Cluster channels to reduce the number of columns from 110 to 10

💬 Support & Discussions
1

**After converting raw EEG signals into epochs, I converted epochs into the panda DataFrame. The data frame consists of 110 columns. These columns are the channel names corresponding to the standard montage GSN-HydroCel-129.

For optimization purpose, I want to reduce the columns from 110 to 10. How can I achieve that?
Any help will be highly appreciated. Thank you!
**

  • MNE-Python version: 0.19.2
  • operating system: Google Colab
epochs = mne.Epochs(raw, evs, event_id, tmin, tmax, proj=True,
                    picks=('eeg'), baseline=(None, 0.), preload=True)
epochs.pick_types(eeg=True, exclude='bads')

df = epochs.to_data_frame(scalings=dict(eeg=1, mag=1, grad=1))
print(df)

channel rows
channel rows946×445 20.8 KB

PS: The picture does not capture all the columns.

2

Try standard dimensionality reduction tools like PCA or ICA? It’s hard to offer any concrete advice without knowing what exactly you’re trying to optimize / what scientific question you’re trying to answer.