First, it is required that the 3D positions of the electrodes in the heart are known in order to make the spatial corre- lation. Second, it should be noted that the amplitude of late potentials in averaged signals in any lead will not only be influenced by the size and orientation of the underlying electrical wave front but also by the distance of the source to the electrode. The amplitude of a signal generated by a dipole will diminish rapidly with the distance between source and recording electrode. This means that irrespec- tive of the location of the substrate in the heart the maxi- mum amplitude of late potentials will usually be recorded in the precordial region. To avoid underestimation of the significance of late potentials in other chest areas, a correc- tion for the effect of electrode-source distance is needed. Since the potential measured at the surface is a result of the potential distribution at all points in the heart and a transfer matrix, this transfer matrix should be estimated. The transfer matrix could be computed from volume con- ductance theory and information derived from MR images. For practical purposes, however, a measure that can be de- rived from the data itself is preferable.