Brain-Computer Interfaces (BCIs) are direct connections between the brain and a computer. They translate brain signals into operational commands for technical devices. With BCIs severely impaired people can communicate (Kübler et al., 2001; Nijboer et al., 2008) or control arm movement (Pfurtscheller et al., 2003). Different electroencephalogram (EEG) signals have been applied to control a BCI, e.g. slow cortical potentials, sensorimotor rhythms, event-related potentials in healthy and severely motor impaired individuals (Hinterberger et al., 2004; Kübler et al., 2005; Sellers et al., 2006). But also after about 30 years of research on BCIs and multiple improvement e.g., in signal acquisition and classification, there is little knowledge about the phenomenon, that some people – healthy as well as individuals with disease – are not able to learn BCI-control. We refer to this phenomenon as “BCI-illiteracy”. The number of users from whom we cannot record a classifiable brain signal varies between 10% to 30% depending on the input signal and health status of participants.
There is little research about the determinants and predictors of learning how to control a BCI within a neurofeedback paradigm. Daum and colleagues (1993) showed that memory span and attention were correlated to the regulation of slow cortical potentials (SCP) in a sample of epileptic patients. Holzapfel (1998) replicated these results partially; memory span, personality factors and “dealing with stressful situations” predicted best the BCI performance.
Neumann and Birbaumer (2003) showed that the initial performance could predict future performance in a BCI based on SCP in a sample of five severely paralysed patients. Burde and Blankertz (2006) reported a significant correlation of r = 0.59 between the SMR-BCI feedback performance and the variable “locus of control by dealing with technology”. Subjects who felt confident with controlling a technical device performed better in the SMR-BCI.
To elucidate the “BCI-illiteracy” phenomenon, we want to investigate whether physiological and psychological parameters, such as heart rate variability (HRV), imagery abilities, visual-motor coordination abilities, attention, intelligence, verbal and non-verbal learning abilities, personality traits, psychological well-being, motivation and mood, could predict BCI performance in a BCI controlled by modulation of sensorimotor rhythms (SMR) with motor imagery.
This work is supported by the DFG KU 1453/3-1.