Branimir Leskošek

EMG archive: a local solution and a global proposal

Abstract:
The idea for local smooth muscle uterine EMG database/archive arose from the fact that within the Slovenian EMG group, EMG recordings were collected at different locations and dispersed on different media. With a centralised archive, those problems could easily be overcome. The only reasonable technical solution seemed the use of web technologies since the it is the most standardised and widely available technology for connecting distant locations.

The local EMG archive was built using open source LAMP (Linux OS, Apache web server, MySQL relational database, Perl/PHP programming language) technologies. The server and the access to the server are adequately protected regarding security and safety (compliant to ISO information security standards as ISO17799 and ISO27001). On the client side, only XHTML browser and internet connection is required.

The existing EMG archive database includes raw EMG recordings, raw and filtered images of EMG recordings and EMG recordings meta-data such as EMG data file name, date of measurement, measurement and channel description, sampling frequency, measured subject/animal and other descriptors.

Currently, data of 2338 uterine smooth muscles EMG recordings are collected in the EMG archive. Data browsing is enabled via user-friendly search engine. EMG recordings can be selected using text search in meta-data fields, setting recording time limits, subject/animal ID limits and other descriptors limits. Up to five hits per page can be browsed with interactive image zooming capability. Raw EMG data in Matlab/ASCII form can be downloaded for each hit separately or for all hits at once as a ZIP compressed file. Up-to-date EMG analysis scripts for Matlab can also be downloaded from the EMG archive web interface.

To enable easier and more efficient international collaboration, we propose to establish a global EMG archive (GEA) in which all (smooth-muscle) EMG research groups around the world would be invited to participate. The GEA database structure should be expanded from our local EMG archive at least with analyses descriptors. Furthermore, GEA should enable user-friendly EMG recording uploads (with validation), server-side EMG analyses for digital signal processing and statistical analyses, third-party analitical plugins (in different signal processing and statistical languages, e.g. Matlab, SPSS and R) and download of results. This would turn GEA into a proper data-warehouse, with OLAP and knowledge discovery capabilities.

Contact:
Dr. Branimir Leskošek
University of Ljubljana, Faculty of Medicine
Institute of Biomedical Informatics
Vrazov trg 2, SI-1104 Ljubljana, Slovenia
e-mail: brane.leskosek@mf.uni-lj.si

Copyright 2007 You.com. All Rights Reserved.
Dreamweaver Templates Resources