but longer edge artifacts.

Visualizing Filter Effects#

After filtering, compare before and after:

  1. Use the sidebar version selector to switch between original and filtered data

  2. Plot → Plot power spectral density to see frequency content changes

Performance Issues#

Problem: MNELAB is slow with large files
Solution:

  • Close other applications to free memory

  • Consider downsampling: Tools → Resample

  • Process data in shorter segments

Next Steps#

After filtering your Wearable Sensing data:

  1. Remove artifacts - Clean data with ICA

  2. Create epochs - Extract event-related segments

  3. Save filtered data: File → Save as (use FIF format for compatibility)

Resources#