Better Ways of Recording and Averaging Event-related Potentials
The development of analogue BCIs requires an interdisciplinary approach where engineering principles come to terms with (and exploit) the psychophysics of perception and attention. This is a big challenge.
The problem is not simply the need to plough the immense literature accumulated over more than 50 years of event-related-potential (ERP) research in psychology, psychophysiology, etc., in search for results applicable to the stimuli and tasks of a specific BCI. The biggest problem is that all studies use averaging techniques, which suffer from a magnifying-glass effect: they increase the resolution of specific ERPs at the cost of blurring other ERPs, thereby severely misrepresenting what really goes on in the brain.
Within our Analogue Evolutionary Brain-Computer Interfaces EPSRC
Project (EP/F033818/1) we devised, theoretically analysed and tested a simple technique which can radically improve the situation: binning trials based on response times and averaging. Results indicated that the method provably produces clearer representations of ERPs than any previous technique, revealing finer details and helping in the evaluation of the amplitude and other properties of ERP waves, which is very important in an analogue approach to BCI. We also found that we could further improve the resolving power of this method by combining it with an evolutionary algorithm.