Blind Source Separation: Independent Component Analysis for EEG data with python-MNE package and SSVEP
This is an example package performing ICA noise removal for SSVEP data.
I. Required data format. Due to MNE-python policy, the data format has to be strictly abided.
For our purpose it is Biosemi128 data format (csv).
II. Usage.
ICA artifacts removal consists of two consecutive phases:
1. Extract and plot indepentent components.
*** Decide which of the components to remove ***
2. Remove components.
In this package these operations are devided (for tutorial purposes)
into separate scripts:
01_plot_ica_components.py
02_plot_ica_selected_channels.py
core file is bieg.py (Blind source separation Independent component
analysis for EeG)
User has to specify the following information/parameters:
* Input file location.
* Number of components to devide the data into.
* Electrodes (channels) to pick.
* INDICES of the components to remove (starting at 0.)
* (optionally) output path (by default it takes the input file's
path and simply adds "_cleaned" before a file extention.