Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
A model for types and levels of human interaction with automation
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Objective and subjective evaluation of online error correction during P300-based spelling
Advances in Human-Computer Interaction - Special issue on Using Brain Waves to Control Computers and Machines
| Hi-index | 0.00 |
A Brain-Computer Interface (BCI) directly translates patterns of brain activity to input for controlling a machine. The introduction of methods from statistical machine learning [1] to the field of brain-computer interfacing (BCI) had a deep impact on classification accuracy. It also minimized the effort needed to build up the skill of controlling a BCI system [2]. This enabled other fields of research to adapt methods from BCI research for their own purposes [3, 4]. Team PhyPA, the research group for physiological parameters of the chair for Human-Machine Systems (HMS) of the Technical University of Berlin, focuses on enabling new communication channels for HMS. Especially the use of passive BCIs (pBCI) [3, 4], not dependent on any intended action of the user, show a high potential for enhancing the interaction in HMS [5]. Additionally, as actual classification rates are still below the threshold for efficient primary control [6, 7] in HMS, we focus on establishing a secondary, BCI-based communication channel. This kind of interaction does not necessarily disturb the primary mode of interaction, providing a low usage cost and hence an efficient way of enhancement. We have designed several applications following this approach. Here we are going to present briefly the results from two studies, which show the capabilities arising from the use of passive and secondary BCI interaction. First, we show that a pBCI can be utilized to gain valuable information about HMSs, which are hard to detect by exogeneous factors. By mimicking a typical BCI interaction, we have been able to identify and isolate a factor inducing non-stationarities with a deep impact on the feature dynamics. The retained information can be utilized for automatically triggered classifier adaptation. And second, we show that pBCIs are indeed capable to enhance common HMS interaction outside the laboratory. With this, we would like to feed back our experiences made with the use of BCIs for HMS. We hope to povide new and useful information about brain dynamics which might be helpful for ongoing research in augmented cognition.