lina-usc / pylossless

🧠 EEG Processing pipeline that annotates continuous data
https://pylossless.readthedocs.io/en/latest/
MIT License
15 stars 8 forks source link

pipelines correlation calculation seems off #141

Open scott-huberty opened 10 months ago

scott-huberty commented 10 months ago

If you run this notebook: https://colab.research.google.com/drive/1XR867ptDDFnxmC62qFnSlZ6IbticR0G7?usp=sharing

You'll see that In the raw data, neighboring channels "EEG 015" and "EEG 016" have a correlation coefficient of .90.

According the pipeline's correlation calculation, both channels have an average max correlation of about .30

When even highly correlated channels come out with low correlations, it makes them group together with actual uncorrelated channels, making the bar for flagging any channel as low_r difficult.