Overview of neuroinformatics infrastructure in Pilsen, CZ

Roman Mouček (Department of Computer Science and Eng..), Petr Ježek (Department of Computer Science and Eng..)

This abstract deals with a present neuroinformatics infrastructure at Department of Computer Science and Engineering, Pilsen, the Czech Republic. The research group at our department works within the Czech INCF National Node and focuses on EEG/ERP experiments. It includes storage, management, processing and analysis of EEG/ERP data.  As the result a number of software solutions have been developed. 

The core activity is development of the EEG/ERP portal. The portal was primarily developed as a custom software solution for storage and management of EEG/ERP experiments. Since then the portal has significantly changed. At present EEG/ERP experiments can be not only stored and managed, but e.g. the following functionalities have been added: management of experimental scenarios, a content management system and management of user groups, full text search, download history, or a laboratory booking system. At present, the portal is being registered within Neuroscience Information Framework (NIF).    

The inner structure of the portal (data/metadata structure) is provided as ontology using semantic web languages and technologies. Since the semantic web and its realization is still a matter of extensive research, two different approaches have been investigated and subsequently two independent software solutions have been introduced. The first approach uses the developed software tool that transforms the structure and content of underlying relational database into RDF/OWL. The second approach uses the software tool that adds the missing semantics into Java annotations (at the level of the object oriented code) and the corresponding RDF/OWL output is generated.      

Currently the experiments with storage of EEG/ERP data/metadata by using XML trees combined together with the relational database structure are performed.  The convertors of widely used EEG/ERP data/metadata formats (edf, vhdr, ascii) have been developed; they are tested on a set of various datasets.  

The developed library of methods suitable for EEG/ERP signal processing includes, in addition to basic methods, the wavelet transform, matching pursuit algorithm, and the modification of Hilbert-Huang transform. The library is being integrated into the EEG/ERP portal. Since signal processing methods usually have a high computational complexity, their local execution could be necessary.  They can be added as plugins into the JERPA/JUIGLE software.

The JERPA/JUIGLE is a local software application which has been developed to provide a user with EEG/ERP signal visualization and processing. It includes the following features:  signal visualization, epochs and artifacts marking, baseline correction, epochs averaging, integration of signal processing methods and their application to EEG/ERP signal, and export of results. Experiments can be downloaded directly from the EEG/ERP portal.

 Acknowledgements

This work was supported by Ministry of Education, Youth and Sport of the Czech Republic under the grant ME 949 and UWB grant SGS-2010-038 Methods and Applications of Biomedical Informatics.

Preferred presentation format: Poster
Topic: Electrophysiology

Document Actions