Equivalent current dipole (ECD)
The equivalent current dipole (ECD) models a focal patch of synchronously active cortex as a single point current dipole with a position, orientation and magnitude. Dipole fitting is a parametric inverse method: one posits a small number of dipoles and estimates their parameters by nonlinearly minimizing the mismatch between the model's forward-predicted field and the measured field, usually over a short time window.
ECD modelling is powerful when the true generator really is focal and few in number — early sensory evoked responses, interictal epileptic spikes — giving a compact, interpretable answer. Its weaknesses are that the number of dipoles must be chosen in advance, the nonlinear fit can converge to local minima, and the model is misleading when activity is genuinely distributed. Fixed-position, rotating-dipole and moving-dipole variants trade off between stationarity assumptions and the number of free parameters.
A single tangential dipole fit to the N20 somatosensory evoked response reliably localizes to area 3b in the posterior wall of the central sulcus, matching its known anatomy.
ECD shines when the generator is genuinely focal and its count is known a priori.