This paper introduces a complete and elegant solution to the problem of inherent incompatibility of different formats used for digital storage of biomedical time series (in particular EEG) and their annotations.
We define a simple XML-based language, in which information on the structure of binary data files can be simply and efficiently coded. In most cases, the description of an existing format takes relatively few lines of XML code. Once written, this information can be used by any software, which, owing to this meta-description, may read the original data files, thus eliminating the need for conversions and duplication of data.
This proposition is hereby submitted to an open discussion within the
community involved in relevant research, clinical and commercial
applications. Links to the current version of the XML Schema defining
the language and the mailing list signalml-l@fuw.edu.pl dedicated to
this topic are located at http://eeg.pl/SignalML/. This site
offers also the first implementation: an Open Source multiplatform
software for display/annotation of the biomedical time series, which,
owing to the idea presented in this paper, can be easily adopted to
new data formats.