Towards a symbiotic brain–computer interface: exploring the application–decoder interaction

Abstract

For BCIs in general, knowing the dominant factor that affects the decoder performance and being able to respond to it is of utmost importance to improve system performance. For the P300 speller, the proposed tunable paradigm offers the possibility to tune the application to the decoder's needs at any time and, as such, fully exploit this application-decoder interaction.

Riemannian geometry for EEG-based brain-computer interfaces; a primer and a review

“…In this article we have provided a primer on the simpler Riemannian classification method, the minimum distance to mean (MDM), providing rationale on its efficacy without requiring any specific knowledge on differential geometry. 30 Instead, we have highlighted the simplicity of its application in practice and we have relied mainly on intuitive (geometrical) explanations. The Riemannian MDM approach is based entirely on two simple concepts: the distance between two data points and a mean of a number of them.…”

“…For P300-based BCI these lines of research have led to, inter varia, the introduction of language models for letter and word prediction [23][24][25], automatic pause detection [26], the use of faces for flashing symbols [24], the use of random groups or pseudo-random groups flashing instead of row-column flashing [27][28][29][30], the use of inter-stimulus intervals randomly drawn from an exponential distribution instead of constant [27], the dynamic stopping of flashing sequences [25,31], etc. For SSVEP-based BCI improvements of the interface include the use of precise tagging of the flickering so as to use phase information (e.g., [32]) and the use of smart flickering sequences such as code modulation [33], multi-phase cycle coding [34], etc.…”

“…The ability to use a BCI without calibration or after a short calibration has been recognized in the community as a priority for improving BCI usability. Toward this objective effort is put in the conception, analysis and testing of generic model classifiers and/or domain adaptation methods, allowing the so-called transfer learning, whereas data from other sessions and/or other subjects are used to initialize a BCI so as to start using it without calibration and also to increase the performance of low-performance users [25,30,[57][58][59][60][61][62][63].…”

“…It also allows keeping optimal performance by adapting to mental and environmental changes during the session, enabling continuous adjustments ("pursuing"), thus ensuring reliability and robustness throughout the session and across-sessions [30,62,[64][65][66].…”

Using the detectability index to predict P300 speller performance

“…While alternative stimulus presentation paradigms have been proposed to improve performance, e.g. [10, 46], these stimulus paradigms are usually constructed based on a set of rules using criteria where potential performance improvements are inferred rather than evaluated objectively. Sometimes, the stimulus presentation paradigms developed with these adhoc rules might not be the optimum or best configuration, e.g.…”