The amount of event-related decrease of increase in alpha-band power is termed "Event-Related Desynchronization" or ERD. In this study, ERD is used as an index of cortical activation, which not only can be applied to monitor poststimulus but prestimulus processes as well. Usually, ERD is analyzed within a fixed range of alpha-frequency, typically within a range of 8-12 Hz. alpha-frequency, however, shows a considerable between subject variation, even in a sample of subjects with comparable age. The purpose of this study, thus is to show as to how the range of alpha-frequency affects the calculation of ERD. Two different methodological approaches were compared. First, ERD was calculated within a fixed range of 8-12 Hz. Second, ERD was calculated within an individual range, which was adjusted to the individually determined alpha-frequency of each subject. A sample of ten right handed subjects participated in an episodic memory experiment, in which target words, shown in a earlier session of the experiment, were to be identified. A time span of 375 ms preceding and following the presentation of a word was defined as the "pre-" and "poststimulus" interval. EEG-signals were recorded unipolarly from 29 leads. The results show a rather localized increase of ERD in the poststimulus interval. In the prestimulus interval on the other hand, the average magnitude of ERD is much lower and the topographical distribution less localized. Most importantly, however, when comparing the two methods of analyzing ERD, those ERD-values calculated within an individually determined alpha-band, are superior in order to differentiate between the pre- and poststimulus interval.