This presentation will take you on a 10-year journey and tells you about the challenges and opportunities we have faced in developing spontaneous electroencephalogram (EEG)-based ‘Plug and Play’ Brain-Computer Interfaces (BCIs). The objective of a BCI is to translate EEG patterns into messages. BCI user generate EEG patterns by performing associated mental tasks that elicit different mental states. Which mental task is best? How to ensure user reliably generate these patterns? How to train the BCI to robustly detect the pattern? How to do this outside of the lab? How with patients? How fully automatically?