Best practices for preprocessing low-density (8-channel) EEG for frontal theta and parietal alpha

Support & Discussions
, , ,
1

Hi!

I have implemented a preprocessing pipeline in Python (MNE) and would appreciate guidance on best practices for artifact correction and bad segment handling with this low-density setup. Here are the steps I currently use:

  1. Filtering: 1–40 Hz band-pass + 50 Hz notch filter

  2. Bad channel detection: PyPREP (flat channels, high-variance, low correlation)

  3. Artefact correction: currently considering methods such as wavelet-thresholding or ASR, but I am unsure which is most appropriate for 8-channel EEG.

  4. Bad segment detection: Amplitude (>±100 µV / 4 s), step (>50 µV / 2 ms), joint probability (Z>3)

  5. Re-referencing: currently using A1 as reference; we are considering REST but are unsure if it is advisable or feasible with only 8 channels.

My questions are:

  • What are the recommended artifact correction methods for low-density (8-channel) EEG, especially for preserving frontal theta and parietal alpha features?

  • Any guidelines or suggestions for detecting and handling bad segments reliably in this setup?

Any documentation, advice, or example scripts specifically tailored for low-density 8-channel EEG data would be extremely helpful.

Thank you very much for your support.

Best regards,

Carmen