Brain-computer interface (BCI) systems allow people to send messages or commands without moving, and hence can provide an alternative communication and control channel for people with limited motor function. In this study, we demonstrate a BCI system for orthosis control. Our BCI was asynchronous, meaning that subjects could move the orthosis whenever they wanted, instead of pacing themselves to external cues. Seven subjects each performed two tasks with a BCI that relied on steady state visual evoked potentials (SSVEPs). Although none of the subjects had any training, six subjects showed good control with a positive predictive value (PPV) higher than 60%. The overall PPV for all subjects reached 78% ±10%. However, the false positive rate was high, and some subjects dislike the flickering lights required in SSVEP BCIs. In follow-up work, we hope to reduce both the false positive rate and the annoyance produced by flickering lights by hybridizing this BCI with a "brain switch," which could allow people to turn the SSVEP system on or off using a second type of brain activity when they do not wish to control the orthosis. We also hope to validate this approach with people with tetraplegia.

No abstract

EEG-based discrimination between different motor imagery states has been subject of a number of studies in healthy subjects. We investigated the EEG of 15 patients with complete spinal cord injury during imagined right hand, left hand, and feet movements. In detail we studied pair-wise discrimination functions between the 3 types of motor imagery. The following classification accuracies (mean ± SD) were obtained: left versus right hand 65.03% ± 8.52, left hand versus feet 68.19% ± 11.08, and right hand versus feet 65.05% ± 9.25. In 5 out of 8 paralegic patients, the discrimination accuracy was greater than 70% but in only 1 out of 7 tetraplagic patients. The present findings provide evidence that in the majority of paraplegic patients an EEG-based BCI could achieve satisfied results. In tetraplegic patients, however, it is expected that extensive training-sessions are necessary to achieve a good BCI performance at least in some subjects.

The "Graz Brain-Computer Interface (BCI)" transforms changes in oscillatory EEG activity into control signals for external devices and feedback. These changes are induced by various motor imageries performed by the user. For this study, 2 different types of motor imagery (movement of the right vs. left hand or both feet) were classified by processing 2 bipolar EEG-channels (derived at electrode positions C3 and C4). After a few sessions, within some weeks, 4 young paraplegic patients learned to control the BCI. In accordance with the participants, decision-speed (trial length) was varied and the information transfer rate (ITR) was calculated for each run. All experimental runs have been feedback-runs employing a simple computer-game-like paradigm. A falling ball had to be led into a randomly marked target halfway down the screen. The horizontal position was controlled by the BCI-output signal and the trial length was varied by the investigator across runs. The goal was to find values for trial length enabling a maximum ITR. Three out of 4 participants had good results after a few runs. Analysis of their last 2 experimental sessions, each containing between 10 and 16 runs, showed that the trial length can be reduced to values around 2 s to obtain the highest possible information transfer. Attainable ITRs were between 5 and 17 bit/min depending on the participant's performance and condition.

No abstract