Decoding kinetic features of hand motor preparation from single-trial EEG using convolutional neural networks.

Building accurate movement decoding models from brain signals is crucial for many biomedical applications. Predicting specific movement features, such as speed and force, before movement execution may provide additional useful information at the expense of increasing the complexity of the decoding problem. Recent attempts to predict movement speed and force from the electroencephalogram (EEG) achieved classification accuracies at or slightly above chance levels, highlighting the need for more accurate prediction strategies. Thus, the aims of this study were to accurately predict hand movement speed and force from single-trial EEG signals and to decode neurophysiological information of motor preparation from the prediction strategies. To these ends, a decoding model based on convolutional neural networks (ConvNets) was implemented and compared against other state-of-the-art prediction strategies, such as support vector machines and decision trees. ConvNets outperformed the other prediction strategies, achieving an overall accuracy of 84% in the classification of two different levels of speed and force (four-class classification) from pre-movement single-trial EEG (100 ms and up to 1,600 ms prior to movement execution). Furthermore, an analysis of the ConvNet architectures suggests that the network performs a complex spatiotemporal integration of EEG data to optimize classification accuracy. These results show that movement speed and force can be accurately predicted from single-trial EEG, and that the prediction strategies may provide useful neurophysiological information about motor preparation.

A comparison of subject-dependent and subject-independent channel selection strategies for single-trial P300 brain computer interfaces.

Brain computer interfaces (BCI) represent an alternative for patients whose cognitive functions are preserved, but are unable to communicate via conventional means. A commonly used BCI paradigm is based on the detection of event-related potentials, particularly the P300, immersed in the electroencephalogram (EEG). In order to transfer laboratory-tested BCIs into systems that can be used by at homes, it is relevant to investigate if it is possible to select a limited set of EEG channels that work for most subjects and across different sessions without a significant decrease in performance. In this work, two strategies for channel selection for a single-trial P300 brain computer interface were evaluated and compared. The first strategy was tailored specifically for each subject, whereas the second strategy aimed at finding a subject-independent set of channels. In both strategies, genetic algorithms (GAs) and recursive feature elimination algorithms were used. The classification stage was performed using a linear discriminant. A dataset of EEG recordings from 18 healthy subjects was used test the proposed configurations. Performance indexes were calculated to evaluate the system. Results showed that a fixed subset of four subject-independent EEG channels selected using GA provided the best compromise between BCI setup and single-trial system performance.

Genetic feature selection to optimally detect P300 in brain computer interfaces.

A Brain Computer Interface is a system that provides an artificial communication between the human brain and the external world. The paradigm based on event related evoked potentials is used in this work. Our main goal was to efficiently solve a binary classification problem: presence or absence of P300 in the registers. Genetic Algorithms and Support Vector Machines were used in a wrapper configuration for feature selection and classification. The original input patterns were provided by two channels (Oz and Fz) of resampled EEG registers and wavelet coefficients. To evaluate the performance of the system, accuracy, sensibility and specificity were calculated. The wrapped wavelet patterns show a better performance than the temporal ones. The results were similar for patterns from channel Oz and Fz, together or separated.

DETECCIÓN DE POTENCIALES EVOCADOS RELACIONADOS A EVENTOS EN INTERFACES CEREBRO-COMPUTADORA MEDIANTE TRANSFORMADA WAVELET / DETECTION OF EVENT-RELATED POTENTIALS IN BRAIN-COMPUTER INTERFACES USING THE WAVELET TRANSFORM / DETECÇÃO DE POTENCIAIS EVOCADOS RELACIONADOS A EVENTOS EM INTERFACES CÉREBRO-COMPUTADOR USANDO TRANSFORMADA WAVELET

Una interfaz cerebro computadora (ICC) es un sistema que provee una forma de comunicación directa entre el cerebro de una persona y el mundo exterior. Para el presente trabajo se utilizaron ICC basadas en EEG utilizando el paradigma de potenciales evocados relacionados con eventos (PRE). El objetivo de este trabajo es resolver en forma eficiente el problema de clasificación, en el cual se tienen dos clases posibles: registros con respuesta (PRE) y registros sin respuesta. Para esto se propone evaluar el desempeño de una ICC utilizando la transformada wavelet diádica discreta (DDWT, del inglés Dyadic Discrete Wavelet Transform) y la transformada wavelet packet (WPT, del inglés Wavelet Packet Transform) como métodos de extracción de características para la detección de la señal de PRE. La base de datos utilizada posee registros de EEG de época única de diez sujetos sanos. A partir de los patrones temporales (registros sin post-procesamiento) se generaron cinco conjuntos de patrones wavelet luego de aplicar la DDWT y WPT mediante diferentes técnicas. Se evaluó el desempeño de cada conjunto de patrones wavelet y de los patrones temporales mediante un clasificador lineal de Fisher. Se encontró que los patrones DDWT filtrados a 16 Hz presentan resultados de clasificación superiores a los patrones temporales. De esta manera al mejorar la etapa de extracción de características se mejora la clasificación, y consecuentemente, el desempeño del sistema completo de una ICC.

A brain-computer interface (BCI) is a system that provides a direct communication between the brain of a person and the outside world. For the present work we used an EEG-based event-related evoked potentials BCI. This paper aims to efficiently solve the problem of classification, which has two possible classes: recordings with evoked-potentials (ERP) and recordings without them. We proposed to evaluate the performance of a BCI using the discrete dyadic wavelet transform (DDWT) and the wavelet packet transform (WPT) as feature extraction methods for ERP signal detection. The database consisted of single-epoch EEG recordings from ten healthy subjects. From temporal patterns (recordings without any post-processing), five wavelet patterns were generated after applying DDWT and WPT via different techniques. The performance of the wavelet and temporal patterns were analyzed with the Fisher linear classifier finding that DDWT patterns, filtered at 16 Hz, presented better classification results than temporal patterns. This means that improving the feature extraction step, improves classification, and consequently, the performance of the entire BCI system.

Uma interface cérebro-computador (BCI) é um sistema que fornece uma forma de comunicação direta entre o cérebro de uma pessoa e o mundo exterior. Para este trabalho foram utilizados ICC baseado EEG evocados usando o paradigma de potenciais relacionados a eventos (ERP). O objetivo deste trabalho é resolver de forma eficiente o problema de classificação, em que há duas classes possíveis: registros Respondidas (PRE) e registros sem resposta. Para isso é avaliar o desempenho de uma ICC usando a wavelet diádica transformada discreta (DDWT, Discrete Wavelet Diádica Inglês Transform) e transformar pacote wavelet (WPT Transformada Wavelet Packet Inglês) como métodos de extração de características para a detecção de sinal PRE. A base de dados utilizada tem apenas EEG registra o tempo de dez indivíduos saudáveis. A partir dos padrões temporais (sem registros de pósprocessamento), cinco conjuntos de padrões após a aplicação wavelet e WPT DDWT gerado por várias técnicas. O desempenho de cada conjunto de padrões de wavelet e padrões temporais usando um classificador linear Fisher foi avaliado. Descobrimos que os padrões DDWT filtrados para 16 Hz apresentaram resultados acima da classificação padrões temporais. Assim, para melhorar a classificação de estágio de extração de características é melhorada, e, consequentemente, o desempenho de todo o sistema no ICC.