Hi 輝大
I am assuming that you are working with “EEGLab” and want to remove artifacts via “FastICA”.
The main difference between “FastICA” and “standard ICA” algorithms (“runica” in EEGLAB) is computational efficiency.
“Standard ICA” algorithms iteratively adjust weight matrix for all components at once until they find a solution. This process is slow when number of channels for EEG device are large. “FastICA” uses a fixed-point iteration scheme. Instead of updating the entire system at once, it can find the independent components one by one (“deflation”) or all at once (“symmetric”) in a way that converges much more quickly.
Please refer to the following MATLAB documentation:
In order to perform “FastICA” with “EEGLab” please refer to following MATLAB answer
