The reliability of event-related desynchronisation: a generalisability study analysis.

The aim of this study was to examine the spatio-temporal reliability of EEG power during the baseline period of an event-related desynchronisation (ERD) experiment. The data were analysed using ideas from generalisability theory. The largest sources of variance were found to be associated with differences between subjects (subject main effect), and to idiosyncratic topographical distributions (subject x electrode interaction). The resulting spatio-temporal reliabilities exceeded 0.7, except in theta and the absolute power values were relatively stable. Overall, this study gives further support for the reliability of quantitative EEG measures in general and for the method of ERD in particular.