@article {3359, title = {A general method for assessing brain{\textendash}computer interface performance and its limitations.}, journal = {Journal of Neural Engineering}, volume = {11}, year = {2014}, month = {03/2014}, abstract = {Objective. When researchers evaluate brain{\textendash}computer interface (BCI) systems, we want quantitative answers to questions such as: How good is the system{\textquoteright}s performance? How good does it need to be? and: Is it capable of reaching the desired level in future? In response to the current lack of objective, quantitative, study-independent approaches, we introduce methods that help to address such questions. We identified three challenges: (I) the need for efficient measurement techniques that adapt rapidly and reliably to capture a wide range of performance levels; (II) the need to express results in a way that allows comparison between similar but non-identical tasks; (III) the need to measure the extent to which certain components of a BCI system (e.g. the signal processing pipeline) not only support BCI performance, but also potentially restrict the maximum level it can reach. Approach. For challenge (I), we developed an automatic staircase method that adjusted task difficulty adaptively along a single abstract axis. For challenge (II), we used the rate of information gain between two Bernoulli distributions: one reflecting the observed success rate, the other reflecting chance performance estimated by a matched random-walk method. This measure includes Wolpaw{\textquoteright}s information transfer rate as a special case, but addresses the latter{\textquoteright}s limitations including its restriction to item-selection tasks. To validate our approach and address challenge (III), we compared four healthy subjects{\textquoteright} performance using an EEG-based BCI, a {\textquoteright}Direct Controller{\textquoteright} (a high-performance hardware input device), and a {\textquoteright}Pseudo-BCI Controller{\textquoteright} (the same input device, but with control signals processed by the BCI signal processing pipeline). Main results. Our results confirm the repeatability and validity of our measures, and indicate that our BCI signal processing pipeline reduced attainable performance by about 33\% (21 bits/min). Significance. Our approach provides a flexible basis for evaluating BCI performance and its limitations, across a wide range of tasks and task difficulties.}, keywords = {brain-computer interface, information gain, information transfer rate, Neuroprosthetics, performance evaluation}, doi = {10.1088/1741-2560/11/2/026018}, url = {http://www.ncbi.nlm.nih.gov/pubmed/24658406}, author = {Jeremy Jeremy Hill and H{\"a}user, Ann-Katrin and Gerwin Schalk} } @article {2862, title = {Communication and control by listening: towards optimal design of a two-class auditory streaming brain-computer interface.}, journal = {Frontiers in Neuroscience}, volume = {6}, year = {2012}, month = {12/2012}, abstract = {Most brain-computer interface (BCI) systems require users to modulate brain signals in response to visual stimuli. Thus, they may not be useful to people with limited vision, such as those with severe paralysis. One important approach for overcoming this issue is auditory streaming, an approach whereby a BCI system is driven by shifts of attention between two simultaneously presented auditory stimulus streams. Motivated by the long-term goal of translating such a system into a reliable, simple yes-no interface for clinical usage, we aim to answer two main questions. First, we asked which of two previously published variants provides superior performance: a fixed-phase (FP) design in which the streams have equal period and opposite phase, or a drifting-phase (DP) design where the periods are unequal. We found FP to be superior to DP (p = 0.002): average performance levels were 80 and 72\% correct, respectively. We were also able to show, in a pilot with one subject, that auditory streaming can support continuous control and neurofeedback applications: by shifting attention between ongoing left and right auditory streams, the subject was able to control the position of a paddle in a computer game. Second, we examined whether the system is dependent on eye movements, since it is known that eye movements and auditory attention may influence each other, and any dependence on the ability to move one{\textquoteright}s eyes would be a barrier to translation to paralyzed users. We discovered that, despite instructions, some subjects did make eye movements that were indicative of the direction of attention. However, there was no correlation, across subjects, between the reliability of the eye movement signal and the reliability of the BCI system, indicating that our system was configured to work independently of eye movement. Together, these findings are an encouraging step forward toward BCIs that provide practical communication and control options for the most severely paralyzed users. }, keywords = {auditory attention, auditory event-related potentials, brain-computer interface, dichotic listening, N1 potential, P3 potential}, issn = {1662-453X}, doi = {10.3389/fnins.2012.00181}, url = {http://www.ncbi.nlm.nih.gov/pubmed/23267312}, author = {Jeremy Jeremy Hill and Moinuddin, Aisha and H{\"a}user, Ann-Katrin and Kienzle, Stephan and Gerwin Schalk} } @article {2144, title = {Robust EEG Channel Selection across Subjects for Brain-Computer Interfaces.}, journal = {EURASIP Journal on Advances in Signal Processing}, volume = {2005}, year = {2005}, month = {01/2005}, pages = {3103{\textendash}3112}, abstract = {

Most EEG-based brain-computer interface (BCI) paradigms come along with specific electrode positions, for example, for a visual-based BCI, electrode positions close to the primary visual cortex are used. For new BCI paradigms it is usually not known where task relevant activity can be measured from the scalp. For individual subjects, Lal et al. in 2004 showed that recording positions can be found without the use of prior knowledge about the paradigm used. However it remains unclear to what extent their method of recursive channel elimination (RCE) can be generalized across subjects. In this paper we transfer channel rankings from a group of subjects to a new subject. For motor imagery tasks the results are promising, although cross-subject channel selection does not quite achieve the performance of channel selection on data of single subjects. Although the RCE method was not provided with prior knowledge about the mental task, channels that are well known to be important (from a physiological point of view) were consistently selected whereas task-irrelevant channels were reliably disregarded.

}, keywords = {brain-computer interface, channel selection, Electroencephalography, feature selection, recursive channel elimination, support vector machine}, doi = {10.1155/ASP.2005.3103}, url = {http://www.researchgate.net/publication/26532072_Robust_EEG_Channel_Selection_across_Subjects_for_Brain-Computer_Interfaces}, author = {Schr{\"o}der, Michael and Lal, T.N and Hinterberger, T. and Bogdan, Martin and Jeremy Jeremy Hill and Niels Birbaumer and Rosenstiel, W. and Sch{\"o}lkopf, B}, editor = {Vesin J M, T EbrahimiEditor} }