Comparison Between Discrete and Continuous Motor Imageries: toward a Faster Detection

International audienceA large number of Brain-Computer Interfaces (BCIs) are based on the detection of changes in sensorimotor rhythms within the electroencephalographic signal [1]. Moreover, motor imagery (MI) modifies the neural activity within the primary sensorimotor areas of the cortex in a similar way to a real movement [2]. In most MI-based BCI experimental paradigms, subjects realize a continuous MI, i.e. one that lasts for a few seconds, with the objective of facilitating the detection of event-related desynchronization (ERD) and event-related synchronization (ERS) [3]. Currently, improving efficiency such as detecting faster a MI is a major issue in BCI to avoid fatigue and boredom. In this regards, a recent article showed that a brief intention of movement corresponding to a 2s-MI, leads to more informative ERS features than continuous motor imageries [4]. Thus, in this study, we are investigating differences between continuous MIs and discrete, i.e. simple short, MIs. Material, Methods and Results: 17 healthy subjects carried out real movements, discrete and continuous MIs, in the form of an isometric flexion movement of their right hand index finger. Each subject realized first a session of real movements, and then a discrete and a continuous sessions of motor imageries in a randomized order. Each session is divided into runs for a total number of 100 trials. Beeps were used as go and stop signals. Finally we computed ERD/ERS% for 9 electroencephalographic channels (FC3, C3, CP3, FCz, Fz, CPz, FC4, C4, CP4) using the " band power method " [3] (Fig. 1), topographic and time-frequency representations. Figure 1. Grand average (n = 17) ERD/ERS% curves estimated for the real movement (blue), the discrete motor imagery (red) and the continuous motor imagery (black) within the beta band (18-25 Hz) for electrode C3