Development and usage of odML based OpenEHR archetypes in Electroencephalography
-
1
University of West Bohemia, Department of Computer Science and Engineering, Czechia
-
2
University of West Bohemia, NTIS - New Technologies for Information Society, Czechia
Electroencephalography (EEG) as a non-invasive brain waveforms measurement is being used even outside laboratories or hospital environment. Behaviour of our brain during daily mental activities could be an interesting subject of a long term recording and analysing. Affordable headsets (e.g. Mindwave, Brain link, one-purposed-like devices as Mindflex games etc.) are good examples of home EEG peripherals to BCI (Brain-Computer Interface) applications. Neurofeedback can be used not only to handle BCI apps or games, but even for brain training itself (e.g. software home-of-attention comes with the idea of virtual coach for brain). A user can train intentional evocation of his/her alpha and lower beta waveforms (meditation, concentration) and improve his/her attention. Another situation is, e.g., when a sport shooter wants to know a relation between the waveforms produced by his/her brain during various mental and physical states (training, competition, illness, stress etc.) and his/her shot score. Such results can be very useful, but to make this a reality, standardized data description (a proper set of metadata included) and suitable application are crucial.
An electrophysiology metadata set is a frequently debated topic. EEGBase (Ježek, Mouček 2012) presents a metadata set designed besides others by domain experts from Pilsen University Hospital. OdML (Grewe et al. 2011) provides a metadata format and also well described and respected set of metadata (definitions, datatypes, restrictions etc. included). The NIX project (Stoewer et al, 2014) implements odML terminology within the universal data format for electrophysiological data. Last but not least OEN (Le Franc 2014) (Ontology for Experimental Neurophysiology), which is currently under development, extends the scope of metadata set and increases its expressive power.
Nevertheless, the metadata set can never be finite. If the above mentioned shooter wants consider data from other domains as well (e.g. diet), the data becomes a metadata for the current domain (electrophysiology in that example). For that reason an idea of personal electronic health record system was proposed (Papež, Mouček 2015). It is a system, where the user will be able to store “any” bio data for which a particular domain description (in form of a module) exists. The modules would extend the system with new domains, analytic functions etc. An openEHR concept was chosen to solve this modularity. Its idea is to describe various domains (or their pieces) by three layers: 1) generic reference model (RM) (general data properties and structure); 2) specific domain archetype (concrete data structure, ontology bindings, and data restrictions); 3) data input templates (solves problems of specific cases like complementary restrictions, archetype subset, archetype composition etc.). Since reference models are immutable and templates are very specific for each implementation, archetypes are the core of a well described domain. Currently, no archetypes for electrophysiology exist in the public openEHR archetype repositories (so called Clinical Knowledge Managers, CMKs).
A first set of our electrophysiology archetypes was designed according to EEGBase metadata, which was merged with the odML metadata subset. Terms binding with any ontology was not considered in the solution. Current version is proposed as fully based on odML. Because odML is not full-fledged ontology its terms do not have the URIs. This issue had to be solved as well as “odML to openEHR archetypes” transformation technique. There are three transformation proposals – automatic, semiautomatic and manual.
Automatic approach considers that all archetypes are based on cluster reference model. Cluster RM provides loose tree structure of attributes (so-called datapoints) without any mandatory parts or restrictions (like e.g. protocol or history in case of Observation RM). It means that one archetype is created for one odML root section). Subsections are represented by nested clusters. OdML attributes can be mapped directly to the archetype datapoints, where complex data types (non-primitive data types) are substituted by slot datapoints (reference to other archetype, which represents complex data type). An explicit mapping between odML and openEHR data types is necessary.
Absence of specific RMs can be solved by semi-automatic approach. An odML section is manually mapped to a specific RM. Its subsections and attributes have to be assigned to a proper part of archetype (e.g. if the experiment is classified as an observation RM, than its protocol must be specified). This process could be nontrivial because odML attributes, crucial for a given RM, may be absent. This has to be fixed manually. The rest of the attributes can be mapped as in the previous approach.
Manual design of archetypes starts from scratch and uses odML only for its list of terms and definitions. This approach presupposes new description of the electrophysiology domain. The process is similar to the development of a new ontology. Therefore, an already existing ontology could simplify the development process.
While the first way limits the openEHR power, the third approach in connection with already existing and well-designed ontology could bring more benefits. Semi-automatic approach provides suitable solution for most cases.
As it was mentioned before, each datapoint in the archetype should be bound to a particular ontology/terminology term using a public unique ID/URI. Since single terms (properties) from odML terminologies do not have their own IDs, for that reason, their identification was proposed. It consists of the URL to the (root) section of odML repository (e.g http://portal.g-node.org/odml/terminologies/v1.0/experiment/electrophysiology.xml) and XPath to the given term (e.g. /odML[@version="1"]/section/property/name[text()="Type"]/../value[text()="EEG"] for property of type “EEG”).
For the next version, OEN is expected to be used. Before archetypes will be published in public CKM, their draft forms are available at https://github.com/NEUROINFORMATICS-GROUP-FAV-KIV-ZCU/sehr/tree/master/CKM.
Acknowledgements
The work was supported by the European Regional Development Fund (ERDF), Project "NTIS - New Technologies for Information Society", European Centre of Excellence, CZ.1.05/1.1.00/02.0090 and UWB grant SGS-2013-039 Methods and Applications of Bio and Medical Informatics.
References
Grewe J, Wachtler T and Benda J (2011). A bottom-up approach to data annotation in neurophysiology. Front. Neuroinform. 5:16. doi: 10.3389/fninf.2011.00016
Stoewer A, Kellner CJ, Benda J, Wachtler T and Grewe J (2014). File format and library for neuroscience data and metadata. Front. Neuroinform. Conference Abstract: Neuroinformatics 2014.
doi: 10.3389/conf.fninf.2014.18.00027
Papež V and Mouček R (2015). Archetypes Development in Electrophysiology Domain - Electroencephalography as a Personal EHR System Module. In HEALTHINF 2015. 8th International Conference on Health Informatics. Setúbal: SciTePress, 2015. s. 611-616. ISBN: 978-989-758-068-0
Jezek P and Moucek R (2012). System for EEG/ERP Data and metadata storage and management. Neural Network World, 22, 277-290. ISSN: 1210-0552
Le Franc Y, Bandrowski A, Brůha P, Papež V, Grewe J, Mouček R, Tripathy SJ and Wachtler T (2014). Describing neurophysiology data and metadata with OEN, the Ontology for Experimental Neurophysiology. Front. Neuroinform. Conference Abstract: Neuroinformatics 2014. doi: 10.3389/conf.fninf.2014.18.00044
Keywords:
OpenEHR,
odml,
Archetype,
ontology,
EEGbase,
metadata
Conference:
Neuroinformatics 2015, Cairns, Australia, 20 Aug - 22 Aug, 2015.
Presentation Type:
Poster, not to be considered for oral presentation
Topic:
General neuroinformatics
Citation:
Papež
V and
Mouček
R
(2015). Development and usage of odML based OpenEHR archetypes in Electroencephalography.
Front. Neurosci.
Conference Abstract:
Neuroinformatics 2015.
doi: 10.3389/conf.fnins.2015.91.00055
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
20 Apr 2015;
Published Online:
05 Aug 2015.
*
Correspondence:
Mr. Václav Papež, University of West Bohemia, Department of Computer Science and Engineering, Pilsen, Please Select, 30614, Czechia, v.papez@ucl.ac.uk