Extending Brain-Computer Interface Access with a Multilingual Language Model in the P300 Speller.

Brain-computer interfaces (BCI) such as the P300 speller have the potential to restore communication to advanced-stage neuromuscular disease patients. Research has improved typing speed and accuracy through innovations including the use of language models. While significant advances have been made, implementations have largely been restricted to a single language, primarily English. It is unclear whether these improvements would extend to other languages that present potential technical hurdles due to different alphabets and grammatical structures. Here, we adapt a language model-based classifier designed for English to two other languages, Spanish and Greek, to demonstrate the generalizability of these methods. Online experimental trials with 30 healthy native English, Spanish, and Greek speakers showed no significant difference between performances using the different versions of the system (66.20 vs. 61.97 vs. 60.89 bits/minute). Extending these methods across languages allows for expanding access to BCI systems to other populations, particularly in the developing world.

Local surface conductivity of transition metal oxides mapped with true atomic resolution.

The introduction of transition metal oxides for building nanodevices in information technology promises to overcome the scaling limits of conventional semiconductors and to reduce global power consumption significantly. However, oxide surfaces can exhibit heterogeneity on the nanoscale e.g. due to relaxation, rumpling, reconstruction, or chemical variations which demands for direct characterization of electronic transport phenomena down to the atomic level. Here we demonstrate that conductivity mapping is possible with true atomic resolution using the tip of a local conductivity atomic force microscope (LC-AFM) as the mobile nanoelectrode. The application to the prototypical transition metal oxide TiO2 self-doped by oxygen vacancies reveals the existence of highly confined current paths in the first stage of thermal reduction. Assisted by density functional theory (DFT) we propose that the presence of oxygen vacancies in the surface layer of such materials can introduce short range disturbances of the electronic structure with confinement of metallic states on the sub-nanometre scale. After prolonged reduction, the surfaces undergo reconstruction and the conductivity changes from spot-like to homogeneous as a result of surface transformation. The periodic arrangement of the reconstruction is clearly reflected in the conductivity maps as concluded from the simultaneous friction force and LC-AFM measurements. The second prototype metal oxide SrTiO3 also reveals a comparable transformation in surface conductivity from spot-like to homogeneous upon reduction showing the relevance of nanoscale inhomogeneities for the electronic transport properties and the utility of a high-resolution LC-AFM as a convenient tool to detect them.

Electrically controlled transformation of memristive titanates into mesoporous titanium oxides via incongruent sublimation.

Perovskites such as SrTiO3, BaTiO3, and CaTiO3 have become key materials for future energy-efficient memristive data storage and logic applications due to their ability to switch their resistance reversibly upon application of an external voltage. This resistance switching effect is based on the evolution of nanoscale conducting filaments with different stoichiometry and structure than the original oxide. In order to design and optimize memristive devices, a fundamental understanding of the interaction between electrochemical stress, stoichiometry changes and phase transformations is needed. Here, we follow the approach of investigating these effects in a macroscopic model system. We show that by applying a DC voltage under reducing conditions on a perovskite slab it is possible to induce stoichiometry polarization allowing for a controlled decomposition related to incongruent sublimation of the alkaline earth metal starting in the surface region. This way, self-formed mesoporous layers can be generated which are fully depleted by Sr (or Ba, Ca) but consist of titanium oxides including TiO and Ti3O with tens of micrometre thickness. This illustrates that phase transformations can be induced easily by electrochemical driving forces.

Integrating language models into classifiers for BCI communication: a review.

OBJECTIVE: The present review systematically examines the integration of language models to improve classifier performance in brain-computer interface (BCI) communication systems. APPROACH: The domain of natural language has been studied extensively in linguistics and has been used in the natural language processing field in applications including information extraction, machine translation, and speech recognition. While these methods have been used for years in traditional augmentative and assistive communication devices, information about the output domain has largely been ignored in BCI communication systems. Over the last few years, BCI communication systems have started to leverage this information through the inclusion of language models. MAIN RESULTS: Although this movement began only recently, studies have already shown the potential of language integration in BCI communication and it has become a growing field in BCI research. BCI communication systems using language models in their classifiers have progressed down several parallel paths, including: word completion; signal classification; integration of process models; dynamic stopping; unsupervised learning; error correction; and evaluation. SIGNIFICANCE: Each of these methods have shown significant progress, but have largely been addressed separately. Combining these methods could use the full potential of language model, yielding further performance improvements. This integration should be a priority as the field works to create a BCI system that meets the needs of the amyotrophic lateral sclerosis population.

Incorporating advanced language models into the P300 speller using particle filtering.

OBJECTIVE: The P300 speller is a common brain-computer interface (BCI) application designed to communicate language by detecting event related potentials in a subject's electroencephalogram signal. Information about the structure of natural language can be valuable for BCI communication, but attempts to use this information have thus far been limited to rudimentary n-gram models. While more sophisticated language models are prevalent in natural language processing literature, current BCI analysis methods based on dynamic programming cannot handle their complexity. APPROACH: Sampling methods can overcome this complexity by estimating the posterior distribution without searching the entire state space of the model. In this study, we implement sequential importance resampling, a commonly used particle filtering (PF) algorithm, to integrate a probabilistic automaton language model. MAIN RESULT: This method was first evaluated offline on a dataset of 15 healthy subjects, which showed significant increases in speed and accuracy when compared to standard classification methods as well as a recently published approach using a hidden Markov model (HMM). An online pilot study verified these results as the average speed and accuracy achieved using the PF method was significantly higher than that using the HMM method. SIGNIFICANCE: These findings strongly support the integration of domain-specific knowledge into BCI classification to improve system performance.

TiO2--a prototypical memristive material.

Redox-based memristive switching has been observed in many binary transition metal oxides and related compounds. Since, on the one hand, many recent reports utilize TiO(2) for their studies of the memristive phenomenon and, on the other hand, there is a long history of the electronic structure and the crystallographic structure of TiO(2) under the impact of reduction and oxidation processes, we selected this material as a prototypical material to provide deeper insight into the mechanisms behind memristive switching. In part I, we briefly outline the results of the historical and recent studies of electroforming and resistive switching of TiO(2)-based cells. We describe the (tiny) stoichiometrical range for TiO(2 - x) as a homogeneous compound, the aggregation of point defects (oxygen vacancies) into extended defects, and the formation of the various Magnéli phases. Furthermore, we discuss the driving forces for these solid-state reactions from the thermodynamical point of view. In part II, we provide new experimental details about the hierarchical transformation of TiO(2) single crystals into Magnéli phases, and vice versa, under the influence of chemical, electrical and thermal gradients, on the basis of the macroscopic and nanoscopic measurements. Those include thermogravimetry, high-temperature x-ray diffraction (XRD), high-temperature conductivity measurements, as well as low-energy electron diffraction (LEED), x-ray photoelectron spectroscopy (XPS), and LC-AFM (atomic force microscope equipped with a conducting tip) studies. Conclusions are drawn concerning the relevant parameters that need to be controlled in order to tailor the memristive properties.

Application of Bayesian decomposition for analysing microarray data.

MOTIVATION: Microarray and gene chip technology provide high throughput tools for measuring gene expression levels in a variety of circumstances, including cellular response to drug treatment, cellular growth and development, tumorigenesis, among many other processes. In order to interpret the large data sets generated in experiments, data analysis techniques that consider biological knowledge during analysis will be extremely useful. We present here results showing the application of such a tool to expression data from yeast cell cycle experiments. RESULTS: Originally developed for spectroscopic analysis, Bayesian Decomposition (BD) includes two features which make it useful for microarray data analysis: the ability to assign genes to multiple coexpression groups and the ability to encode biological knowledge into the system. Here we demonstrate the ability of the algorithm to provide insight into the yeast cell cycle, including identification of five temporal patterns tied to cell cycle phases as well as the identification of a pattern tied to an approximately 40 min cell cycle oscillator. The genes are simultaneously assigned to the patterns, including partial assignment to multiple patterns when this is required to explain the expression profile. AVAILABILITY: The application is available free to academic users under a material transfer agreement. Go to http://bioinformatics.fccc.edu/ for more details.

Localized metallic conductivity and self-healing during thermal reduction of SrTiO3.

The occurrence of metallic conductivity in SrTiO3 single crystals is reported for reduction under low partial pressure of oxygen at 800 degrees C. This transition is shown to result from the formation of a high concentration of vacancy defects along a network of extended defects within the skin region. A self-healing phenomenon is observed for progressive reduction which causes the concentration of initially introduced defects to decrease in the course of heat treatment and leads to a breakdown of the metallic conductivity as well as a substantial loss of optical subgap absorption.