THE WEIGHTED AVERAGE REFERENCE MONTAGE

In this article we describe the implementation of a linearly weighted average reference montage. This montage differs from the commonly used source derivation montage in that the reference includes all the scalp electrodes of the 10-20 system; and it differs from all other montages implemented thus far in that the weighting of the reference electrodes is based on directly measured interelectrode distances. Using EEG and computer generated signals we have been able to demonstrate that the weighted average reference montage combines topographic selectivity and accuracy in the display of both focal and regional background changes. In comparison to the source derivation montage, ectopic peaks and troughs of localized potential fields are less prominent and interhemispheric symmetry is better preserved. In comparison to the common average reference montage the frequent prominent ectopic displacement of high amplitude (e.g., vertex waves) or widespread potentials (e.g., alpha background activity) is suppressed. We believe that the spatial filtering characteristics of the weighted average reference montage, which are intermediate between those of the common average reference montage and the source derivation and Laplacian montages, will make it a useful alternative for topographic analysis. Our results also indicate that actual scalp measurements should be used to calculate reference electrode weighting factors because such measurements yield values that are substantially different from those derived from other methods of estimation presented thus far. A weighted average montage derived from pooled scalp measurements can be easily implemented using the weighting factors provided herein.