A method for the calculation of induced band power: implications for the significance of brain oscillations W. Klimesch*, H. Russegger, M. Doppelmayr, Th. Pachinger Department of Physiological Psychology, Institute of Psychology, University of Salzburg, Hellbrunnerstr. 34, A-5020 Salzburg, Austria Accepted for publication: 17 September 1997 Abstract A method for the calculation of significant changes in induced band power (IBP) is presented. In contrast to traditional measures of event- related band power (ERBP) which are composed of evoked and not evoked EEG components, the proposed measure for IBP is deprived from phase locked (or evoked) EEG activity. It is assumed that changes in IBP reflect the modulation of brain oscillations that are largely independent from ERPs. The results of a visual oddball task show that significant changes in IBP can be observed in response to the presentation of a warning signal (preceding a target or nontarget) and the imperative stimulus (i.e. a target or nontarget) in the a, v and d band. Only a few significant changes in IBP were obtained for the warning signal in the v band although highly significant changes in ERBP were found. Our findings document that changes in IBP may be considered a phenomenon that is largely independent from the occurrence of ERPs. They underline the significance of oscillatory processes and suggest that induced rhythms are modulated by stimuli and/or events in a not phase locked way.  1998 Elsevier Science Ireland Ltd. Keywords: Induced rhythms; Oscillations; a; v; d; Event-related desynchronization; Synchronization 1. Introduction Quantitative EEG analyses are traditionally categorized into analyses in the time and frequency domain (Lopes da Silva, 1987). Event-related potentials (ERPs), calculated by additive averaging are the most prominent and well known example of analyses in the time domain. In the frequency domain, a variety of different measures such as event- related desynchronization or ERD (Pfurtscheller and Arani- bar, 1977), spectral coherences (Petsche and Rappelsberger, 1992) and special methods of frequency domain analyses of ERPs (Basar et al., 1992; Basar and Schu ¨rmann, 1994) are used. As divergent as these methods are, they all require the selection of certain frequency bands. Band pass filtering offers the possibility to analyze dif- ferent types of EEG responses in different frequency win- dows. As an example, it is well known that with increasing task demands v band power increases (e.g. the summary in Schacter, 1977) whereas the opposite holds true for the a band (Klimesch et al., 1994, 1996). If a broad frequency window is used and ERPs are calculated, these opposite effects of an increase and decrease in band power will tend to cancel each other. However, in contrast to ERPs which only reflect phase locked EEG activity, event-related band power measures such as ERD have the disadvantage to reflect both, phase locked and not phase locked EEG activ- ity. The purpose of this article is to demonstrate the useful- ness of a method for calculating band power changes that are deprived from phase locked EEG activity. The basic idea is to subtract the band pass filtered ERP (which repre- sents ‘pure’ phase locked activity) from the filtered EEG data. In assuming that the proposed measure reflects induced rhythms, it will be termed ‘induced band power’ or IBP. By ‘induced rhythms’ we understand oscillations that are modulated by stimuli or events and which (in con- trast to evoked rhythms) do not respond in a phase locked manner (for a similar but much broader definition see Bul- lock, 1992). Conventional measures of band power (which Electroencephalography and clinical Neurophysiology 108 (1998) 123–130 0168-5597/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0168-5597(97)00078-6 EEP 97563 * Corresponding author. Tel.: +43 662 80445120 or 5100; fax: +43 662 8044 5126; e-mail: Wolfgang.Klimesch@sbg.ac.at