Imagining the P300 Speller: Good idea or nonsense?

Abstract

The so-called P300-BCI provided high information transfer rates if stimuli are presented in the visual modality in healthy participants and those with severe neurological disease alike. Visual presentation of stimuli constitutes a severe obstacle to those with no control of eye movement. The following study investigated a potential alternative to using any sensory modality for stimulation. Instead healthy participants were instructed to imagine the traditional flashing matrix. To facilitate such imagery, stimuli were first presented sequentially in the visual and auditory domain. In Experiment 1 (N = 10) we found a decline in performance when stimuli were presented sequentially as compared to randomly, but still a P300 was identifiable albeit lower in amplitude. In Experiment 2 (N = 23) the matrix was still presented as a support, but no visual stimulation occurred. Instead the stimulation frequency was indicated by auditory clicks in condition 1 and removed altogether in condition 2. The P300 amplitude was significantly higher in target than non-target stimulations in both conditions, and higher in the no stimulation condition than in the auditory condition. Selection accuracy was above chance level for 8 participants in the auditory and for 10 in the no stimulation condition. Taken together results indicate that a P300 can be generated by imagined stimulation, but the paradigm requires further investigation and improvement.