ESPCI web site
SSVEP-Based Brain-Computer Interfaces
Emerging neuro-technologies, such as Brain-Computer Interfaces (BCI), are becoming increasingly important for society, as devoting more resources and attention to the elderly and to the disabled is becoming a stable social trend. A BCI platform has the fundamental task to identify reliably its user’s intentions from a limited set of brain activities in near-real-time. Such capability enables bypassing the normal executive functions of the body, and creation of a direct thought-powered link from the brain to a machine which can transfer information.
Designers of non-invasive Brain-Computer Interfaces face substantial challenges in their efforts to adapt this new mode of communication for robust and widely accepted use in society. Several experimental BCI paradigms exist each offering their unique set of advantages, among which SSVEP-BCI systems (based on visual flicker and selective attention) have demonstrated substantial promise due to their capability to provide a large number of commands with very high reliability, design flexibility, and little or no user training necessary. SSVEP-based BCI systems have been used, for example, as a two-command flight simulator control device, or as a four-command NASA Earth viewer. Nevertheless, signal-to-noise ratios of single-trial SSVEP brain responses are typically very low and transient, especially for small flicker stimuli, which requires the application of highly adaptive signal processing approaches.
Two different new eight-command SSVEP-BCI designs will be presented, one featuring very small, clustered, moving SSVEP stimuli for 2-D navigation of objects on a computer screen, and the other using a novel concept for brain response enhancement through flickering emotionally charged face videos for 3-D control of a multi-joint robotic arm. The affective-face SSVEP-BCI paradigm achieved a mean information transfer rate of 64 bits/min, a success rate of 99%, and minimal user fatigue. This work shows that new practical protocols for brain-based technologies should be optimized using two approaches. First, a systematic neurophysiological selection of essential stimulus features should be performed, and second, combined adaptive signal processing algorithms should be used to harness phase- and energy methods (e.g. for fast robust detection of SSVEP onset), especially in multi-command BCI systems. The demonstrated improvements in BCI efficiency would be beneficial for the disabled and the elderly, especially in unsupervised home settings, as well as in gaming- and hazardous remote-control BCI applications for healthy users.
Future approaches in BCI research will be also discussed, including Hybrid BCI, e-Home BCI, and BCI-based neuro-rehabilitation.