iBOTS
MNE-Python is an open-source Python package for exploring, visualizing, and analyzing human neurophysiological data (MEG, EEG, sEEG, ECoG, and more). It provides a complete pipeline from raw data to publication-ready figures.
Why MNE-Python?
- Comprehensive: Full analysis pipeline in one package
- Open Source: Free, transparent, and community-driven
- Well-Documented: Extensive tutorials and examples
- Publication-Ready: High-quality visualizations
- Active Development: Regular updates and new features
Key Features
Data Import and Preprocessing
import mne
# Load raw data
raw = mne.io.read_raw_fif('data.fif', preload=True)
# Filter
raw.filter(l_freq=1.0, h_freq=40.0)
# Find and interpolate bad channels
raw.info['bads'] = ['EEG 001']
raw.interpolate_bads()
# Re-reference
raw.set_eeg_reference('average')
Epoching
# Create events from annotations
events, event_id = mne.events_from_annotations(raw)
# Extract epochs around events
epochs = mne.Epochs(
raw, events, event_id,
tmin=-0.2, tmax=0.5,
baseline=(None, 0),
preload=True
)
# Compute evoked responses
evoked = epochs.average()
evoked.plot()
Time-Frequency Analysis
from mne.time_frequency import tfr_morlet
# Compute power over time and frequency
freqs = np.arange(8, 40, 2)
power = tfr_morlet(
epochs, freqs=freqs,
n_cycles=freqs/2,
return_itc=False
)
power.plot()
Source Localization
# Compute forward solution
fwd = mne.make_forward_solution(
raw.info, trans, src, bem
)
# Inverse solution
inv = mne.minimum_norm.make_inverse_operator(
raw.info, fwd, noise_cov
)
# Apply to data
stc = mne.minimum_norm.apply_inverse(evoked, inv)
Machine Learning
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression
from mne.decoding import Vectorizer
# Build classifier
clf = make_pipeline(
Vectorizer(),
StandardScaler(),
LogisticRegression()
)
# Decode from epochs
X = epochs.get_data()
y = epochs.events[:, -1]
clf.fit(X, y)
Analysis Pipeline
1. Preprocessing
- Import raw data
- Filter (highpass, lowpass, notch)
- Mark and interpolate bad channels
- Re-reference
- ICA for artifact removal
2. Epoching
- Define events of interest
- Extract time-locked epochs
- Reject bad trials
- Baseline correction
3. Analysis
- Evoked responses (ERPs)
- Time-frequency representations
- Connectivity analysis
- Statistical testing
4. Visualization
- Topographic maps
- Source space plots
- Time-frequency plots
- Statistical clusters
Getting Started
Install MNE-Python:
conda install -c conda-forge mne
# or
pip install mne
Basic example:
import mne
from mne.datasets import sample
# Load sample data
data_path = sample.data_path()
raw_fname = data_path + '/MEG/sample/sample_audvis_raw.fif'
raw = mne.io.read_raw_fif(raw_fname, preload=True)
# Basic preprocessing
raw.filter(1, 40)
raw.set_eeg_reference('average')
# Visualize
raw.plot()
Tips
- Start with tutorials on the MNE website
- Use
raw.infoto check recording parameters - Always visualize data at each processing step
- Use ICA to remove eye blinks and heartbeat
- Leverage MNE’s interactive plotting tools
- Check the MNE discourse forum for help
- Consider using MNE-BIDS for data organization