EPILEPSIAE – A European epilepsy database

https://doi.org/10.1016/j.cmpb.2010.08.011Get rights and content

Abstract

With a worldwide prevalence of about 1%, epilepsy is one of the most common serious brain diseases with profound physical, psychological and, social consequences. Characteristic symptoms are seizures caused by abnormally synchronized neuronal activity that can lead to temporary impairments of motor functions, perception, speech, memory or, consciousness.

The possibility to predict the occurrence of epileptic seizures by monitoring the electroencephalographic activity (EEG) is considered one of the most promising options to establish new therapeutic strategies for the considerable fraction of patients with currently insufficiently controlled seizures.

Here, a database is presented which is part of an EU-funded project “EPILEPSIAE” aiming at the development of seizure prediction algorithms which can monitor the EEG for seizure precursors. High-quality, long-term continuous EEG data, enriched with clinical metadata, which so far have not been available, are managed in this database as a joint effort of epilepsy centers in Portugal (Coimbra), France (Paris) and Germany (Freiburg).

The architecture and the underlying schema are here reported for this database. It was designed for an efficient organization, access and search of the data of 300 epilepsy patients, including high quality long-term EEG recordings, obtained with scalp and intracranial electrodes, as well as derived features and supplementary clinical and imaging data. The organization of this European database will allow for accessibility by a wide spectrum of research groups and may serve as a model for similar databases planned for the future.

Introduction

Epilepsy is one of the most common diseases of the human brain, with a prevalence of more than 3 million patients in Europe alone. Epilepsy is characterized by sudden changes in brain dynamics that lead to abnormal synchronization of extended brain networks, the so-called “seizures”. These seizures are characterized by transient impairments of sensation, thinking and motor control. In most patients, seizures are infrequent, occupying much less than 0.1% of the time. Due to their apparently unpredictable occurrence, patients are, however, suffering from restrictions in several domains, e.g., physically due to the risk of trauma, socially due to driving and occupational restrictions, psychologically due to a feeling of helplessness [1], [2]. Furthermore, a continuous prophylactic medical treatment is presently being offered that renders patients liable to side effects [3].

One third of the patients presently do not respond to a continuous prophylactic treatment in maximally tolerated dosages. Particularly for this large patient group, new treatment concepts have to be developed. Such concepts could be based on prediction-based warnings to patients or on prediction-based closed-loop interventions. These would change continuous treatment to timely targeted medical or electrical interventions prior to seizures [4].

For this purpose, the EU-funded project “EPILEPSIAE” (www.epilepsiae.eu) aims at the development of EEG-based seizure prediction algorithms that capture changes in electroencephalography (EEG) dynamics and use these for a warning of the patient. Up to now, visual inspection did not reveal precursors for seizures. So far, one of the factors limiting the evaluation of prediction algorithms have been limitations in the quality and duration of long-term EEG data available for a valid evaluation of seizure prediction methods and their performances [5], [6], and limited access of groups with knowledge in the field of time series analysis to such EEG data.

Accordingly, both in Europe and in the USA efforts are made to overcome this obstacle for the development of seizure prediction. The EPILEPSIAE project will gather the largest and most comprehensive epilepsy database existing worldwide. It is based on the common effort of three European epilepsy centers (Freiburg, Germany; Paris, France; and Coimbra, Portugal), which contribute EEG data from long-term monitoring of epilepsy patients as well as standardized annotations and clinical metadata. After its completion, it is planned that this database project will offer access to data for research groups throughout the world and will collaborate with a US database that is presently being designed.

We here report on the design and the resulting schema of the database. Additionally, we present client applications and the current status regarding content: the input and management of large data sets.

Section snippets

Background

Seizure prediction is based on the identification of precursors in long-term EEG time series. A major concern is access to suitable data for the systematic application and evaluation of seizure prediction methods. Particularly for research groups with no direct contact to epilepsy centers, access to clinical data often imposes severe constraints for the progress of research. This has been pinpointed already in 2001 [7] when the importance of freely available datasets was emphasized: “One recent

Design considerations

Since large EEG data sets are important for the statistical validity of prediction results, one of the most important goals of the project was to constitute the largest pool of data, by far surpassing currently available sources. The project's plan schedules a database content of 300 datasets, thereof 50 with invasive electrodes, within a three-years period. The database should be open to additional datasets integrated after the project duration, possibly also including data from experimental

Database schema

The initial step for the design of a database is the conceptual design that defines and describes the excerpt of the real world that is of particular interest for the considered database application. The most commonly used conceptual model is the Entity-Relation (ER) model [14], [15], which models the real world in form of entities, usually recognizable concepts, either concrete or abstract, and defines relationships between them. Furthermore, it associates the entities with describing

Status report

While the database schema was evolving during the early stages of the project, it has reached a mostly steady state where changes are rare and mostly concern domain table updates, for instance the addition of a seizure pattern. The focus of the work is currently shifting towards the preparation of the datasets and the development of client applications.

Discussion and future plans

Albeit still in the process of completion, the EPILEPSIAE database is already by far the most comprehensive and complete epilepsy database currently existing. The arousing interest for access to the database presently offered by the Freiburg Epilepsy Center [11] shows the general need for publicly available, high quality databases of long-term, continuous EEG recordings, not only for seizure prediction but also for numerous related research communities. Apart from that, numerous participation

Acknowledgements

We would like to sincerely thank the clinical teams at the Epilepsy Centers in Freiburg, Paris, and Coimbra. This work was supported by the European Union (Grant 211713), the German Federal Ministry of Education and Research (BMBF grant 01GQ0420), and the Excellence Initiative of the German Federal and State Governments.

References (31)

  • R. Andrzejak et al.

    Indications of nonlinear deterministic and finite dimensional structures in time series of brain electrical activity: dependence on recording region and brain state

    Phys. Rev. E

    (2001)
  • The Flint Hills Scientific ECoG database...
  • The Freiburg EEG database...
  • H. Feldwisch-Drentrup et al.

    Joining the benefits: Combining epileptic seizures prediction methods

    Epilepsia

    (2010)
  • D. Chamberlin, R.F. Boyce, SEQUEL: A Structured English Query Language. SIGMOD Workshop, vol. 1, 1974, pp....
  • Cited by (168)

    • Continental generalization of a human-in-the-loop AI system for clinical seizure recognition

      2022, Expert Systems with Applications
      Citation Excerpt :

      As a continental generalization study framework, we train the AI interface using a US-based dataset from the Temple University Hospital (TUH) (Shah et al., 2018) and run inference on an Australia-based dataset from the Royal Prince Alfred Hospital (RPAH) to recognize seizure. Before applying the AI to a realistic dataset with nearly 200 patients at RPAH, we also verified the model’s generalization by running inference only on the open public EPILEPSIAE dataset (30 patients) (Ihle et al., 2012; Klatt et al., 2012). The EPILEPSIAE and RPAH datasets were kept unseen during the training of the AI.

    • SeizNet: An AI-Enabled Implantable Sensor Network System for Seizure Prediction

      2024, 2024 19th Wireless On-Demand Network Systems and Services Conference, WONS 2024
    • Revisiting Seizure Prediction with Path Signatures: the Limitations of System Identification

      2024, 2024 Australian and New Zealand Control Conference, ANZCC 2024
    View all citing articles on Scopus
    View full text