The feasibility of continuous decoding of self-initiated, self-selected hand movements to three-dimensional (3D) spatial targets from scalp electroencephalography (EEG) using linear decoders has been recently demonstrated. In this paper, we show that it is possible to train linear classifiers to decode hand movement direction to eight 3D spatial targets, in both planning and movement windows, using only the fluctuations in the amplitude of smoothed low-frequency signals from high-density scalp EEG. Taken together these results support the design of brain-computer interfaces (BCI) based on non-invasive scalp EEG signals and suggest that the current perception of the limits of EEG as a source signal for BCI applications merits further examination.
ASJC Scopus subject areas
- Computer Science(all)