Brain–computer-interface research: Coming of age

Brain–computer-interfaces (BCI) or brain–machineinterfaces use brain signals (electric, magnetic, metabolic) to activate external devices such as computers, switches, or prostheses. The last 5 years have seen an exponential increase in publications, particularly new algorithms classifying and transforming different types of brain signals into mechanical or electronic output (for an overview, see a special issue of IEEE transactions in biomedical engineering, edited by Nicolelis, Birbaumer and Mueller, 2004) and animal experiments using cellular neuroelectric patterns to reconstruct motor trajectories (for a summary, see Riehle and Vaaida, 2005). Patterns of spike trains and field potentials from multielectrode recordings represent astonishingly well-imagined or intended movements. These results have created enormous public interest and hope for a rapid solution of critical clinical problems such as communication in locked-in patients and movement restoration in patients with spinal cord lesions and chronic stroke. BCI research is characterized by a relative excess of methodological and experimental approaches and a lamentable lack of studies with clinical populations. The two papers published in this issue by Piccione et al. (2006) and Sellers and Donchin (2006) are notable exceptions. Both employed severely handicapped patients with amyotrophic lateral sclerosis (ALS) and other forms of motor paralysis demonstrating the successful application of a P300-guided BCI in four-directional curser-control (Piccione et al.) and spelling (Sellers and Donchin). The two excellent papers underlining the conviction (Wolpaw and McFarland, 2005) that non-invasive EEG-driven BCIs offer a realistic perspective for communication in paralyzed patients initially demonstrated in 1999 (Birbaumer et al., 1999) with two completely paralyzed patients suffering from ALS. The P300-based approaches proposed here, following the first P300-BCI was developed with a group of healthy volunteers by Farwell and Donchin (Farwell and Donchin, 1988; Donchin et al., 2000), have several advantages over the other non-invasive BCI-techniques particularly, the slow cortical potential BCIs used for paralyzed ALS patients (Kübler et al., 2002):