Alvaro Fuentes Cabrera

Auditory and spatial navigation imagery in Brain–Computer Interface using optimized wavelets

Features extracted with optimized wavelets were compared with standard methods for a Brain-Computer Interface driven by non-motor imagery tasks. Two non-motor imagery tasks were used, Auditory Imagery of a familiar tune and Spatial Navigation Imagery through a familiar environment. The aims of this study were to evaluate which method extracts features that could be best differentiated and determine which channels are best suited for classification. EEG activity from 18 electrodes over the temporal and parietal lobes of nineteen healthy subjects was recorded. The features used were autoregressive and reflection coefficients extracted using autoregressive modeling with several model orders and marginals of the wavelet spaces generated by the Discrete Wavelet Transform (DWT). An optimization algorithm with 4 and 6 taps filters and mother wavelets from the Daubechies family were used. The classification was performed for each single channel and for all possible combination of two channels using a Bayesian Classifier. The best classification results were found using the marginals of the Optimized DWT spaces for filters with 6 taps in a 2 channels classification basis. Classification using 2 channels was found to be significantly better than using 1 channel (p<<0.01). The marginals of the optimized DWT using 6 taps filters showed to be significantly better than the marginals of the Daubechies family and autoregressive coefficients. The influence of the combination of number of channels and feature extraction method over the classification results was not significant (p=0.97).

EEG based BCI-towards a better control. Brain-computer interface research at aalborg university

This paper summarizes the brain-computer interface (BCI)-related research being conducted at Aalborg University. Namely, an online synchronized BCI system using steady-state visual evoked potentials, and investigations on cortical modulation of movement-related parameters are presented.

Comparison of feature selection and classification methods for a brain-computer interface driven by non-motor imagery.

The aim of this study was to compare methods for feature extraction and classification of EEG signals for a brain–computer interface (BCI) driven by auditory and spatial navigation imagery. Features were extracted using autoregressive modeling and optimized discrete wavelet transform. The features were selected with exhaustive search, from the combination of features of two and three channels, and with a discriminative measure (r2). Moreover, Bayesian classifier and support vector machine (SVM) with Gaussian kernel were compared. The results showed that the two classifiers provided similar classification accuracy. Conversely, the exhaustive search of the optimal combination of features from two and three channels significantly improved performance with respect to using r2 for channel selection. With features optimally extracted from three channels with optimized scaling filter in the discrete wavelet transform, the classification accuracy was on average 72.2%. Thus, the choice of features had greater impact on performance than the choice of the classifier for discrimination between the two non-motor imagery tasks investigated. The results are relevant for the choice of the translation algorithm for an on-line BCI system based on non-motor imagery.

Brain-Computer Interfacing: How to control computers with thoughts

Brain-Computer Interface (BCI) technology aims at providing communication and control facilities to severely paralyzed people. These patients are not able to manipulate objects or communicate their needs, even though their mental capabilities are intact. Electroencephalographic (EEG) signals recorded from the scalp can be used to decode wishes and intentions. BCI approaches are based on a variety of strategies to generate control signals. For example, the control signals may be the result of visual or auditive stimulation or of imaginary motor tasks. The control signals are analyzed by a translation algorithm which associates a signal to a command. Thus, BCI provides a communication channel not based on nerves and muscles. This paper describes the BCI systems developed at the Center for Sensory-Motor Interaction of Aalborg University, with special emphasis on strategies based on non-motor imagery tasks.

Single-trial classification of auditory event-related potentials elicited by stimuli from different spatial directions

This study is focused on the single-trial classification of auditory event-related potentials elicited by sound stimuli from different spatial directions. Five naϊve subjects were asked to localize a sound stimulus reproduced over one of 8 loudspeakers placed in a circular array, equally spaced by 45°. The subject was seating in the center of the circular array. Due to the complexity of an eight classes classification, our approach consisted on feeding our classifier with two classes, or spatial directions, at the time. The seven chosen pairs were 0°, which was the loudspeaker directly in front of the subject, with all the other seven directions. The discrete wavelet transform was used to extract features in the time-frequency domain and a support vector machine performed the classification procedure. The average accuracy over all subjects and all pair of spatial directions was 76.5%, σ = 3.6. The results of this study provide evidence that the direction of a sound is encoded in single-trial auditory event-related potentials.

Smario: a toolbox for brain-computer interfacing analysis and design

In this paper we introduce Smario, a MATLAB open source toolbox for the analysis of BCI signals and implementation of translation algorithms for BCI systems. The Smario functions have been created based on the design of EEGLAB [1], they are accessible through the graphic user interface but they can also be run and edited using MATLAB syntax. Smario reads BCI2000 files in DAT and MAT formats, and documentation is available to implement functions to read other formats.

Autonomous wheelchair for patients with severe motor disabilities

This paper presents the design and implementation of a control strategy for an autonomous wheelchair to assist individuals suffering from severe motor disabilities. The user is presented with a pre-generated map of a known area (e.g. home, office) displayed on a computer screen, on which the location of the wheelchair is shown. Using a specially design man-machine interface the user can select the desired point to be transported to. After the the desired point has been selected on the map, the control algorithm calculates the path and transports the user to the destination, avoiding any obstacles on its way. A Bayesian estimation method, which takes into account the uncertainty inherent in the sensor measurements, is used to fuse the sensory information obtained from a laser, and to generate and update the occupancy grid map. The proposed system uses data from a probabilistic laser map to feed a Kullback Leiber Divergence KLD localization algorithm and path planning based on the solution of the Laplace’s equation. The system described in this manuscript is simulated in Matlab using actual measurements from a laser mounted on a mobile robot.