Pharmacoelectroencephalogra. phy Main Editor:W.M. Herrm!lnn (Berlin) Original Paper Neuropsychobiology 1999;40:150-157 Artifact Processing in Computerized Analysis of Sieep EEG - A Review Peter Anderer a Stephen Roberts g Alois Schlögl f Georg Gruber a Gerhard Klösch b Werner Herrmann h Peter Rappelsberger C Oliver Filz d Manel J. Barbanoji Georg Dorffner e Bernd Saletu a Departments of ·Psychiatry and bNeurology, and <institute for Neurophysiology, School of Medicine, University of Vienna, dlnstitute for Information Processing, Austrian Academy of Sciences, and eAustrian Research Institute for Artificiallntelligence, Vienna, (Department of Medicallnformatics, University of Technology, Graz, Austria; gDepartment of Electrical and Electronic Engineering, Imperial College, London, UK; hLaboratory of Clinical Psychophysiology, Department of Psychiatry, Free University of Berlin, Germany; iDepartment of Pharmacology and Psychiatry, University of Barcelona, Spain KeyWords Sieep EEG analysis· Artifact minimization . Artifact identification Abstract Quantitative analysis of sleep EEG data can provide valu- able additional information in sleep research. However, analysis of data contaminated by artifacts can lead to spurious results. Thus, the first step in realizing an auto- matic sleep analysis system is the implementation of a reliable and valid artifact processing strategy. This strat- egy should include: (1) high-quality recording techniques in order to minimize the occurrence of avoidable artifacts (e.g. technical artifacts); (2) artifact minimization proce- dures in order to minimize the loss of data by estimating the contribution of different artifacts in the EEG record- ings, thus allowing the calcl!lation ofthe 'corrected' EEG (e.g. ocular and ECG interference), and finally (3) artifact identification procedures in order to define epochs con- taminated by remaining artifacts (e.g. movement and muscle artifacts). Therefore, after a short description of KAR.GER. © 1999 S. Karger AG, Basel 0302-282X/99/0403-O 150$17 .50/0 Fax+4161 3061234 E-Mail karger@karger.cb Accessible online al : www.karger.com hup://BioMedNel.com/ karger the types of artifacts in the sleep EEG and some typical examples obtained in different sleep stages, artifact minimization and identification procedures will be re- viewed. Introduction In reeent years, inereasing efforts have been made to develop eomputer-assisted sleep analysis systems [for re- view see 1-3]. In 1993, a consensus report of the EC eon- eerted action 'Methodology for analysis ofthe sleep-wake- fulness continuum' was published by Kemp [4], reeom- mending the use of a standard format for digitized poly- graphie reeordings (EDF-format, Kemp et al. [5]) and of a time resolution of 1 s for deseribing sleep/wake-related signal eharaeteristies on a eontinuous seale. New ap- proaehes to the automatie analysis of human sleep were presented by Roberts and Tarassenko [6], Schaltenbrand et al. [7] and Pardey et al. [8]. Based on these previous results, a European projeet (,SIESTA - A new standard DipL·Jng. Peler Anderer Departmenl ofPsychialry, UniversilY ofVienna Währinger Gürtel 18-20 A-1090 Vienna (Auslria) Tel. +43 I 40400 3628, Fax +43 I 4025909, E-Mail peler.anderer@akh·wien.ac.al