Talk:BCILAB

Some note from Arno about meeting with Christian

For preprocessing the data: bci_preproc

- This is mostly to preprocess the data - There are some pre-made preprocessing pipeline - It is possible to overwrite some parameters in these pipelines - Use of causal filter for instance (instead of bidirectional ones for offline data) - This return s

For classifying data: bci_train

- This is mostly to preprocess the data - There are some pre-made BCI pipeline - It is possible to overwrite some parameters in these pipelines

CSP visualization function

- Plot classification performance as a function of time and frequency - Uses a FFT window which position increases with time - Use CSP then LDA to find the best classification and plot performance accuracy

Classification function

ml_train - used to train the classifier ml_test - used to test the data using the model returned above