– S. Barbieri (LATP) : Optimal Time-Frequency Bases for EEG Signal Classification in the Context of BCI.

Carte non disponible

Date/heure
Date(s) - 27/06/2013
14 h 00 min - 15 h 00 min

Catégories Pas de Catégories


Optimal Time-Frequency Bases for EEG Signal Classification in the Context of BCI. by Sebastiano Barbieri Abstract: We consider the problem of classifying multi-sensor signals, more precisely EEG signals in the context of Brain Computer Interfaces (BCI), by selection of time-frequency features. The features are determined among local cosine bases (MDCT) by a “best basis” type algorithm adapted to the classification context. In the BCI domain, or more generally in neuroscience, many classification algorithms are based upon automatic approaches (artificial neural networks, SVM, …) which do not allow a simple interpretation of the features. In the talk we propose an approach which allows such interpretation, since the features are determined in the form of time-frequency atoms, similarly to classic analyses of EEG signals which involve specific frequency bands and time intervals. The proposed algorithm generalizes the best discriminant basis algorithm by Saito, employing pairwise comparisons between the signals belonging to two classes of data. Results on artificial data show that the method is able to determine simulated differences between signals. Results on real data are competitive with state of the art classification algorithms and more easily interpretable.

– S. Barbieri (LATP) : Optimal Time-Frequency Bases for EEG Signal Classification in the Context of BCI.

Carte non disponible

Date/heure
Date(s) - 27/06/2013
14 h 00 min - 15 h 00 min

Catégories Pas de Catégories


Optimal Time-Frequency Bases for EEG Signal Classification in the Context of BCI. by Sebastiano Barbieri\n\nAbstract :\nWe consider the problem of classifying multi-sensor signals, more\nprecisely EEG signals in the context of Brain Computer Interfaces (BCI), by selection of\ntime-frequency features. The features are determined among local cosine bases (MDCT)\nby a “best basis” type algorithm adapted to the classification context.\n\nIn the BCI domain, or more generally in neuroscience, many classification algorithms\nare based upon automatic approaches (artificial neural networks, SVM, …) which do not\nallow a simple interpretation of the features. In the talk we propose an approach which allows such interpretation, since the features are determined in the form of time-frequency atoms, similarly to classic analyses of EEG signals which involve specific frequency\nbands and time intervals.\n\nThe proposed algorithm generalizes the best discriminant basis algorithm\nby Saito, employing pairwise comparisons between the signals belonging to two\nclasses of data. Results on artificial data show that the method is able to determine\nsimulated differences between signals. Results on real data are competitive with state of\nthe art classification algorithms and more easily interpretable.[