meeg-utils

MEG/EEG quality control, preprocessing and epoching on top of MNE-Python: scikit-learn style steps with documented defaults, QC metrics and figures for every step, BIDS in and out.

import meeg_utils as meu

raw = meu.io.read("bids/sub-01/eeg/sub-01_task-oddball_eeg.vhdr")
meu.qc.inspect(raw)                          # how well was it recorded?

pipe = meu.Pipeline.preset("eeg-erp")        # every parameter has a cited source
clean = pipe.fit_transform(raw)
pipe.plot(inst=raw)                          # figures of every step

Or a whole BIDS dataset from the command line:

meu qc bids/ --out qc/
meu run --preset eeg-erp --epochs --option "event_id=[target, standard]" \
    --sources bids/ --out bids/derivatives/meu

• Acquisition QC: bridged electrodes, flat channels, line noise, head movement, SQUID jumps, empty room, BIDS metadata, ...
• Steps for every system MNE reads: PREP, ZapLine-plus, SSS/tSSS, HFC, ICA with ICLabel/MEGnet, autoreject, ...
• Presets: eeg-erp, eeg-rest, meg-erp, meg-rest, had-meeg.

Installation

pip install meeg-utils

Documentation

Tutorials, user guide and API · design notes (Chinese) · validation on real data (Chinese)

MIT License.